As filed with the U.S. Securities and Exchange Commission on September 30, 2024.

Registration No. 333-

UNITED STATES

SECURITIES AND EXCHANGE COMMISSION

Washington, D.C. 20549

FORM S-1

REGISTRATION STATEMENT

UNDER

THE SECURITIES ACT OF 1933

Cerebras Systems Inc.

(Exact name of registrant as specified in its charter)


Delaware	3674	81-2256092
(State or other jurisdiction of incorporation or organization)	(Primary Standard Industrial Classification Code Number)	(I.R.S. Employer Identification Number)

1237 E. Arques Avenue

Sunnyvale, California 94085

(650) 933-4980

(Address, including zip code, and telephone number, including area code, of registrant’s principal executive offices)

Andrew D. Feldman

Chief Executive Officer and President

1237 E. Arques Avenue

Sunnyvale, California 94085

(650) 933-4980

(Name, address, including zip code, and telephone number, including area code, of agent for service)

Copies to:


Tad J. Freese Sarah B. Axtell Latham & Watkins LLP 140 Scott Drive Menlo Park, California 94025 (650) 328-4600			Shirley X. Li General Counsel and Secretary Cerebras Systems Inc. 1237 E. Arques Avenue Sunnyvale, California 94085 (650) 933-4980			Alan F. Denenberg Elizabeth W. LeBow Davis Polk & Wardwell LLP 1600 El Camino Real Menlo Park, California 94025 (650) 752-2000

Approximate date of commencement of proposed sale to the public: As soon as practicable after the effective date of this registration statement.

If any of the securities being registered on this Form are to be offered on a delayed or continuous basis pursuant to Rule 415 under the Securities Act of 1933, check the following box. o

If this Form is filed to register additional securities for an offering pursuant to Rule 462(b) under the Securities Act, please check the following box and list the Securities Act registration statement number of the earlier effective registration statement for the same offering. o

If this Form is a post-effective amendment filed pursuant to Rule 462(c) under the Securities Act, check the following box and list the Securities Act registration statement number of the earlier effective registration statement for the same offering. o

If this Form is a post-effective amendment filed pursuant to Rule 462(d) under the Securities Act, check the following box and list the Securities Act registration statement number of the earlier effective registration statement for the same offering. o

Indicate by check mark whether the registrant is a large accelerated filer, an accelerated filer, a non-accelerated filer, a smaller reporting company or an emerging growth company. See the definitions of “large accelerated filer,” “accelerated filer,” “smaller reporting company” and “emerging growth company” in Rule 12b-2 of the Exchange Act.


Large accelerated filer	☐	Accelerated filer	☐
Non-accelerated filer	☒	Smaller reporting company	☐
		Emerging growth company	☒

If an emerging growth company, indicate by check mark if the registrant has elected not to use the extended transition period for complying with any new or revised financial accounting standards provided pursuant to Section 7(a)(2)(B) of the Securities Act. o

The Registrant hereby amends this registration statement on such date or dates as may be necessary to delay its effective date until the registrant shall file a further amendment which specifically states that this registration statement shall thereafter become effective in accordance with Section 8(a) of the Securities Act of 1933, or until the registration statement shall become effective on such date as the Securities and Exchange Commission, acting pursuant to said Section 8(a), may determine.

The information in this preliminary prospectus is not complete and may be changed. We and the selling stockholders may not sell these securities until the registration statement filed with the Securities and Exchange Commission is effective. This preliminary prospectus is not an offer to sell these securities and neither we nor the selling stockholders are soliciting offers to buy these securities in any jurisdiction where the offer or sale is not permitted.

SUBJECT TO COMPLETION, DATED , 2024

PRELIMINARY PROSPECTUS

Cerebras Systems Inc.

Class A Common Stock

$ per share

This is the initial public offering of our Class A common stock and no public market currently exists for shares of our common stock. We are selling shares of our Class A common stock and the selling stockholders identified in this prospectus are selling an aggregate of shares of Class A common stock. We will not receive any proceeds from the sale of shares of Class A common stock by any of the selling stockholders. We currently expect the initial public offering price to be between $ and $ per share of our Class A common stock.

We intend to grant the underwriters an option to purchase up to additional shares of our Class A common stock from us to cover over-allotments, if any, at the initial public offering price less the underwriting discount.

We have applied to list our Class A common stock on the Nasdaq Global Market under the symbol “CBRS.”

Following completion of this offering, we will have two classes of authorized common stock: Class A common stock and Class N common stock. The rights of the holders of Class A common stock and Class N common stock are identical, except with respect to voting and conversion rights. Each share of Class A common stock is entitled to one vote. Each share of Class N common stock is non-voting and is convertible into one share of Class A common stock. See the section titled “Certain Relationships and Related Party Transactions” for additional information regarding issuances of our Class N common stock following the completion of this offering.

We are an “emerging growth company” as defined under the federal securities laws and, as such, have elected to comply with certain reduced public company reporting requirements for this prospectus and future filings.

Investing in our Class A common stock involves risks. See “Risk Factors” beginning on page 20 to read about factors you should consider before deciding to invest in our Class A common stock.

Neither the Securities and Exchange Commission nor any state securities commission has approved or disapproved of these securities or determined if this prospectus is truthful or complete. Any representation to the contrary is a criminal offense.


	Per Share	Total
Initial public offering price	$	$
Underwriting discounts and commissions(1)	$	$
Proceeds to us (before expenses)	$	$
Proceeds to selling stockholders (before expenses)	$	$

_______________

(1)See the section titled “Underwriting” for a description of the compensation payable to the underwriters.

The underwriters expect to deliver the shares to purchasers on or about , 2024 through the book-entry facilities of The Depository Trust Company.


Citigroup						Barclays


UBS Investment Bank						Wells Fargo Securities


Mizuho						TD Cowen

Needham & Company

Craig-Hallum

Wedbush Securities

Rosenblatt

Academy Securities

, 2024

TABLE OF CONTENTS


			Page

Prospectus Summary			1

The Offering			12

Summary Consolidated Financial Data			15

Risk Factors			20

Special Note Regarding Forward-Looking Statements			66

Market and Industry Data			68

Use of Proceeds			69

Dividend Policy			71

Capitalization			72

Dilution			75

Management’s Discussion and Analysis of Financial Condition and Results of Operations			79

Business			102

Management			130

Executive and Director Compensation			138

Certain Relationships and Related Party Transactions			150

Principal and Selling Stockholders			155

Description of Capital Stock			159

Shares Eligible for Future Sale			167

Material U.S. Federal Income Tax Consequences to Non-U.S. Holders			172

Underwriting			176

Legal Matters			188

Experts			188

Where You Can Find Additional Information			188

Index to Consolidated Financial Statements			F-1

Through and including , 2024 (the 25th day after the date of this prospectus), all dealers that buy, sell, or trade shares of our Class A common stock, whether or not participating in this offering, may be required to deliver a prospectus. This delivery requirement is in addition to the obligation of dealers to deliver a prospectus when acting as underwriters and with respect to their unsold allotments or subscriptions.

As used in this prospectus, unless the context otherwise requires, references to “Cerebras Systems,” “Cerebras,” the “company,” “we,” “us,” “our,” and similar terms refer to Cerebras Systems Inc. and, where appropriate, its subsidiaries, taken as a whole.

“Cerebras,” “Cerebras Systems,” the Cerebras logos, and other trade names, trademarks, or service marks of Cerebras appearing in this prospectus are the property of Cerebras Systems Inc. Other trade names, trademarks, or service marks appearing in this prospectus are the property of their respective holders. Solely for convenience, trade names, trademarks, and service marks referred to in this prospectus appear without the ®, ™, and SM symbols, but those references are not intended to indicate, in any way, that we will not assert, to the fullest extent under applicable law, our rights or that the applicable owner will not assert its rights, to these trade names, trademarks, and service marks.

Numerical figures included in this prospectus have been subject to rounding adjustments. Accordingly, numerical figures shown as totals in various tables may not be arithmetic aggregations of the figures that precede them.

References to www.cerebras.ai in this prospectus are inactive textual references only, and the information contained on, or that can be accessed through, our website does not constitute part of this prospectus.

Neither we, the selling stockholders, nor the underwriters have authorized anyone to provide you any information or to make any representations other than those contained in this prospectus or in any free writing prospectus prepared by or on behalf of us or to which we have referred you. Neither we, the selling stockholders, nor the underwriters take responsibility for, or provide any assurance as to the reliability of, any other information others may give you. This prospectus is an offer to sell only the shares offered hereby, and only under circumstances and in jurisdictions where it is lawful to do so. We, the selling stockholders, and the underwriters are not making an offer to sell these securities in any jurisdiction where the offer or sale is not permitted. The information contained in this prospectus is accurate only as of the date of this prospectus, regardless of the time of delivery of this prospectus or any sale of the shares of our Class A common stock. Our business, financial condition, results of operations, and prospects may have changed since that date.

For investors outside the United States: Neither we, the selling stockholders, nor the underwriters have done anything that would permit this offering or the possession or distribution of this prospectus or any free writing prospectus in connection with this offering in any jurisdiction where action for that purpose is required, other than in the United States. Persons outside the United States who come into possession of this prospectus must inform themselves about, and observe any restrictions relating to, the offering of the shares of our Class A common stock and the distribution of this prospectus outside the United States. See the section titled “Underwriting” for additional information.

GLOSSARY OF CERTAIN TERMS

The following are abbreviations, acronyms, and definitions of certain terms used in this prospectus:

•“8-way GPU system” means a single, salable system within which eight GPUs are connected. Examples of an 8-way GPU system include the NVIDIA DGX and HGX system configurations.

•“AI” stands for artificial intelligence. AI includes GenAI and other artificial intelligence tools, systems, products and related technologies.

•“API” stands for Application Programming Interface. An API is a set of rules, protocols, and tools that allow different software applications to communicate and interact with each other.

•“Chassis” means the metal frame that supports and houses the components of an electronic device, including the circuits that connect the components.

•“CPU” stands for Central Processing Unit. A CPU is the brain of a computer, responsible for executing instructions and carrying out computations. It is a complex IC that fetches, decodes, and executes instructions, typically from main memory under the control of software programs.

•“CSP” stands for Cloud Service Provider, a company that offers various computing services, such as storage, processing power, and applications, over the internet.

•“CUDA” stands for Compute Unified Device Architecture. It is a parallel computing platform and programming model developed by NVIDIA, enabling developers to program NVIDIA GPUs directly at a low level of abstraction.

•“Customers” refers to our end customers. When the context requires, we may use “end customers,” which include leading enterprises, Sovereign AI initiatives, cloud service providers, government agencies, and research institutions. When used in our audited consolidated financial statements included elsewhere in this prospectus, “customers” means parties we directly invoice for products or services.

•“Ethernet” refers to a standardized network protocol used in data centers and other environments to connect servers, switches, and other devices in a Local Area Network (LAN). Ethernet standards also include features for security and network diagnostics.

•“Fine-tuning” is a process in machine learning where a pre-trained model is further trained on a smaller, task-specific dataset to improve its performance on a specific task. Fine-tuning leverages the knowledge already learned by the model during the initial training phase, allowing it to adapt quickly and effectively to new but related tasks.

•“FLOPS” stands for Floating Point Operations Per Second. It is a measure of a computer’s performance, specifically its ability to perform floating-point arithmetic operations (such as addition, subtraction, multiplication, and division) within one second.

•“GenAI” stands for generative AI. GenAI is a type of AI technology that can produce various types of content, including text, imagery, audio, and synthetic data.

•“GPU” stands for Graphic Processing Unit. GPU is a specialized IC with a high degree of parallelism used to accelerate the rendering of complex graphics onto a screen. Due to their ability to perform numerous computations simultaneously, GPUs outperform CPUs on certain tasks and are increasingly used for scientific computing and accelerating AI workloads, such as the training of large language models.

iii

•“HBM” stands for High Bandwidth Memory. HBM is a type of computer memory designed to provide high bandwidth and low latency for GPUs, other AI accelerators, and CPUs. HBM is significantly faster and more expensive than traditional DRAM memory and is typically integrated within the IC package.

•“HPC” stands for high performance computing. HPC involves using powerful computing systems to perform complex tasks quickly and efficiently, often through parallel processing. HPC is used in applications such as scientific research, weather forecasting, and financial modeling.

•“HuggingFace” is a machine learning and data science platform that enables users to build, train and deploy ML models using open-source code and technologies. It provides tools and libraries for developing ML models and integrating them into live applications. It is a popular and leading open-source repository of AI models and pretrained model checkpoints.

•“Hyperscalers” means large technology companies that offer highly scalable cloud computing services, utilizing extensive data centers. They offer a wide range of dynamically-provisioned services, including computing infrastructure, software platforms, and, increasingly, AI model training and inferencing. These services are available on an as-needed basis, managed and scaled via software by the users.

•“IC” stands for an Integrated Circuit. IC is a miniaturized electronic circuit that combines multiple transistor components and other elements into a single small package. ICs are the fundamental building blocks of modern electronics, and they are used in a wide variety of applications, including computers, servers, networking equipment, smartphones, automobiles, and medical devices.

•“Inference” means the process of using a trained machine learning model to make predictions or decisions based on new data. It involves applying the patterns and knowledge the model learned during training to analyze and interpret new, unseen inputs.

•“I/O” stands for input/output.

•“IT” stands for information technology.

•“LLM” stands for Large Language Model. LLMs are a class of artificial intelligence models that are trained on vast amounts of text data to understand, interpret, and generate human-like language.

•“ML” stands for Machine Learning, a subset of artificial intelligence that involves developing algorithms and statistical models that enable computers to perform specific tasks without explicit programming.

•“MW” stands for megawatt.

•“Node” has multiple meanings:

◦In the context of AI models, a node in a neural network typically means a computational unit within the network graph. Neural networks consist of interconnected layers of nodes, each representing a mathematical operation that transforms input data into meaningful output.

◦In the context of chip manufacturing, “node” is often used as shorthand for “process node,” which refers to specific semiconductor manufacturing processes corresponding to different circuit generations and architectures, for example, 14nm and 5nm nodes.

◦In the context of a supercomputer or computer cluster, a “node” refers to one individual server or system within the cluster.

•“PCIe” stands for Peripheral Component Interconnect Express. PCIe is a high-speed serial data transfer technology widely used in servers. The PCIe protocol is commonly used in data centers for interconnects between CPUs, GPUs, other AI accelerators, storage devices, and Network Interface Cards.

•“Rack” means an open-frame cabinet of standard dimensions used to organize and house servers, networking equipment, power supplies, and other IT hardware. A data center typically houses thousands of racks interconnected by networking switches typically using Ethernet protocol.

•“Sovereign AI” refers to AI systems that are developed, controlled, and managed by a particular nation or established in furtherance of such nation’s public interests.

•“SRAM” stands for Static Random-Access Memory. SRAM is a type of memory that stores data within transistors so long as power is being supplied. Compared to DRAM (Dynamic Random-Access Memory), another common type of RAM used in computers, SRAM is faster and consumes less power during active use. However, it is more expensive and takes up more space than DRAM due to its complex architecture. SRAM is often used on-chip in processors for cache memory because of its speed and efficiency, providing quick access to frequently used data.

•“Tape-out” is the final phase of the chip design process for integrated circuits, where the completed design is released to manufacturing.

•“Time-to-solution” refers to the time period from creation of a customer problem statement to delivery of a model solution with high performance on evaluation metrics. Time-to-solution generally includes the time to train a model, as well as data preprocessing, experiment setup, hyperparameter tuning and evaluation runs, and any required adjustments based on results.

•“Training” refers to the process of teaching an artificial intelligence model to make accurate predictions or decisions by feeding it large amounts of data and adjusting its internal parameters based on identified patterns. During training, the AI model uses algorithms to learn from the input data, iteratively refining its accuracy by adapting its behavior to minimize errors.

•“TSMC” stands for Taiwan Semiconductor Manufacturing Company Limited.

•“Wafer” means a thin slice of a semiconductor material, typically made of silicon, upon which integrated circuits are fabricated. Wafers serve as the foundation for the production of electronic components, including microchips and microprocessors.

•“x86” refers to a family of instruction set architectures based on the Intel 8086 CPU and its successors. The x86 architecture is widely used in personal computers and servers due to its compatibility, performance, and extensive software ecosystem.

PROSPECTUS SUMMARY

This summary highlights selected information contained elsewhere in this prospectus. This summary does not contain all of the information that you should consider before deciding to invest in our Class A common stock. You should read this entire prospectus carefully, including the sections titled “Risk Factors,” “Special Note Regarding Forward-Looking Statements,” “Management’s Discussion and Analysis of Financial Condition and Results of Operations,” and “Business,” and our consolidated financial statements and related notes included elsewhere in this prospectus before making an investment decision.

Our Mission

We believe AI is the most transformative technology of our generation. Our mission is to accelerate AI by making it faster, easier to use, and more energy efficient, making AI accessible around the world.

Company Overview

Cerebras is an AI company. We design processors for AI training and inference. We build AI systems to power, cool, and feed the processors data. We develop software to link these systems together into industry-leading supercomputers that are simple to use, even for the most complicated AI work, using familiar ML frameworks like PyTorch. Customers use our supercomputers to train industry-leading models. We use these supercomputers to run inference at speeds unobtainable on alternative commercial technologies. We deliver these AI capabilities to our customers on premise and via the cloud.

AI compute is comprised of training and inference. For training, many of our customers have achieved over 10 times faster training time-to-solution compared to leading 8-way GPU systems of the same generation and have produced their own state-of-the-art models. For inference, we deliver over 10 times faster output generation speeds than GPU-based solutions from top CSPs, as benchmarked on leading open-source models. This enables real-time interactivity for AI applications and the development of smarter, more capable AI agents. The Cerebras solution requires less infrastructure, is simpler to use, and consumes less power than leading GPU architectures. It enables faster development and eliminates the complex distributed compute work required when using thousands of GPUs. Cerebras democratizes AI, enabling organizations that have less in-house AI or distributed computing expertise to leverage the full potential of AI.

The rise of AI presents a unique set of compute challenges. Unlike other computational workloads, both training and inference require a huge number of relatively simple calculations, the results of which necessitate constant movement to and from memory, and to and from millions or tens of millions of compute cores. This traditionally demands hundreds or thousands of chips, and puts tremendous pressure on memory, memory bandwidth, and the communication fabric linking them all together.

Cerebras started with a simple question: How can we design a processor, purpose-built to meet these exact challenges? If we were to start with a clean sheet, how would we avoid carrying forward the tradeoffs made for graphics and other workloads, and ensure that every transistor is optimized for the specific challenges presented by AI?

Our answer is wafer-scale integration. Cerebras solved a problem that was open for the entire 75-year history of the computer industry: building chips the size of full silicon wafers. The third-generation Cerebras Wafer-Scale Engine (the “WSE-3”) is the largest chip ever sold. It is 57 times larger than the leading commercially available GPU. It has 52 times more compute cores, 880 times more on-chip memory (44 gigabytes), and 7,000 times more memory bandwidth (21 petabytes per second). The sheer size of the wafer-scale chip allows us to keep more work on-silicon and minimize the time-consuming, power-hungry movement of data. This enables Cerebras customers to solve problems in less time and using less power. Our AI compute platform combines processors, systems, software, and AI expert services, to deliver massive acceleration on even the largest, most capable AI models. It substantially reduces training times and inference latencies, while reducing programming complexity.

Our business model is designed to meet the needs of our customers. Organizations seeking control over their data and AI compute infrastructure can purchase Cerebras AI supercomputers for on-premise deployment. Those that want the flexibility of a cloud-based platform can purchase Cerebras high-performance AI compute via a consumption-based model through the Cerebras Cloud, or via our partner’s cloud. We offer customers the flexibility to choose the solution that best aligns with their budgetary, security, and scalability requirements, and some customers choose to use both options simultaneously.

We have established a growing set of customer engagements spanning CSPs, leading enterprises, Sovereign AI programs, national laboratories, research institutions, and other innovators at the forefront of AI. While a substantial portion of our current business is supported by one primary customer, we are actively seeking to expand our reach and diversify our customer base. We collaborate with our customers to harness the power of AI to tackle their most significant challenges and drive breakthroughs across industries.

Bloomberg Intelligence estimates that the AI market will grow to $1.3 trillion by 2032. Consumer and enterprise models like Google’s Gemini, Meta’s Llama, and OpenAI’s ChatGPT have driven demand for AI infrastructure training and inference solutions, powering AI applications such as specialized assistants, agents, and services. We believe that our AI compute platform addresses a large and growing AI hardware and software opportunity across training and inference, as well as software and expert services. We believe that further adoption of AI, accelerated by the advent of GenAI, and the widespread integration of AI into business processes, will rapidly expand our total addressable market (“TAM”) from an estimated $131 billion in 2024 to $453 billion by 2027, a compounded annual growth rate (“CAGR”) of 51%.

We have experienced rapid growth, with revenue of $78.7 million and $24.6 million for the years ended December 31, 2023 and 2022, respectively, representing year-over-year growth of 220%. During the six months ended June 30, 2024 and 2023, we generated $136.4 million and $8.7 million in revenue, respectively. Since our inception, we have incurred operating losses and negative cash flows to develop, market, and expand our product portfolio and to continue our research and development activities. Our net loss for the years ended December 31, 2023 and 2022 was $127.2 million and $177.7 million, respectively, representing a year-over-year reduction of 28%. Our net loss for the six months ended June 30, 2024 and 2023 was $66.6 million and $77.8 million, respectively, representing a year-over-year reduction of 14%.

Industry Background

Over the past 40 years, the computer industry has followed a clear pattern: as major new computational workloads with distinct characteristics emerged, so too have new compute architectures. With each new compute paradigm, technologists first attempted to adapt existing compute architectures to these workloads. But in each case, a new purpose-built architecture was ultimately needed to unlock the potential of the new paradigm.

We believe this pattern is continuing with the rise of AI – the next major compute workload. This growth has been accelerated by the emergence of GenAI, a new class of powerful AI models that can create new content and reason across broad domains and multiple data types. These breakthrough capabilities translate to tremendous potential value creation and have driven rapid GenAI adoption.

The Computational Demands of Training and Inference

Both training and inference demand immense compute, each with unique compute, memory, and memory bandwidth requirements. They represent two stages in the continuous lifecycle of AI models. Once trained, a model is “served” in production and used for inference. As part of this cycle, models in production are continuously being optimized to use fewer compute resources, and while that is happening, new and more powerful models are being trained, leading to the eventual obsolescence of the previous model—starting the cycle over.

trillions of tokens. As the industry has pushed to achieve greater AI capabilities, the size of GenAI models has grown, and we expect this trend to continue.

The increase in model and data sizes has led to a dramatic surge in computational demand. Today, training for GenAI requires enormous GPU clusters, sophisticated engineering teams, and months of time for a single run. A training run can cost over a hundred million dollars, and improving the model and keeping it fresh with new data requires additional fine-tuning and regular re-training.

For inference, the required compute is a function of model size, user demand, and time spent on inference. Larger models use more computational resources, and each user request also increases the compute need, contributing to significant and ongoing operational costs. We expect the demand for inference to grow, especially as larger and more capable models become more widely adopted. There is currently a direct tradeoff between model capability and responsiveness of user experience because the largest and most capable models demand more inference compute and therefore run more slowly during inference. We believe that as both inference speed and reasoning capabilities advance, GenAI models will support more demanding applications, thereby expanding the market opportunity for inference.

Existing Compute Architectures Are Fundamentally Limited for GenAI Training and Inference

GPUs, though better than CPUs for AI workloads, face fundamental limitations when processing the unique characteristics of large GenAI models. GenAI models are complex, interconnected compute graphs that require the constant communication of intermediate calculations to train. This requires a high amount of data movement to and from memory and across cores. Similarly, during inference, these models generate outputs sequentially – each dependent on the previous output – requiring the full model to be constantly moved in and out of memory to produce successive outputs, and again requiring massive data movement between cores and memory. GPUs face inherent scalability and complexity challenges when faced with the distinct, communication-heavy requirements of GenAI workloads.

For Training – Individual GPUs Are Too Small, and Scaling to Many GPUs is Highly Inefficient

Large GenAI models far exceed the memory and processing limits of a single GPU. For example, to train GPT-3, it would take a single NVIDIA H100 more than eight years of running at peak theoretical performance to train the model. Recent models like GPT-4 and Gemini are over 10 times larger in parameter size than GPT-3. Consequently, training a large GenAI model on GPUs in a tractable amount of time requires breaking up the model and calculations, and distributing the pieces across hundreds or thousands of GPUs, creating extreme communication bottlenecks and power inefficiencies.

This distributed compute problem also creates a high level of complexity for developers, who are responsible for partitioning and coordinating the compute, memory, and communication across GPUs, so that they can work together in a complex choreography. This is an ongoing cost and slows down time-to-solution, as the delicate balance of bottlenecks needs to be reconfigured every time the ML developer wants to change the model architecture, model size, or run on a different number of GPUs.

For Inference – GPU Efficiency is Low and Limited by Memory Bandwidth

During generative inference, the full model must be run for each word that is generated. Since large models exceed on-chip GPU memory, this requires frequent data movement to and from off-chip memory. GPUs have relatively low memory bandwidth, meaning that the rate at which they can move information from off-chip HBM to on-chip SRAM, where the computation is done, is severely limited. This leads to low performance as GPUs cores are idle while waiting for data – they can run at less than 5% utilization on interactive generative inference tasks. Low utilization and limited memory bandwidth impact the responsiveness and throughput of GPU-based systems and hinders real-time applications for larger models. This inefficiency also necessitates larger GPU deployments and dramatically drives up the cost of inference.

GPU companies have tried to address these challenges, but the issues of small core count, memory size, and memory bandwidth are fundamental hardware limitations that persist. Interconnect technologies like InfiniBand, PCIe, and NVLink are limited by their physical interfaces, and moving data across them is thousands of times slower and more power-hungry than keeping the data on silicon. Software libraries intended to simplify distributed computing still require developers to manage complex parallelism strategies and extensive codebases. Realizing the physical challenges of repurposing small chips for a large compute problem, the GPU industry has recently announced new multi-chip packaging techniques, but these also yield only minimal expansions of GPU chip size.

Accelerating GenAI requires a dedicated compute solution, designed for the unique requirements of GenAI, that can deliver faster training times, real-time inference speeds, and simple developer workflows, at reasonable cost.

Our Solution

•Cerebras Wafer-Scale Engine (WSE). Our third generation WSE, the WSE-3, is 57 times larger than the leading commercially available GPU and has 52 times more compute cores, totaling 900,000. It features 880 times more on-chip memory (44 gigabytes SRAM) and 7,000 times more memory bandwidth (21 petabytes per second) than the leading commercially available GPU. The immense scale of the WSE delivers significant acceleration, efficiency, and simplicity for AI training and inference.

For Training. Each Cerebras WSE has enough compute and on-chip memory to run even the largest, multi-trillion parameter GenAI models on a single chip, thus avoiding the complexities of chip-to-chip data movement. To further speed up training time-to-solution, Cerebras users can simply add more WSEs to the problem, scaling to near-linear performance.

•Cerebras System (CS). The CS AI computer system houses the WSE and delivers innovative power and cooling to the chip. Our third generation CS (“CS-3”) delivers three times more compute per unit power than the leading 8-way GPU system. This compact AI powerhouse is designed to easily integrate into standard data centers, and connects into the network via standards-based 100G Ethernet.

•Cerebras AI Supercomputer. The Cerebras AI Supercomputer is designed to streamline scaling up to 2,048 CS-3 systems for maximum AI acceleration, with more efficiency and simplicity than scaling up to large GPU clusters. Our AI Supercomputer delivers near-linear performance increase as CS systems are added to a problem, takes only seconds to configure, and does not incur the overhead or complexity of heavy inter-chip, inter-system communication.

•Cerebras Software Platform (CSoft). Our proprietary software platform, CSoft, is core to our solution and provides intuitive usability and improved developer productivity.

CSoft seamlessly integrates with industry-standard ML frameworks like PyTorch and with popular software development tools, allowing developers to easily migrate to the Cerebras platform.

CSoft eliminates the need for low-level programming in CUDA, or other hardware-specific languages. Starting from a user’s PyTorch model, the CSoft graph compiler automatically maps model operations to the WSE, creating an optimized executable without user-level intervention.

CSoft allows ML users to accelerate training and inference on models of any size, scaled across any configuration of the Cerebras AI Supercomputer, just by changing one number in a configuration file, simulating a single-device programming experience without the complexities of distributed programming. This drastically reduces operational overhead and speeds up developer iteration time and business impact.

•Cerebras Inference Serving Stack. The Cerebras Inference Stack is built on top of CSoft and is designed to allow customers to easily deploy even the largest GenAI models at industry-leading inference speeds. The Cerebras Inference API is designed to facilitate rapid developer adoption and ease of use. Our serving software automatically handles system-level optimizations for our inference solution and is designed to enable low latency and high cost effectiveness.

•AI Model Services. Our AI model services further amplify speed to solution. Our team of AI practitioners helps customers design research experiments, train models, and optimize processes designed to achieve the fastest time-to-solution. These services complement our advanced hardware and software platform, providing an end-to-end solution for rapid and efficient AI development and deployment that helps our customers translate AI potential to business impact.

Summary of key customer benefits include:

•Enables over 10 times faster training time-to-solution compared to leading 8-way GPU systems of the same generation, as reported by many of our customers. This dramatically accelerates AI model time-to-solution, enabling businesses to test ideas faster, iterate more quickly, and bring next-generation GenAI-powered products and services to market, faster and more cost effectively.

•Delivers over 10 times faster GenAI inference compared to GPU-based solutions from top CSPs, as benchmarked on leading open-source models. The WSE’s massive on-chip SRAM capacity keeps the vast majority of memory-to-compute communication on-silicon, thereby avoiding the memory bandwidth bottleneck faced by GPU-based solutions. Our resulting ultra-low latency delivers industry-leading inference speeds and real-time responses back to users, even on large, cutting-edge GenAI models. Ten times more speed compared to GPU-based solutions from top CSPs also allows developers to make ten times more inference calls in the same amount of time. This supports a new level of model capability delivered by techniques like multi-step inference and agentic AI flows, which leverage more inference calls to produce higher reasoning capability for more complex tasks in domains such as coding, math, and science applications.

•End-to-end solution. Cerebras offers a unified platform purpose-built to accelerate fundamental compute characteristics of both AI training and inference. Cerebras excels across the axes of compute, memory, memory bandwidth, and simplicity of use, made possible by wafer-scale integration. This allows customers to swiftly transition from training to fine-tuning to deploying high-quality GenAI models on the same platform, eliminating the need for investing in and maintaining separate hardware infrastructure.

•Zero distribution complexity. Users can effortlessly run a GenAI model of any size, and it takes no additional code to achieve automatic near-linear performance scaling across the nodes of a Cerebras Supercomputer.

•Low migration cost. Our proprietary CSoft platform integrates seamlessly with familiar ML frameworks and tools, like PyTorch, eliminating the need for AI teams to learn new languages or adapt to new development environments and easing the transition from other hardware platforms.

•Power and operational efficiency. Cerebras outperforms GPU systems in power efficiency due to both hardware and architectural advantages. Wafer-scale integration allows the majority of AI workload communication to remain on-chip, significantly reducing data movement distance and power consumption; moving a bit of data on the WSE-3 takes less than 1% of the energy needed to do the same over off-chip GPU interconnects. This drives significant operational cost savings and streamlined management for organizations deploying AI at scale.

•Expert-led model training and AI integration services. We offer expert-led foundation model training, fine-tuning, and retrieval-augmented generation services to customers. Our team provides guidance on cutting-edge AI methods that work on top of our hardware, assisting customers to derive the maximum value from their AI investment.

Our Customers

We have an expanding customer base that includes leading enterprises, Sovereign AI initiatives, cloud service providers, government agencies, and research institutions at the forefront of AI and at the intersection of AI and HPC. These customers leverage Cerebras to tackle complex AI challenges and achieve previously unattainable business outcomes, even with the most advanced GPU systems. Our customers find fundamental value in the simplicity of the Cerebras solution. It unlocks breakthrough business and scientific use cases by removing limitations in development time, programming complexity, and runtime speed.

Our Business Model

We use a combination of direct sales and partnerships to address the rapidly expanding AI market. We offer both on-premise solutions and cloud-based solutions to provide maximum flexibility to our customers. We offer a collection of services from data center deployment to AI expert services and AI Supercomputer operation and management, to provide our customers with the support they need to train, deploy, and accelerate GenAI time-to-value.

•On Premise. We sell our AI Supercomputers to leading organizations who seek maximum control over their data and their AI infrastructure, fulfilling their needs for high-performance AI compute on premise. Our on-premise AI Supercomputers support both training and inference. They can be configured either for both workloads, or to be further optimized for only one, depending on our customers’ needs.

•Cloud-Based Computing Services. We sell Cerebras solutions via our cloud offering as well as via the Condor Galaxy Cloud owned by Group 42 Holding Ltd (together with its affiliates, “G42”), our partner CSP. Our cloud solutions provide customers fast and flexible access to our powerful AI acceleration hardware. This offering gives our customers the ability to train LLMs with extraordinary speed and deploy them for inference at ultra-low latencies, all without the complexity or time needed to build and manage on-premise infrastructure.

•Cerebras Inference Cloud. Our real-time inference solution is also available via a dedicated inference cloud service. Leveraging our Cerebras Inference Serving Stack, this cloud API offering allows developers to directly point their applications towards efficient and reliable model serving endpoints. On Cerebras Inference Cloud, we host both popular open-source models and proprietary customer models. For customers who do not need direct compute access and are not interested in managing their own inference serving software stack, our inference cloud offering is the quickest and simplest way to leverage our fast model inference services.

We provide a combination of these offerings to customers who may benefit from leveraging both on-premise and cloud-based options. This flexibility allows customers to choose the solution that best aligns with their budgetary, security, and scalability requirements.

Additionally, customers can train models on-premise and then leverage our inference cloud for production, benefiting from flexible serving resources that can adapt to fluctuating demand. This end-to-end solution allows seamless integration from training to production, serving the entire lifecycle of a customer’s AI needs.

We also provide professional services to assist customers throughout the AI workflow.

Our Market Opportunity

We participate in a large and growing AI market. Our full suite of AI computing solutions addresses use cases for training, inference, software, and expert services. We estimate the TAM for our AI computing platform to be approximately $131 billion in 2024, growing to $453 billion in 2027, a 51% CAGR. This TAM is comprised of the following core markets:

•AI Compute for Training. As businesses continue to evaluate and deploy solutions, we believe the market for training new models will continue to grow. Based on market estimates in Bloomberg Intelligence research, our estimate of the TAM for AI Training Infrastructure is $72 billion in 2024, growing to $192 billion in 2027, a 39% CAGR.

•AI Compute for Inference. While GenAI training is essential to developing models that are powerful and accurate, we believe inference is the next phase of the ongoing wave of AI disruption. As more enterprises develop models and start to deploy their models in applications at scale, the need for high performance and efficient inference is becoming critical to fully realize the commercial potential of ML. We believe that the inference opportunity is enormous, as the market is in the early phases in its adoption cycle. Our estimate of the TAM for AI Inference is $43 billion in 2024, growing at an estimated 63% CAGR to $186 billion in 2027.

•Software and Expert Services. Based on market estimates in Bloomberg Intelligence research, our estimate of the TAM for GenAI Software and Services is $16 billion in 2024, growing to $75 billion in 2027, a 67% CAGR.

We believe we are at the very early stages of a large and fast-growing market opportunity. As adoption of AI continues to accelerate, we expect numerous new applications will be identified, and we believe our solutions are well-positioned to capitalize on the wave of disruption that will come in the coming years.

Competitive Strengths

We believe our design is capable of meeting today’s needs and is scalable to address tomorrow’s challenges. Our competitive strengths include:

•The world’s first and only wafer-scale chip in the market. Our wafer-scale chip architecture eliminates the need for distributed computing. This enables AI developers to use up to 97% less code when working with large models on our platform compared to on clusters of GPUs and greatly accelerates the speed of AI model development for larger-scale models.

•Full system solution that is easy to deploy and efficient to operate. Our system is co-designed with the wafer-scale processor, leveraging proprietary technology to address thermal and power delivery challenges. This allows us to keep the system operating efficiently, optimizing the energy consumption and underlying operating costs for our customers.

•Comprehensive software suite that leads to ease of adoption and shortens time to deployment. The Cerebras Software platform allows for seamless programmability of AI models through our integration with PyTorch. Our tools allow users to bring models that have been trained on other hardware onto our platform for training or for inference. Likewise, users can train models on our platform and deploy those models for inference elsewhere.

•Our AI platform addresses both training and inferencing markets. For training, our AI platform enables customers to effortlessly and swiftly use the most advanced GenAI models available on the market, without the need for specialized software frameworks or help from distributed computing experts. For inferencing, our AI platform delivers ultra-low latencies and high generation throughput. This helps our customers to unlock cutting edge performance for ultra-low-latency use cases leveraging GenAI.

•Scalable architecture. We have developed our solutions to support models and data sets of large and varying sizes. Our current generation CS-3 is designed to support models with up to 24 trillion parameters, much larger than even today’s state-of-the-art GenAI models. Our platform is designed to seamlessly scale from 1 to 2,048 CS-3 systems, forming an AI supercomputer that is even further differentiated by our proprietary interconnect and memory technology. This enables customers to seamlessly increase compute resources from small-scale experiments to large-scale deployments.

•AI model services help customers translate AI potential to business impact. We provide customers with AI model services to help them develop solutions that are customized to meet their needs and help them realize the full value of their AI investments. These services include model selection, data preparation, training, and solutions integration.

•World-class AI talent with a proven track record of innovation and execution. We have assembled a world-class team of industry leaders in integrated circuit design, processor architecture, power delivery, cooling, system engineering and software. Over the last five years, we have introduced three generations of our WSE, each time achieving two times the performance of its predecessor, and bringing new IC, power, and cooling technologies to bear.

Growth Strategies

We believe we are positioned for sustained growth in the rapidly expanding market for AI acceleration solutions. We have designed our focused strategies to drive continued success and establish ourselves as a long-term leader:

•Increase sales to our existing customers. We have established a strong land-and-expand track record with our existing customer base and we intend to deepen these relationships by expanding our product and service offerings tailored to their evolving needs. Our strategy focuses on demonstrating the value proposition of our solutions through initial engagements, cross-selling complementary products, and identifying new use cases within existing customers.

•Expand our customer base. We plan to aggressively pursue opportunities in relevant sectors such as healthcare, pharmaceutical, biotechnology, government, financial services, sovereign, and energy, to name a few, where our AI acceleration capabilities can address critical computational bottlenecks. We will seek to drive this expansion by focused sales and marketing initiatives, highlighting the transformative potential of our technology with targeted use cases. We intend to leverage our existing success stories and strategic partnerships to both bolster our credibility within new markets and establish key channels for customer acquisition.

•Further penetrate into the rapidly growing inference market. We recently launched our inference solution. The immense amount of memory bandwidth and capacity on our chip allows us to deliver significantly lower generation latency and higher generation throughput over GPUs. By making API-based

inferencing available through our cloud offering, we could significantly accelerate our adoption into inferencing use cases.

•Benefit from opportunities in large adjacent AI and compute-intensive markets. We are actively enabling applications in fields like life sciences, materials science, and financial modeling, where our cutting-edge AI computing solutions can unlock new discoveries and solve complex problems. Geographically, our strategy includes deepening our investment in partnerships with large Sovereign AI initiatives and new markets, to accelerate AI adoption.

•Accelerate our existing product roadmap as well as develop new products for emerging use cases. As AI infrastructure requirements scale, we expect emerging use cases to require new products with added functionalities to solve data, networking, and memory bottlenecks. With our continued focus on innovation, we intend to develop and introduce new products and form factors that will enable us to service a larger portion of our market opportunity.

•Advance product adoption by proliferating cloud deployment of our AI solutions. We intend to accelerate our growth by expanding access to our revolutionary AI systems through cloud deployment, which we believe will significantly broaden our customer base.

Risk Factors Summary

Our business is subject to a number of risks and uncertainties of which you should be aware before making a decision to invest in our Class A common stock. These risks are more fully described in the section titled “Risk Factors.” These risks include, among others, the following:

•We may not sustain our growth rate, and we may not be able to manage future growth effectively.

•We have a history of generating net losses, and if we are unable to achieve adequate revenue growth while our expenses increase, we may not achieve and maintain profitability in the future.

•We have a limited operating history, and we may have difficulty accurately predicting our future revenue for the purpose of appropriately budgeting and adjusting our expenses.

•We currently generate a significant majority of our revenue from one customer, G42, and a significant portion of our revenue from a limited number of customers. A reduction in demand from, or a material adverse development in our relationship with, G42 or any of our other significant customers may harm our business, financial condition, results of operations, and prospects.

•Our business and our products and services are subject to various governmental regulations, and compliance with these regulations may cause us to incur significant expense. Similarly, we are required to obtain export licensing to sell our products to various jurisdictions where we have customers, and we cannot guarantee that we will be successful in obtaining all required licenses in the future. If we fail to comply with applicable regulations, we could be subject to civil or criminal penalties, and if we are unable to obtain licenses to export our products, our business, financial condition, results of operations, and prospects may be harmed.

•If the market does not adopt our products, we will be unable to grow our business.

•If we are unable to expand the application of our products, or if the new products we develop and introduce into the market are not successful, our business, financial condition, results of operations, and prospects may be harmed.

•The market for AI computing solutions is competitive and evolving, and if we do not compete effectively, our business, financial condition, results of operations, and prospects may be harmed.

•We depend on third-party suppliers, and substantially all of our manufacturing services and components are procured on a purchase order basis without capacity commitments, which may harm our ability to bring products to market and our reputation, business, financial condition, results of operations, and prospects.

•Our supply chain is long, complex, and global, with many interdependencies. Any significant fluctuations of supply and demand or disruption to our supply chain may harm our ability to manufacture and deliver our products to our customers.

•No public market for our Class A common stock currently exists and an active liquid market may not develop or be sustained following this offering.

•We identified material weaknesses in our internal control over financial reporting. If we are unable to remediate these material weaknesses, or if we identify additional material weaknesses in the future or otherwise fail to maintain an effective system of internal controls, we may not be able to accurately or timely report our financial condition or results of operations, which may adversely affect investor confidence in us and, as a result, the value of our Class A common stock.

Corporate Information

We were incorporated in April 2016 as a Delaware corporation. Our principal executive offices are located at 1237 E. Arques Avenue, Sunnyvale, California 94085, and our telephone number is (650) 933-4980. Our website address is www.cerebras.ai. Information contained on, or that can be accessed through, our website does not constitute part of this prospectus, and the inclusion of our website address in this prospectus is an inactive textual reference only.

Implications of Being an Emerging Growth Company

We are an emerging growth company as defined in the Jumpstart Our Business Startups Act of 2012 (the “JOBS Act”). We will remain an emerging growth company until the earliest of: (i) the last day of the fiscal year following the fifth anniversary of the completion of this offering; (ii) the last day of the fiscal year in which we have total annual gross revenue of at least $1.235 billion; (iii) the last day of the fiscal year in which we are deemed to be a “large accelerated filer” as defined in Rule 12b-2 under the Securities Exchange Act of 1934, as amended (the “Exchange Act”), which would occur if the market value of our Class A common stock held by non-affiliates exceeded $700.0 million as of the last business day of the second fiscal quarter of such year; or (iv) the date on which we have issued more than $1.0 billion in non-convertible debt securities during the prior three-year period. An emerging growth company may take advantage of specified reduced reporting requirements and is relieved of certain other significant requirements that are otherwise generally applicable to public companies. As an emerging growth company:

•we will present in this prospectus only two years of audited annual financial statements, plus any required unaudited condensed consolidated financial statements, and related management’s discussion and analysis of financial condition and results of operations;

•we will avail ourselves of the exemption from the requirement to obtain an attestation and report from our independent registered public accounting firm on the assessment of our internal control over financial reporting pursuant to the Sarbanes-Oxley Act of 2002;

•we will provide less extensive disclosure about our executive compensation arrangements; and

•we will not require stockholder non-binding advisory votes on executive compensation or golden parachute arrangements.

In addition, the JOBS Act provides that an emerging growth company can take advantage of an extended transition period for complying with new or revised accounting standards. This provision allows an emerging growth

company to delay the adoption of some accounting standards until those standards would otherwise apply to private companies. We have elected to use the extended transition period for any other new or revised accounting standards until the date that we are no longer an emerging growth company or affirmatively and irrevocably opt out of the extended transition period. As a result, our financial statements may not be comparable to companies that comply with new or revised accounting pronouncements as of public company effective dates.

The Offering


Class A common stock offered by us						shares.

Class A common stock offered by the selling stockholders						shares.

Option to purchase additional shares of Class A common stock from us						shares.

Class A common stock to be outstanding immediately after this offering						shares (or shares if the underwriters exercise their option to purchase additional shares of Class A common stock from us in full).

Class N common stock to be outstanding immediately after this offering						None.

Total Class A common stock and Class N common stock to be outstanding after this offering						shares (or shares if the underwriters exercise their option to purchase additional shares of Class A common stock from us in full).

Use of proceeds						We estimate that we will receive net proceeds from this offering of approximately $ (or $ if the underwriters exercise their option to purchase additional shares of Class A common stock in full), based upon the assumed initial public offering price of $ per share of Class A common stock, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. The principal purposes of this offering are to obtain additional capital to fund our operations, create a public market for our Class A common stock, facilitate our future access to the public equity markets, and increase awareness of our company among potential partners. We currently intend to use the net proceeds from this offering, together with our existing cash, cash equivalents, and investments, for general corporate purposes, including working capital, operating expenses, and capital expenditures. We may also use a portion of the net proceeds to in-license, acquire, or invest in complementary technologies, assets, businesses, or intellectual property. We periodically evaluate strategic opportunities; however, we have no current commitments to enter into any such acquisitions or make any such investments. We intend to use approximately $ of the net proceeds to satisfy tax withholding and remittance obligations related to the RSU Net Settlement (as defined below) for restricted stock units (“RSUs”) that will vest in connection with this offering. We will have broad discretion in the way that we use the net proceeds of this offering. See the section titled “Use of Proceeds” for additional information. We will not receive any proceeds from the sale of Class A common stock in this offering by the selling stockholders.


Voting rights						We have two classes of common stock: Class A common stock and Class N common stock. Class A common stock is entitled to one vote per share and Class N common stock is non-voting and is convertible into one share of Class A common stock. See the section titled “Description of Capital Stock” for additional information.

Risk factors						See the section titled “Risk Factors” and other information included in this prospectus for a discussion of factors you should carefully consider before deciding whether to invest in our Class A common stock.

Proposed Nasdaq Global Market trading symbol						“CBRS”

In this prospectus, the number of shares of our common stock to be outstanding after this offering is based on shares of our Class A common stock and no shares of our Class N common stock outstanding as of June 30, 2024, after giving effect to the Preferred Stock Conversion, the Option Exercise, and the RSU Net Settlement (each as defined below), and excludes:

• shares of our Class A common stock issuable upon the exercise of outstanding stock options as of June 30, 2024, with a weighted-average exercise price of $ per share, after giving effect to the Option Exercise;

• shares of our Class A common stock issuable upon the exercise of stock options granted after June 30, 2024, with a weighted-average exercise price of $ per share;

• shares of our Class A common stock issuable upon the vesting and settlement of RSUs subject to service-based and liquidity-based vesting conditions outstanding as of June 30, 2024, for which the service-based vesting condition was not yet satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering, after giving effect to the RSU Net Settlement;

• shares of Class A common stock issuable upon the vesting and settlement of RSUs subject to service-based and liquidity-based vesting conditions granted after June 30, 2024, for which the service-based vesting condition was not yet satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering, after giving effect to the RSU Net Settlement;

•22,851,296 shares of our Class N common stock reserved for future purchase pursuant to the G42 Primary Purchase (see the section titled “Certain Relationships and Related Party Transactions” for additional information);

•a variable number of shares of our Class N common stock that may be issued pursuant to the G42 Option (see the sections titled “Dilution—G42 Option” and “Certain Relationships and Related Party Transactions” for additional information);

• shares of our Class A common stock reserved for future issuance under our 2024 Incentive Award Plan (the “2024 Plan”), which will become effective on the business day immediately prior to the date of effectiveness of the registration statement of which this prospectus forms a part, including new shares and the number of shares (i) that remain available for grant of future awards under our 2016 Equity Incentive Plan (as amended, the “2016 Plan”) at the time the 2024 Plan becomes effective, which shares will cease to be available for issuance under the 2016 Plan at such time and (ii) underlying outstanding stock-based compensation awards granted under the 2016 Plan (such awards outstanding under such plans, the “Prior Plan Awards”) that expire, or are cancelled, forfeited, reacquired, or withheld; and

• shares of our Class A common stock reserved for future issuance under our 2024 Employee Stock Purchase Plan (the “ESPP”), which will become effective on the business day immediately prior to the date of effectiveness of the registration statement of which this prospectus forms a part.

The 2024 Plan and the ESPP also provide for automatic annual increases in the number of shares reserved thereunder. See the section titled “Executive and Director Compensation—Equity Compensation Plans” for additional information.

Except as otherwise indicated, all information in this prospectus assumes or gives effect to:

•the adoption, filing, and effectiveness of our amended and restated certificate of incorporation and the adoption of our amended and restated bylaws, each of which will occur immediately prior to the completion of this offering;

•the automatic conversion of all outstanding shares of our redeemable convertible preferred stock into an aggregate of 82,899,159 shares of our Class A common stock, which will occur prior to the completion of this offering, including 68,213 shares issued upon the exercise of a warrant (the “Preferred Stock Conversion”);

•the cash exercise of stock options to purchase shares of our Class A common stock in connection with this offering by certain selling stockholders (the “Option Exercise”), with a weighted-average exercise price of $ per share, for total gross proceeds to us of approximately $ , by certain selling stockholders in connection with the sale of all or a portion of such shares by such selling stockholders in this offering, as described in the section titled “Principal and Selling Stockholders”;

•the net issuance of shares of our Class A common stock issuable upon the vesting and settlement of RSUs subject to service-based and liquidity-based vesting conditions outstanding as of , 2024, for which the service-based vesting condition was satisfied as of , 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering, after giving effect to the withholding of an estimated shares to satisfy estimated tax withholding and remittance obligations (based on an assumed % tax withholding rate) (the “RSU Net Settlement”);

•no repurchase of outstanding shares of our capital stock after June 30, 2024;

•no exercise of outstanding stock options or settlement of outstanding RSUs after June 30, 2024, except for the Option Exercise and the RSU Net Settlement;

•no issuance of shares of our capital stock pursuant to the G42 Primary Purchase or the G42 Option (see the sections titled “Dilution—G42 Option” and “Certain Relationships and Related Party Transactions” for additional information); and

•no exercise by the underwriters of their option to purchase additional shares of our Class A common stock from us in this offering.

Summary Consolidated Financial Data

The following tables set forth our summary consolidated financial data. The summary consolidated statements of operations data for the years ended December 31, 2023 and 2022 have been derived from our audited consolidated financial statements included elsewhere in this prospectus. The summary consolidated statements of operations data for the six months ended June 30, 2024 and 2023 and the summary consolidated balance sheet data as of June 30, 2024 have been derived from our unaudited interim consolidated financial statements included elsewhere in this prospectus. In our opinion, the unaudited interim consolidated financial statements have been prepared on a basis consistent with our audited consolidated financial statements and, in our opinion, contain all adjustments, consisting only of normal and recurring adjustments, necessary for a fair presentation of such interim financial statements. Our historical results are not necessarily indicative of results that may be expected in the future, and our results for the six months ended June 30, 2024 are not necessarily indicative of results that may be expected for the year ending December 31, 2024 or any future period.

You should read the following summary consolidated financial data in conjunction with the section titled “Management’s Discussion and Analysis of Financial Condition and Results of Operations” and our consolidated financial statements and related notes included elsewhere in this prospectus. The summary consolidated financial data in this section are not intended to replace, and are qualified in their entirety by, the consolidated financial statements and related notes.


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands, except per share amounts)
Consolidated Statement of Operations:
Revenue
Hardware	$	57,114	$	15,599	$	104,269	$	1,559
Services and other	21,630		9,020		32,133		7,105
Total revenue	78,744		24,619		136,402		8,664
Cost of sales(1)
Hardware	45,559		19,195		66,442		1,980
Services and other	6,827		2,534		13,941		2,306
Total cost of sales	52,386		21,729		80,383		4,286
Gross profit	26,358		2,890		56,019		4,378
Operating expenses
Research and development(1)	140,057		155,408		77,742		76,295
Sales and marketing(1)	9,642		9,401		7,237		4,176
General and administrative(1)	10,593		16,902		12,851		4,922
Total operating expenses	160,292		181,711		97,830		85,393
Loss from operations	(133,934)		(178,821)		(41,811)		(81,015)
Interest income	5,683		1,076		3,809		2,349
Other income (expense), net	1,228		230		(28,284)		918
Loss before income taxes	(127,023)		(177,515)		(66,286)		(77,748)
Income tax expense	132		204		319		72
Net loss	$	(127,155)	$	(177,719)	$	(66,605)	$	(77,820)
Net loss per share – basic and diluted(2)	$	(2.92)	$	(4.28)	$	(1.42)	$	(1.82)
Weighted average number of common shares outstanding, basic and diluted	43,552		41,485		46,945		42,857


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands, except per share amounts)
Pro forma net loss per share attributable to common stockholders, basic and diluted(3)
Pro forma weighted-average shares used in calculating pro forma net loss per share attributable to common stockholders, basic and diluted(3)
Other Financial Information:
Non-GAAP operating loss(4)	$	(107,303)	$	(155,777)	$	(9,482)	$	(71,755)
Non-GAAP net loss(5)	$	(100,524)	$	(154,675)	$	(3,949)	$	(68,560)
Net cash (used in) provided by operating activities	$	(78,977)	$	(164,402)	$	311,813	$	(70,185)

_______________

(1)Includes stock-based compensation expense as follows:


	Year Ended December 31,				Six Months Ended, June 30,
	2023		2022		2024		2023
	(in thousands)
Cost of sales	$	309	$	223	$	420	$	51
Research and development	21,187		17,732		23,905		7,367
Sales and marketing	3,563		832		3,195		1,075
General and administrative	1,572		4,257		4,809		767
Total stock-based compensation expense	$	26,631	$	23,044	$	32,329	$	9,260

Stock-based compensation expense included $9.0 million, $8.6 million, $18.5 million, and $0.9 million for the years ended December 31, 2023, and 2022 and for the six months ended June 30, 2024 and 2023, respectively, related to secondary transactions in each period and a common stock repurchase from employees during the year ended December 31, 2022. See Note 12 to our audited consolidated financial statements and our unaudited interim condensed consolidated financial statements included elsewhere in this prospectus for additional details on the secondary transactions.

(2)See Note 7 to our audited consolidated financial statements included elsewhere in this prospectus for an explanation of the method used to calculate our basic and diluted net loss per share and the weighted-average number of shares used in the computation of per share amounts.

(3)The pro forma weighted-average shares used in computing pro forma net loss per share gives effect to (i) the Preferred Stock Conversion, (ii) the Option Exercise, and (iii) the RSU Net Settlement. The pro forma net loss used to calculate pro forma net loss per share reflects stock-based compensation expense of approximately $ that we will recognize upon the completion of this offering related to RSUs subject to service-based and liquidity-based vesting conditions for which the service-based vesting condition was satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering.

(4)See “Non-GAAP Operating Loss” below for more information and for a reconciliation of Non-GAAP operating loss to loss from operations, the most directly comparable financial measure calculated and presented in accordance with U.S. generally accepted accounting principles (“U.S. GAAP”).

(5)See “Non-GAAP Net Loss” below for more information and for a reconciliation of Non-GAAP operating loss to net loss, the most directly comparable financial measure calculated and presented in accordance with U.S. GAAP.


	As of June 30, 2024
	Actual		Pro Forma(1)	Pro Forma As Adjusted(2)(3)
	(in thousands)
Consolidated Balance Sheet Data:
Cash and cash equivalents	$	90,931	$	$
Working capital(4)	129,089
Total assets	622,862
Total liabilities	479,394
Redeemable convertible preferred stock	722,780
Stockholders’ deficit	(579,312)

_______________

(1)The pro forma column above gives effect to (i) the filing and effectiveness of our amended and restated certificate of incorporation, which will occur immediately prior to the completion of this offering; (ii) the Preferred Stock Conversion; (iii) the Option Exercise; (iv) the RSU Net Settlement; (v) the increase in accrued expenses and other current liabilities and an equivalent decrease in additional paid-in capital of $ in connection with the estimated tax withholding and remittance obligations related to the RSU Net Settlement; and (vi) stock-based compensation expense of approximately $ that we will recognize upon the completion of this offering related to RSUs subject to service-based and liquidity-based vesting conditions for which the service-based vesting condition was satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering.

(2)The pro forma as adjusted column above gives further effect to (i) the pro forma adjustments set forth above; (ii) the issuance and sale of shares of Class A common stock by us in this offering at an assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us; (iii) the receipt by us of gross proceeds of approximately $ in connection with the Option Exercise; and (iv) the use of a portion of the net proceeds from this offering to satisfy the estimated tax withholding and remittance obligations related to the RSU Net Settlement.

(3)Each $1.00 increase or decrease in the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would increase or decrease, as applicable, each of cash and cash equivalents, working capital, total assets, and stockholders’ deficit by $ , assuming that the number of shares of Class A common stock offered by us, as set forth on the cover page of this prospectus, remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. Similarly, each increase or decrease of 1.0 million shares in the number of shares of Class A common stock offered by us would increase or decrease, as applicable, each of cash and cash equivalents, working capital, total assets, and stockholders’ deficit by $ , assuming the assumed initial public offering price of $ per share remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. In addition, each 1.0% increase or decrease in the assumed tax withholding rate would increase or decrease, as applicable, the amount of estimated tax withholding and remittance obligations related to the RSU Net Settlement by $ . Pro forma adjustments in the footnotes above and the related information in the consolidated balance sheet data are illustrative only and will be adjusted based on the actual initial public offering price and other terms of this offering determined at pricing, the actual tax withholding rate, as well as the actual amount of RSUs settled in connection with this offering (including after accounting for forfeitures prior to the settlement date).

(4)Working capital is defined as total current assets less total current liabilities. See our unaudited interim consolidated financial statements and the related notes thereto included elsewhere in this prospectus for further details regarding our current assets and current liabilities.

Non-GAAP Financial Measures

We use certain non-GAAP financial measures to supplement the performance measures in our consolidated financial statements, which are presented in accordance with GAAP. These non-GAAP financial measures include non-GAAP operating loss and non-GAAP net loss. We use these non-GAAP financial measures for financial and operational decision-making and as a means to assist us in evaluating period-to-period comparisons. By excluding certain items that may not be indicative of our recurring core operating results, we believe that non-GAAP operating loss and non-GAAP net loss provide meaningful supplemental information regarding our performance. Accordingly, we believe these non-GAAP financial measures are useful to investors and others because they allow for additional information with respect to financial measures used by management in its financial and operational decision-making and they may be used by our institutional investors and the analyst community to help them analyze the health of our business. However, there are a number of limitations related to the use of non-GAAP financial measures, and these non-GAAP measures should be considered in addition to, not as a substitute for or in isolation from, our financial results prepared in accordance with GAAP. Other companies, including companies in our industry, may calculate these non-GAAP financial measures differently or not at all, which reduces their usefulness as comparative measures.

Non-GAAP Operating Loss

We define non-GAAP operating loss as operating loss presented in accordance with GAAP, adjusted to exclude stock-based compensation expenses. We have presented non-GAAP operating loss because we consider non-GAAP operating loss to be a useful metric for investors and other users of our financial information in evaluating our operating performance because it excludes the impact of stock-based compensation, a non-cash charge that can vary from period to period for reasons that are unrelated to our core operating performance. This metric also provides investors and other users of our financial information with an additional tool to compare business performance across companies and periods, while eliminating the effects of items that may vary for different companies for reasons unrelated to core operating performance.

A reconciliation of our GAAP operating loss, the most directly comparable GAAP financial measure, to non-GAAP operating loss is presented below:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
GAAP operating loss	$	(133,934)	$	(178,821)	$	(41,811)	$	(81,015)
Add: Stock-based compensation expense	26,631		23,044		32,329		9,260
Non-GAAP operating loss	$	(107,303)	$	(155,777)	$	(9,482)	$	(71,755)

Non-GAAP Net Loss

We monitor non-GAAP net loss for planning and performance measurement purposes. We define non-GAAP net loss as net loss reported on our consolidated statements of operations, excluding the impact of stock-based compensation expenses and change in fair value of forward contract liability. We have presented non-GAAP net loss because we believe that the exclusion of these charges allows for a more relevant comparison of our results of operations to other companies in our industry and facilitates period-to-period comparisons as it eliminates the effect of certain factors unrelated to our overall operating performance. Our calculation of non-GAAP net loss does not currently include the tax effects of the stock-based compensation expense adjustment because such tax effects have not been material to date.

A reconciliation of our GAAP net loss, the most directly comparable GAAP financial measure, to our non-GAAP net loss is presented below:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
GAAP net loss	$	(127,155)	$	(177,719)	$	(66,605)	$	(77,820)
Add: Stock-based compensation expense(1)	26,631		23,044		32,329		9,260
Add: Change in fair value of forward contract liability	—		—		30,327		—
Non-GAAP net loss	$	(100,524)	$	(154,675)	$	(3,949)	$	(68,560)

_______________

(1)Non-GAAP net loss does not include the tax effects of the stock-based compensation expense adjustment because such tax effects were not material during the periods presented.

RISK FACTORS

Investing in our Class A common stock involves a high degree of risk. You should carefully consider the risks described below as well as the other information in this prospectus, including our consolidated financial statements and the notes thereto, and the section titled “Management’s Discussion and Analysis of Financial Condition and Results of Operations,” before deciding whether to invest in our Class A common stock. The occurrence of any of the events or developments described below may harm our business, financial condition, results of operations, and prospects. In such an event, the price of our Class A common stock could decline, and you may lose all or part of your investment. Additional risks and uncertainties not presently known to us or that we currently believe are not material may also harm our business, financial condition, results of operations, and prospects.

Risks Related to Our Business and Our Industry

We may not sustain our growth rate, and we may not be able to manage future growth effectively.

We have experienced significant growth in a short period of time. Our revenue increased from $24.6 million for the year ended December 31, 2022 to $78.7 million for the year ended December 31, 2023, and from $8.7 million for the six months ended June 30, 2023 to $136.4 million for the six months ended June 30, 2024. We may not achieve similar growth rates in future periods. You should not rely on our results of operations for any prior quarterly or annual periods as an indication of our future operating performance. If we are unable to maintain adequate revenue growth, our financial results could suffer, and our stock price could decline.

To manage our growth successfully and handle the responsibilities of being a public company, we believe we must effectively, among other things:

•recruit, hire, train, and manage additional qualified personnel for our research and development activities;

•continue to make significant investments in our new and existing products;

•add additional sales personnel; and

•implement and improve our administrative, financial, and operational systems, procedures, and controls.

If we are unable to manage our growth effectively, we may not be able to take advantage of market opportunities or develop new products, and we may fail to satisfy customer requirements, maintain product quality, execute our business plan, or respond to competitive pressures, which may harm our business, financial condition, results of operations, and prospects.

We have a history of generating net losses, and if we are unable to achieve adequate revenue growth while our expenses increase, we may not achieve and maintain profitability in the future.

We have a history of generating net losses. We incurred net losses of $127.2 million and $177.7 million for the years ended December 31, 2023 and 2022, respectively, and our net loss for the six months ended June 30, 2024 and 2023 was $66.6 million and $77.8 million, respectively. As of June 30, 2024 and December 31, 2023, we had an accumulated deficit of $728.2 million and $661.6 million, respectively. These losses and our accumulated deficit are a result of the substantial investments we have made to grow our business. We expect our costs will increase over time and our losses may continue if such increases in costs are not more than fully offset by increases in our revenue. We expect to continue to invest significant additional funds in expanding our business and research and development activities as we continue to develop new products. We also expect to incur additional general and administrative expenses as a result of our growth and expect our costs to increase to support our operations as a public company.

If our revenue or revenue growth rate declines or our operating expenses exceed our expectations, our financial performance will be adversely affected. We will need to generate and sustain increased revenue levels in future

periods in order to achieve and maintain profitability. If we cannot successfully grow our revenue at a rate that exceeds the costs associated with our business, we will not be able to achieve and maintain profitability, and the trading price of our Class A common stock could decline.

We have a limited operating history, and we may have difficulty accurately predicting our future revenue for the purpose of appropriately budgeting and adjusting our expenses.

We were established in 2016 and began generating revenue in 2019. Our limited operating experience, a dynamic and rapidly evolving market in which we sell our products, our dependence on one customer, as well as numerous other factors beyond our control, could impede our ability to forecast quarterly and annual revenue accurately. As a result, we could experience budgeting and cash flow management problems, unexpected fluctuations in our results of operations, and other challenges, any of which could make it difficult for us to achieve and maintain profitability and could increase the volatility of the price of our Class A common stock.

We currently generate a significant majority of our revenue from one customer, G42, and a significant portion of our revenue from a limited number of customers. A reduction in demand from, or a material adverse development in our relationship with, G42 or any of our other significant customers may harm our business, financial condition, results of operations, and prospects.

Group 42 Holding Ltd (together with its affiliates, “G42”) accounted for 83% and 87%, respectively, of our total revenue for the year ended December 31, 2023 and six months ended June 30, 2024. Our dependence on our relationship with G42 subjects us to a number of risks. Any negative changes in the demand from G42, in G42’s ability or willingness to perform under its contracts with us, in laws or regulations applicable to G42 or the regions in which it operates, or in our broader strategic relationship with G42 would harm our business, financial condition, results of operations, and prospects. Even if G42 remains satisfied with our offerings, it is possible that it will no longer need to purchase additional AI compute or services at the same quantity as prior periods, or that G42’s ability to purchase our products may change for reasons outside of its control. G42 may also choose to purchase more of its AI compute from our competitors.

Further, as of December 31, 2023, customers representing 10% or more of total accounts receivable consisted of four customers (including G42) who accounted for 43%, 22%, 15%, and 15% of our accounts receivable balance. Two customers accounted for 68% and 16%, respectively, of our accounts receivable balance as of June 30, 2024. This customer concentration increases the risk of quarterly fluctuations in our results of operations and our sensitivity to any material adverse developments experienced by, or in our relationships with, our significant customers. The loss of, any substantial reduction in sales to, or the default on payments by, any of our significant customers may harm our business, financial condition, results of operations, and prospects.

Our business and our products and services are subject to various governmental regulations, and compliance with these regulations may cause us to incur significant expense. Similarly, we are required to obtain export licenses to sell our products to various jurisdictions where we have customers, and we cannot guarantee that we will be successful in obtaining all required licenses in the future. If we fail to comply with applicable regulations, we could be subject to civil or criminal penalties, and if we are unable to obtain licenses to export our products, our business, financial condition, results of operations, and prospects may be harmed.

Our business and our products and services are subject to various domestic and international laws and other legal requirements, including packaging, product content, and labor regulations. Further, we must conduct our business in compliance with international trade laws and regulations, including various import and export control laws and regulations, such as the U.S. Export Administration Regulations (“EAR”), which are administered by the U.S. Department of Commerce’s Bureau of Industry and Security (“BIS”), as well as economic and trade sanctions, including those administered by the U.S. Department of the Treasury’s Office of Foreign Assets Control (“OFAC”). OFAC administers and enforces economic and trade sanctions that restrict U.S. companies from selling or supplying products and services to embargoed jurisdictions and sanctioned parties. U.S. export control laws and regulations restrict or prohibit the export, re-export, and in-country transfer of certain commodities, software, and technology (including certain AI technologies) to restricted countries, governments, persons, and entities.

Our products and services are subject to U.S. export controls, and our products generally may only be exported to customers located in certain countries with prior licensing from the BIS. In particular, in October 2023, BIS announced updated licensing requirements for exports of certain semiconductors and other items, including certain components of our products, to countries in the EAR’s Country Groups D:1, D:4 (which includes the United Arab Emirates, where our strategic partner and largest customer, G42, is headquartered), and D:5. The licensing requirements also apply to the export of these items to a party headquartered in, or with an ultimate parent headquartered in, Country Group D:5.

While we have obtained an export license from BIS to export, reexport, or transfer (in-country) our CS-2 systems to G42 in the United Arab Emirates, all of the systems we have sold to G42, or for which purchase orders have been placed by G42, to date have been or are expected to be deployed in the United States, which does not require an export license from BIS. To the extent that we cannot export to a specific customer without a license from BIS, we may seek a license for the customer. However, the licensing process is time-consuming. There is no assurance that BIS will grant such a license or that BIS will act on the license application in a timely manner. Even if BIS issues a license, it may impose burdensome conditions that we or our customer cannot accept or decide not to accept.

We believe our cloud offerings are in compliance with export, sanctions, and other applicable laws, including exemptions therefrom, and we have implemented certain control mechanisms designed to prevent unauthorized dealings with U.S. embargoed or sanctioned countries.

If we or our customers fail to comply with these regulations, or if users in U.S. embargoed or sanctioned countries circumvent our precautions for compliance with these regulations, we may be unable to export our products and services or we may be required to suspend sales to these customers, and we may be subject to civil monetary fines, criminal sanctions, and other penalties (such as loss of export privileges), which may harm our reputation, business, financial condition, results of operations, and prospects. To the extent we increase our business outside the United States, our risks under these laws and regulations would increase.

Managing these new licenses and other requirements is complicated and time consuming. The requirements may disadvantage us relative to competitors who sell products that are not subject to U.S. export control restrictions or who may be able to acquire licenses for their products that we are not able to obtain. Our competitive position and future results may be further harmed over the long-term if there are further changes in BIS export controls, including further expansion of the geographic, customer, or product scope of the controls applicable to our products, if customers purchase products from competitors, if customers develop their own internal solutions to avoid the need to purchase our products, if we are unable to provide contractual warranty or other extended service obligations, if BIS does not grant licenses in a timely manner or denies licenses relating to significant customers, or if we incur significant transition costs. Even if BIS grants requested licenses, the licenses may be temporary or impose burdensome conditions that we or our customers or end users cannot or choose not to fulfill. The licensing requirements may benefit certain of our competitors, as the licensing process makes our pre-sale and post-sale technical support efforts more cumbersome and less certain and could encourage customers to pursue alternatives to our products, including semiconductor suppliers based in China, Europe, and Israel.

Given the increasing strategic importance of AI and rising geopolitical tensions, BIS has changed and may again change the export control rules at any time and may impose additional export control restrictions and elevated licensing requirements on certain components of our products, which would negatively impact our business, financial condition, results of operations, and prospects. For example, in January 2024, BIS proposed regulations requiring, among other things, U.S. providers of Infrastructure-as-a-Service (“IaaS”) products, and resellers of such products, to verify the identity of their foreign customers, and providers of certain IaaS products to submit a report to the U.S. Secretary of Commerce when a foreign person transacts with that provider or one of its resellers to train a large AI model with potential capabilities that could be used in malicious cyber-enabled activity. If the regulations go into effect as proposed and are deemed to apply to our business, we may be required to expend substantial resources to comply with the verification, reporting, recordkeeping, and resale enforcement requirements and may be restricted in our customer base. In the event of changes to export control rules, we may not be able to sell our products and services, including existing inventory of such products, to existing or potential customers. We also may

not be able to develop replacement products and services not subject to licensing and other requirements, effectively excluding us from all or part of certain markets, including the Middle East and China. While our business in the United Arab Emirates and elsewhere has not been negatively impacted to date by export regulations, we cannot provide assurance that heightened attention on relations with the Middle East, including companies based therein, and such region’s AI ambitions, will not cause the U.S. government to adopt new regulations, or deny us necessary regulatory approvals, that limit our ability to sell our products or services there or result in brand or reputational harm, competitive harm, or financial harm.

Further, any deterioration in relations between the United States and China, Taiwan, the Middle East, and other jurisdictions could lead to additional sanctions or export controls on such countries, regions, and specific individuals or entities, which could impact our ability to sell to or source components from such locales or otherwise negatively impact our business. In addition, trade regulations or other governmental actions targeted at one country or entity may impact other countries or entities. Any decreased use of our products or limitation on our ability to export or sell products to certain regions would adversely affect our business, financial condition, results of operations, and prospects.

We are also subject to anti-corruption laws, including the U.S. Foreign Corrupt Practices Act, which have been enforced aggressively in recent years. Although we have implemented policies and procedures designed to support compliance with relevant economic sanctions, export controls, and anti-corruption laws, there can be no assurance that our employees, partners, contractors, or agents will not violate such laws and regulations or our policies and procedures. Any failure by us to comply with these laws or regulations may have adverse consequences for us, including reputational harm, government investigations, possible loss of export or import privileges, and substantial civil and criminal penalties. We may be required to incur significant expense to comply with, or to remedy violations of, these regulations.

Further changes in trade or national security protection policy, tariffs, additional taxes, restrictions on exports, or other trade barriers, including those taken against the United States in retaliation for U.S. policies or those implemented by the United States in connection with election cycles, may limit our ability to obtain components or raw materials, limit our ability to produce products, increase our selling and/or manufacturing costs, decrease margins, reduce the competitiveness of our products, reduce our ability to sell products, or reduce our ability to have investments or mergers and acquisitions approved by governmental agencies, any of which may harm our business, financial condition, results of operations, and prospects.

If the market does not adopt our products and services, we will be unable to grow our business.

Our success depends on our ability to attract new customers and users to our products and services, which in turn depends on the competitiveness and desirability of our offerings. We expect competition to increase from both existing competitors and new market entrants with products that may be lower priced than ours, may be better optimized than our offerings for emerging technologies, or may provide better performance or additional features not provided by our products or services. Potential customers may also develop their own AI computing products and services that they then may commercialize. There can be no assurance that our solution will gain wide market acceptance. Many of our competitors use different semiconductor compute architectures, platforms, tools, and technology stacks from ours in their products and services, including some, like GPUs, that are already widely used in the industry. Potential customers may be unwilling or unable to convert from legacy AI solutions to our solution, and we may be unable to attract new customers. Current customers may further be dissatisfied or stop purchasing our products if our products are not compatible with or able to be used in combination with other AI solutions, including those of our competitors.

In addition, our products require data centers with sufficient power, equipment, and cooling infrastructures, which may not be readily available to our customers. This may cause customers to delay purchases or delay acceptance of delivery of previously purchased goods, which may harm our financial condition and results of operations, including timing of revenue recognition. While most AI computing solutions require power and other special infrastructure, customers or potential customers may choose to purchase alternative solutions from our competitors.

If our products and services are not widely adopted, or if we are unable to successfully develop and deploy additional go-to-market methods or expand the functional areas in which our products and services compete, our business, financial condition, results of operations, and prospects may be harmed, and we will be unable to grow our business.

If we are unable to expand the application of our products, or if the new products we develop and introduce into the market are not successful, our business, financial condition, results of operations, and prospects may be harmed.

Currently, our AI computing solution is focused on accelerating training and inference for GenAI model architectures, including LLMs and multimodal reasoning models. Part of our success will depend on our ability to expand the application of our solution to other areas of use, such as traditional computer vision, HPC use cases, and new ML application areas yet to be discovered, as well as new or expanded verticals, such as multi-lingual language applications, biomedical, pharmaceutical and life sciences, and financial services. If we are unable to service and capture such use cases, we may lose market share and not be able to grow our business.

Further, in addition to releasing next generation versions of existing products, our business strategy may involve introducing new offerings to the market, which are subject to different risks and uncertainties. There can be no assurance that such new offerings will be broadly accepted. For instance, we recently released our inference solution, and if we fail to win broad customer adoption, we will be unable to participate in a significant segment of the AI market. In addition, if market demand for high speed inference, including for real-time, agent-based or other similar applications, does not exist or increase as expected, our inference solution may not be successful. The competitive landscape, customer preferences, market pressures, and technical and product challenges for new products may be different from our existing offerings. We can provide no assurance that new product offerings will achieve success on a timely basis, or at all.

The market for AI computing solutions is competitive and evolving, and if we do not compete effectively, our business, financial condition, results of operations, and prospects may be harmed.

The market for AI computing solutions is highly competitive and evolving. New technologies and solutions are rapidly emerging in the markets in which we compete. We compete against Nvidia Corporation, a dominant market leader, Advanced Micro Devices, Inc., Intel Corporation, Microsoft Corp., and Alphabet Inc., among others, as well as internally developed custom application-specific integrated circuits and a variety of private companies, some of which are focused on inference-only offerings. Many of these companies are large, well established, and have greater financial resources than we do. In addition, many companies, including potential customers, are developing or may decide to develop their own AI computing solutions based on current or new technologies, including wafer-scale technology, and may, in the future, seek to monetize such solutions in a way that competes with our product and service offerings. Many of our current and potential competitors benefit from competitive advantages over us, such as prominent and cutting-edge technology and software tools designed to keep out new market entrants, greater name recognition, established developer communities, established engineering teams with key industry knowledge, longer operating histories, greater financial assets, more varied product offerings, larger sales and marketing resources, more established relationships, wider geographic presence or larger, more established customer bases, more comprehensive and mature intellectual property portfolios and patent protections, a more robust supply chain, including favorable contracts with suppliers, and other resources. These competitors may also be able to more effectively identify and capitalize upon opportunities in new markets and customer trends, be better able to secure suppliers or may have priority access to suppliers, and obtain sufficient research and development operations, design, manufacturing, and fabrication solutions, foundry and assembly and test capacity than we can, which may harm our business. Competitors with greater financial resources may be able to offer lower prices than us, or they may offer additional products, services, or other incentives that we may be unable to match. Additionally, while we compete against many parties, we also operate in a market with a dominant incumbent. This introduces a number of competitive pressures, including that it is more challenging to gain market awareness of our products, or gain market share. Additionally, a dominant incumbent has the ability to influence the direction of the market and user community in ways that are advantageous to them. In such cases, we have less visibility to be able to accurately predict the direction of the market, and such direction may not be consistent with our strengths and roadmap. If we

cannot influence the AI community to engage with our products and services instead, or if we are unable to compete effectively, or are unable to convert new customers, our business, financial condition, results of operations, and prospects may be harmed.

Additionally, customer expectations and requirements are rapidly evolving. As a result, some of our competitors may be better situated to meet changing customer needs and secure design wins. For example, our software offering is currently optimized around certain AI models, such as LLMs and multimodal reasoning models. Should AI model architectures significantly change, it may take us time to build out the software support, and we may lose potential or current customers and developers in the interim.

While we believe that our current processor design and associated software are optimized for a variety of AI and HPC workloads, including our ability to continue to scale performance through investments in our current architectures, disruptive technologies in compute, including “more than Moore” technologies and innovations, have the potential to severely disrupt current computer processing technologies, including ours. While we continue to invest in advanced system and semiconductor techniques, we may not be able to adapt our technologies, intellectual property, and expertise to such new paradigms, while our established competitors and new entrants specializing in such technologies may. If we are unable to innovate, participate in, and leverage such new fields, or if our solutions are not compatible with other AI solutions, our solutions may become obsolete, we may lose market share and customers, and our business, financial condition, results of operations, and prospects may be harmed.

Further, we sell a premium solution, and our competitors’ solutions may be less expensive to purchase by certain measures, which may limit the number of customers who can or will buy our product. While the market in which we operate is evolving rapidly, as it matures, AI computing solutions may become increasingly commoditized, where many different solutions become sufficient for customer requirements. Factors that may contribute to commoditization include advances in AI compute, advances in or simplifications of AI model architectures that require less sophisticated compute, and consumer preferences, which alone or in combination with other factors may lead to price erosion. If we are unable to establish and maintain a competitive lead that supports premium pricing or offset price reductions with increased sales, our business, financial condition, results of operations, and prospects may be harmed.

We may also experience discriminatory or anti-competitive practices by our competitors that may impede our growth, cause us to incur additional expense, or otherwise harm our business. For example, some of these competitors may use their market power to dissuade potential or current customers from purchasing from us. Our competitors may also establish cooperative relationships among themselves or with third parties or may acquire companies that provide similar products as ours. We expect this trend to continue as companies attempt to improve the leverage of growing research and development costs, strengthen or hold their market positions in an evolving industry, or are unable to continue operations. Certain of our competitors have active merger and acquisition pipelines and are well-funded to inorganically accelerate their product development. In addition, companies that are strategic partners, vendors, or customers may acquire or form alliances with our competitors, thereby reducing their business with us. We believe that industry consolidation has and may continue to result in stronger competitors that are better able to compete as sole-source vendors for customers. There can be no assurance that we will not be forced to engage in price-cutting or margin limiting initiatives, or to increase our advertising and other expenses to attract and retain customers in response to competitive pressures. Any of these factors, alone or in combination with others, may harm our business, financial condition, results of operations, and prospects, and result in a loss of market share and an increase in pricing pressure.

We depend on third-party suppliers, and substantially all of our manufacturing services and components are procured on a purchase order basis without capacity commitments, which may harm our ability to bring products to market and our reputation, business, financial condition, results of operations, and prospects.

We operate a fabless manufacturing model that utilizes third-party suppliers, such as a third-party wafer foundry, as well as system assembly and test service providers in a number of countries, including outside the United States. We depend on one third-party foundry, TSMC, to manufacture our proprietary processor using its fabrication equipment and techniques. We purchase components such as servers, field-programmable gate arrays,

network switches, cooling units, interconnects, and chassis, among many others, from third-party providers. Except for certain assembly, packaging, and testing steps that we perform internally, we do not assemble, test, or package our products, but instead contract with independent contract manufacturers. Most of our products are designed to be manufactured in a specific process, typically at one particular fabrication facility with particular suppliers. In addition, we source a number of the components used in our products from sole or single-source suppliers, or use a single supplier to perform certain of the processes involved in the manufacture of the products. Some of our components, such as our wafer, have long lead times. As a result, we are highly reliant on our suppliers, and depend on them to allocate sufficient manufacturing capacity to meet our needs, develop products of acceptable quality at acceptable yields, and deliver those products to us on a timely basis.

Further, we do not generally have long-term capacity commitments with our suppliers. Substantially all of our manufacturing services and component orders are currently transacted on a purchase order basis, and in many instances we negotiate pricing separately for each purchase order, with our contractors and suppliers having no obligation to perform services or supply products to us for any specific period, in any specific quantities, or at any specific price, except as may be provided in a particular purchase order. It is possible that other customers of our suppliers that are larger and better financed than we are, or have long-term agreements with these suppliers, may induce these suppliers to reallocate capacity to them, which could impair our ability to secure manufacturing, assembly, and testing capacity that we need for our products. Further, our competitors may have long-term capacity commitments with our suppliers, and our suppliers may be obligated to prioritize our competitors under these agreements when supply is limited. Some of these suppliers have experienced in the past, and may experience in the future, shortages and other disruptions to their ability to source and obtain deliveries of necessary raw materials, components, or services, or logistics issues in delivering their products or performing their services. We are dependent on the availability of this capacity to manufacture and assemble our products, and our suppliers have not provided assurances that adequate capacity will be available to us in the future. If our suppliers’ operations are curtailed or disrupted, we may need to seek alternate sources of supply, which may not exist. The lead time needed to identify, qualify, and establish reliable production at acceptable yields with a new supplier is typically lengthy, and there is often no readily available alternative supplier. In addition, qualifying new suppliers is often expensive, and they may not develop products or provide services as cost-effectively as our current suppliers. These shortages and disruptions may, in turn, harm our ability to manufacture and timely deliver our products to our customers. For example, shipment delays in the fourth quarter of 2023 by a supplier resulted in delays in our delivery schedule to a customer. While we met our contractual obligations to the customer in this instance, we can provide no assurance that we will be able to do so in other instances under similar circumstances. We have also had to qualify alternative suppliers and have experienced delays in delivering our products when existing suppliers no longer supported our requirements due to changes in their business. If one or more of our suppliers no longer manufactures our products, fails to perform its obligations in a timely manner or at satisfactory quality levels, or is obligated to prioritize other entities in a constrained environment, our ability to bring products to market and to timely deliver products to our customers may be harmed.

Additionally, if our component suppliers cease to, or become unable to, manufacture a component for us, we may also be forced to make a significant “lifetime” purchase of the affected component or part from an existing supplier, in order to enable us to meet our customer demand or to re-engineer a product. Significant lifetime purchases of such discontinued components could significantly increase our inventory and other expenses, such as insurance costs, and expose us to additional risks, such as the loss of, or damage to, products that may not subsequently be available to us from an alternative source. In any such circumstances, we may be unable to meet our customers’ demands and may fail to meet our contractual obligations and may need to redesign our products. This may result in the payment of significant damages by us to our customers and our revenue may decline, harming our business, financial condition, results of operations, and prospects. We may also have excess inventory if we overestimate the “lifetime” purchase quantity, which would result in write-offs.

There are certain components in our products that require long lead times to manufacture and deliver, including the wafer for our WSE. Accordingly, we must make estimates of demand well in advance of certainty of revenue. In addition, the yield on components, including our wafers, is variable and thus we must account for a certain amount of waste when purchasing components. If our estimates of customer demand are ultimately inaccurate, as we have experienced from time to time, there could be a significant mismatch between supply and demand. This mismatch

has in the past resulted, and may in the future result, in both product shortages and excess inventory, with a corresponding impact to our results of operations. If we underestimate our customers’ future demand for our products, our suppliers may not have adequate lead-time or capacity to increase production and we may not be able to obtain sufficient inventory to fill orders on a timely basis or in a cost-effective manner. If we fail to fulfill our customers’ orders on a timely basis, or at all, our customer relationships could be damaged, we may lose revenue and market share and our reputation may be harmed. If we overestimate our customers’ future demand, we may have excess inventory with limited or no resale value, which would harm our business, financial condition, results of operations, and prospects.

Some of our customer agreements provide for, and future customer agreements may also require, certain remedies if we do not deliver our products and services in a timely manner or by a certain date. If our third-party suppliers do not perform services or deliver components on our expected timeline, including due to force majeure or other reasons beyond their control, we may owe damages or incur other liabilities with our customers. In addition, certain of our customers prepay for hardware purchases and are entitled to refunds of such advances in certain situations.

Prior to our wafer scale engine, a full wafer-sized processor had not previously been successfully manufactured. We worked with TSMC to develop the processes necessary to manufacture the semiconductor wafers needed for our wafer scale engine, which involve many complexities and proprietary technologies. We are currently dependent on TSMC to produce all of the wafers that we use in our products. We have no formalized long-term supply or allocation commitments from TSMC, and TSMC also fabricates wafers for other companies, including certain of our competitors, many of whom are significantly larger than us and purchase considerably more wafers from TSMC than we do. TSMC could reduce or eliminate deliveries to us on short notice, or raise their prices to us, all of which may result in loss of revenue opportunities, damage our relationships with our customers, and harm our results of operations and gross margin. Many of our suppliers, including TSMC, are located in areas of high geopolitical tension, such as Taiwan and South Korea, or in areas prone to natural disasters, such as earthquakes and typhoons. Any substantial disruption in TSMC’s supply of wafers to us, or in the other contract manufacturing services that we utilize, as a result of a natural disaster, climate change, water shortages, political unrest, military conflict, geopolitical turmoil, trade tensions, inflation, government orders, global pandemics or health crises, economic instability, equipment failure, or other causes, may harm our business, financial condition, results of operations, and prospects.

Our supply chain is long, complex, and global, with many interdependencies. Any significant fluctuations of supply and demand or disruption to our supply chain may harm our ability to manufacture and deliver our products to our customers.

Our products depend on many different electronic components of varying complexity and raw materials, as well as software tools for the design of our products. Accordingly, our supply chain is dependent on many aspects of the global semiconductor industry, which is highly cyclical and subject to wide fluctuations of supply and demand as a result of rapid technological change, rapid product obsolescence and price erosion, evolving standards, and frequent new product introductions, among other reasons. The industry has experienced significant downturns during recent global recessions, characterized by diminished product demand, production overcapacity, and high inventory levels. Any upturn in the semiconductor industry could result in increased competition for access to suppliers. Additionally, the impact of industry-wide trends on specific segments and suppliers is difficult to predict.

Our products have long and complex bills of materials, with several long lead time components and single-source suppliers for certain critical components. For example, TSMC fabricates all of our wafers, and switching to a different foundry would require development of new proprietary processes and technologies, which may be expensive and increase production time. Some of our suppliers are small and are unable to accommodate fluctuations in required capacity. In other cases, our requirements represent a small portion of a supplier’s business, and therefore we may not be a priority for such supplier.

Our supply chain is also subject to complex and frequently changing rules and regulations regarding global trade, such as imports, exports, tariffs, and taxes, as well as the effects of U.S. and foreign government programs and

initiatives regarding the semiconductor supply chain and other geopolitical forces. For example, both the United States and China have national programs directed toward influencing the future of the global semiconductor industry, and the effects of such programs on our business are not easily predictable.

Further, force majeure and other events beyond our and our suppliers’ control, including the COVID-19 pandemic and natural disasters such as earthquakes in Taiwan, have in the past and may in the future result in disruption to parts of our supply chain, including difficulty procuring wafers and other necessary components in a timely fashion, with suppliers increasing lead times or placing products on allocation, requiring committed, non-cancelable purchases and raised prices. Such events may excuse performance of certain suppliers but not excuse our performance to our customers, which may cause us to incur liabilities.

Transportation disruptions, such as shortages in available cargo capacity or labor availability, changes in frequency of service, and shipping channel disruptions, may also increase delivery times and costs for materials and components to our facilities, transfers of our products to our key suppliers, and may affect our ability to timely ship our products to customers. In addition, sometimes we bear the risk of loss during transportation of components from our suppliers and products to our customers.

In periods when broad fluctuations or changes in business conditions occur, it is difficult to assess the impact on our business. We have experienced substantial period-to-period fluctuations in our results of operations and expect, in the future, to experience period-to-period fluctuations in our results of operations due to these changes in business conditions.

Raw material and component price fluctuations or decreased availability of certain raw materials or components can increase the cost of our products, impact our ability to meet customer commitments, and may adversely affect our business, financial condition, results of operations, and prospects.

The cost of raw materials and components is a key element in the cost of our products, both directly and indirectly through our suppliers. Our inability to offset material price inflation may harm our business, financial condition, results of operations, and prospects. Although we believe that sources of supply for raw materials and components are generally adequate, it is difficult to predict what effects shortages or price increases may have in the future. Our inability to meet our supply needs would jeopardize our ability to fulfill obligations under our contracts, which could, in turn, result in reduced sales and profits, contract penalties or terminations, and damage to our customer relationships. Furthermore, increases in the price of certain raw materials may result in increased production costs and may result in a decrease in our gross margins. We may choose not to, or may not be able to, pass on such price increases to our customers. In addition, because we primarily outsource manufacturing of our products, global market trends such as shortage of capacity to fulfill our fabrication needs also may increase our raw material costs and thus decrease our gross margins.

Dependency on third-party suppliers and their technology to manufacture, assemble, test, or design our products reduces our control over product quantity and quality, manufacturing yields, development, enhancement, and product delivery schedules and could harm our business.

Except for certain assembly, packaging, and testing steps that we conduct internally, we depend on third-party suppliers to manufacture, assemble, and test our products, and utilize third-party development tools for design, simulation, and verification of new products. Further, while we primarily generate intellectual property internally, we have, on occasion, leveraged certain third parties for software development purposes. We also face several other risks related to our third-party suppliers that have adversely affected or could adversely affect our ability to meet customer demand and scale our supply chain, negatively impact longer-term demand for our products and services, and harm our business or results of operations, including:

•lack of guaranteed supply of wafer, component, and capacity or decommitment and potential higher wafer and component prices, from long lead times, incorrectly estimating demand, and failing to place orders with our suppliers with sufficient quantities or in a timely manner;

•failure by our foundries or suppliers to procure raw materials or provide adequate levels of manufacturing or test capacity for our products;

•failure by our foundries to develop, obtain, or successfully implement high quality process technologies, including transitions to smaller geometry process technologies such as advanced process node technologies and memory designs needed to manufacture our products;

•failure by our suppliers to comply with our policies and expectations and emerging regulatory requirements;

•limited number and geographic concentration of global suppliers, foundries, contract manufacturers, assembly, and test providers;

•loss of a supplier and additional expense and/or production delays as a result of qualifying a new foundry or other supplier and commencing volume production or testing in the event of a loss, addition, or change of a supplier;

•lack of direct control over product quantity, quality, and delivery schedules;

•responsibility or lack of recourse for loss or damage incurred while our products are in transit;

•failure of our suppliers or their suppliers to supply high-quality products and/or making changes to their products without our qualification;

•delays in product shipments, shortages, a decrease in product quality, and/or higher expenses in the event our subcontractors or foundries prioritize our competitors’ or other customers’ orders over ours;

•requirements to place orders that are not cancellable upon changes in demand or requirements to prepay for supply in advance;

•low manufacturing yields resulting from a failure in our product design or a foundry’s proprietary process technology;

•disruptions in manufacturing, assembly, and other processes due to closures related to heat waves, earthquakes, fires, or other natural disasters and electricity conservation efforts; and

•failure by our third-party software development tools to meet consumer demands for greater functionality from the design, simulation, and verification of new products or product enhancements.

The complexity of our products and services could result in unforeseen delays or expense or undetected defects, bugs, or security vulnerabilities, which may harm the market acceptance of new products and services, damage our reputation with current or prospective customers, cause significant remediation expenses, and may harm our business, financial condition, results of operations, and prospects.

Our AI computing solutions, including both hardware and software, are highly complex to design and complicated to make, involve novel manufacturing processes, and may contain defects, bugs, or security vulnerabilities or experience failures or unsatisfactory performance due to a variety of issues, including design, fabrication, packaging, materials, or use. Our products and services may contain defects when they are first introduced or as new versions or enhancements are released, or their release may be delayed due to unforeseen difficulties during product development. We shipped our first CS-1 system in 2020 and our first CS-2 system in 2021, and first introduced our CS-3 system and our inference solutions in 2024. As a result, we and our customers have not had a significant amount of time to gauge the long-term performance and reliability of our products and services. In addition, while we have designed our platform to be able to seamlessly scale up to 2,048 CS systems, the largest cluster deployed as of September 27, 2024 is comprised of over 100 CS systems. If our AI solutions, third-

party components used in our products, the AI models we create or train on behalf of our customers, or our inference output, contain defects, bugs, or vulnerabilities or have reliability, quality, or compatibility problems, we may not be able to successfully design workarounds or resolve the issues in a timely manner. For example, the AI models may not generate the responses that our customers want or expect, or may be otherwise flawed. Any defects discovered in our products or services, or any components included in our products, prior to their shipping or release could delay our ability to deliver products or services to our customers. Further, defects in our products, services, or the AI models we build or help train may only be discovered after a product or model has been shipped, released, or used by our customers. Our products and services may also have undiscovered vulnerabilities that could be exploited by hackers or other unscrupulous third parties who develop and deploy viruses and other malicious software programs, thereby exposing our customers to potential adverse consequences. Additionally, our products may be difficult to repair in the field and may need to be shipped back to our facility for repair. Failures of our products to perform to specifications, or other product defects, have caused us to incur, and could cause us to incur in the future, significant warranty, support, and repair or replacement costs and divert the attention of our engineering personnel from our product development efforts to find and correct the issue or to design workarounds. Further, issues with, or subpar performance of, AI models that we develop or help develop for customers may require additional and unexpected personnel cost to remedy. In addition, delays in shipping or releasing products or training AI models or the discovery of an error or defect in new products or software after commencement of commercial shipment or release could harm our relationships with customers, as well as the perception of our brand, and may result in failure to achieve market acceptance of our products and services and harm our business, financial condition, results of operations, and prospects.

We have experienced, and may continue to experience, difficulties in transitioning to more advanced process node technologies and in developing advanced packaging solutions.

To remain competitive, we have transitioned, and expect to continue to transition, our processors to be manufactured in accordance with increasingly smaller process node technologies. For instance, the CS-2 was based on TSMC’s 7nm process technology while the CS-3 is based on TSMC’s 5nm process technology. We also evaluate the increased costs of designing and manufacturing to more advanced process nodes, including both actual costs and opportunity costs related to the technologies we choose to forego. These complex transitions are imperative for us to remain competitive with the rest of the industry.

We are dependent on our relationship with our foundry partner to transition to smaller process nodes successfully. We cannot be sure that the third parties we use will be able to effectively manage any future transitions, on our desired schedule, or at all. For example, we currently rely on TSMC for the production of all of our wafers. Developing the necessary technology with another foundry would increase our time to market. We have encountered, and may continue to encounter, delays during tape-out of new designs at smaller process nodes, including securing tape-out slots on our preferred timing. If we cannot timely resolve technical and operational issues, we could experience significant delays in manufacturing our next generation product or fail to deliver a next generation product at the most advanced node. If TSMC or another foundry is not able or not willing to manufacture wafers for our future generation products in sufficient quantities or on our preferred timeline to meet customer demand, it could negatively impact our business. In addition, we could experience reduced manufacturing yields, delays in product deliveries, and increased expenses, all of which may harm our relationships with our customers and our results of operations.

We expended significant resources to develop techniques for power delivery and cooling of our wafer scale chip design. We will need to continue investing in innovations in these areas, as well as other aspects of advanced package design, to achieve our product roadmap objectives and remain competitive. If we reach the limits of being able to power, cool, and provide I/O to our processors, or if we encounter other complexities of packaging our chip, such as advanced integration requirements, we would need to develop alternative solutions to remain competitive. Failure to do so may harm our reputation, business, financial condition, results of operations, and prospects.

Failure at tape-out or failure to achieve the expected final test yields for our products may harm our results of operations.

The tape-out process, which marks the finalization of semiconductor design and the transition to manufacturing, is a critical milestone in our product development cycle and involves inherent risks that could adversely affect our business, financial condition, results of operations, and prospects. The tape-out process requires considerable investment in time and resources and close cooperation with the wafer foundry, and repeated failures can significantly increase our costs, lengthen our product development period, and delay our product launch. If the tape-out or testing of a new chip design fails, either as a result of design flaws by our research and development team or problems with production or the testing process by the wafer foundry, we may incur considerable costs and expenses to fix or restart the design process. Such obstacles may delay the launch of new products and harm our business, financial condition, results of operations, and prospects.

New developments and enhancements of our products and services require significant investment. Our failure to develop new or enhanced products and services and introduce them in a timely and effective manner would undermine our competitiveness.

We have invested, and expect to continue to invest, significant resources in the development of our AI compute solutions, both at a hardware and software level. For example, we periodically publish software releases that may contain significant new features or improvements, which require personnel and IT resources to develop and deploy. Similarly, we periodically develop and release new versions of our hardware, including new generations of processors at smaller process nodes, which also requires significant financial and engineering resources, including design, fabrication, assembly, and testing costs. We recently introduced our inference solution, which has new software and novel techniques in computation that required significant investment.

Our failure to anticipate or timely adapt to changes in AI model architecture and other developments in AI technologies could result in our current and next generation products not meeting the most prevalent customer needs, leading to the loss of customers and decreased demand for our products and services. For instance, we have developed software kernels that are optimized for AI solutions built on transformer architectures and designed our hardware to be optimized for models with high compute demand and model parallel scaling. However, AI architecture is a rapidly evolving field, and the design cycles for our products can be long. Therefore, there can be no assurance that our solutions will be optimized for future AI workloads. Our solution currently integrates with the PyTorch ML framework to enable familiar programmability for developers. New deep-learning frameworks emerge periodically, and while we intend to evolve our solution to support those that become popular and widely adopted, we can provide no assurance that we will be able to support all such frameworks with existing or future versions of our software. Likewise, our inference solution was initially designed to support a limited set of leading third-party foundational models. While we intend to expand the number of supported models, we must make decisions as to the order in which popular foundational models will be integrated into our solution. If we do not timely expand our inference solution to support models that become popular within the AI community, our inference product may lose current and potential customers.

At the same time, the AI compute market is subject to rapid technological change, product obsolescence, design changes, and frequent new product introductions and feature enhancements. For example, continuous and rapid advances in semiconductor technology compel providers of compute to release new and improved products at a regular cadence. Failure to timely introduce new offerings that leverage the latest hardware and software advancements in computing or in AI may result in diminished relevance, decreased competitiveness, and loss of market share, which may harm our reputation, business, financial condition, results of operations, and prospects. Our ability to generate usage of additional products by our customers may also require increasingly sophisticated, more costly and differentiated sales efforts and result in longer sales cycles. In addition, adoption of new products or enhancements may put additional strain on our customer support or compliance teams, which could require us to make additional expenditures related to further hiring and training. Further, the investment required to stay abreast of innovations in AI and AI compute, including research and development and intellectual property expenditures and capital investments, is substantial. If we allocate our resources incorrectly, we may impair our product development cycle and erode our competitive positioning. Resource constraints may also require us to discontinue existing

products or services to support new offerings. The effects of such business model changes may be difficult to predict and may not be evident for a long period of time. If we choose to focus on the wrong products and services, it may harm our business, financial condition, results of operations, and prospects.

As a result of the rapid evolutions in AI technologies and the length of our product development cycle, we must make decisions as to where to dedicate resources and develop product functionality based on our assessment of which evolving AI technologies we think are likely to become widely adopted in the future. Our ability to successfully compete and to continue to grow our business depends in significant part upon our ability to develop, introduce, and sell new and enhanced products on a timely and cost-effective basis, and to correctly anticipate and respond to changing customer requirements. If we are incorrect in our assessments, we will have expended financial and engineering resources that may not generate revenue, and which may harm our competitive position, to the extent we do not have adequate solutions for use cases that become prevalent. Lack of market acceptance for our new products for any reason could jeopardize our ability to recoup substantial research and development expenditures and may harm our reputation, business, financial condition, results of operations, and prospects.

We have experienced, and may experience in the future, delays and unanticipated expenses in the development and introduction of new products. A failure to develop products with required feature sets or performance standards or a delay in bringing a new product or service to market may significantly reduce our return on investment as well as our sales, all of which may harm our business, financial condition, results of operations, and prospects. Delays in the development of new products and services or product and service enhancements could also benefit our competitors and allow them to achieve greater market share. We cannot assure that our future product development efforts will be successful or result in products that gain market acceptance.

Our cloud-based offering is subject to certain risks and challenges. Unfavorable or uncertain conditions in the cloud market, as well as for AI infrastructures, may cause fluctuations in our results of operations.

Our offerings include a cloud-based platform, where customers can purchase our AI computing solutions on a consumption-based model. This business model is intended to allow customers to choose the solution that best aligns with their budgetary, security, and scalability requirements. The risks and challenges of a cloud-based business model may be different from selling our products for on-premise use. For instance, building and scaling a cloud-based platform requires us to lease data centers with sufficient power and cooling infrastructures and to maintain sufficient units on hand, which may require resources to be reallocated from producing units that could otherwise be sold to customers for on-premise use. In order to support a cloud-based business model, we may also incur additional IT and personnel costs to service and manage our clusters, as well as rely on third-party infrastructure, such as Amazon Web Services for certain data transmission needs. There is no assurance that such investments will lead to increased revenue as compared to other business models. In addition, we recently introduced a cloud-based inference solution with a free and paid tier. Such product may present new and additional risks to us, including substantially increasing the number of users accessing our cloud, which may increase cybersecurity risk as well as risk of outages due to overcapacity or otherwise.

Additionally, if domestic and global economic conditions worsen, or adoption, use, and commercialization of AI technology, particularly of the large models that are currently dominating the market, do not progress as expected, overall spending may be reduced, which could adversely impact demand for our solutions in these cloud and AI infrastructure markets and we may not realize our expected growth rates. Furthermore, changes to data privacy laws regarding cross-border transmission of data may disproportionately impact cloud-based platforms. In such cases, we may be left with excess units in our clouds and data center leases and costs that we cannot terminate or recoup. Even if the demand for cloud and AI-related infrastructure develops in the manner or in the time periods we anticipate, if we do not have timely, competitively priced, market-accepted solutions or personnel available to timely meet our customers’ needs, we may miss a significant opportunity and our business, financial condition, results of operations, and prospects may be harmed.

Pricing for the current generation of our existing products often decreases over time, which may harm our business, financial condition, results of operations, and prospects.

We recently launched the CS-3 and no longer commercialize the CS-2. In the future, we expect to release new generations of our AI compute system, which may require us to reduce prices of then-current systems in anticipation of such releases. In addition, customers may delay purchasing an existing system in anticipation of our release of a next generation or newer, more advanced system. Our business, financial condition, results of operations, and prospects may be harmed by a decline in the pricing for the current generation of our existing products. Additionally, as competition increases in our target markets, we may need to reduce the prices of our existing products and services in anticipation of competitive pricing pressures, new product introductions by us or our competitors, new generations of such existing products, or for other reasons. If we are unable to offset any reductions in pricing for existing products by increasing our sales volumes or introducing new products or services or new product generations of such existing products with higher margins, our business, financial condition, results of operations, and prospects may be harmed. To maintain our results of operations, we must develop and introduce new products and services, next-generation products, and product enhancements on a timely basis and continually reduce our costs as well as our customers’ costs. Failure to do so may harm our business, financial condition, results of operations, and prospects.

If our customers do not purchase additional products from us or expand their purchases with us in the future, our business, financial condition, results of operations, and prospects could be harmed.

In order for us to maintain or improve our results of operations, it is important that our customers continue to purchase products from us on similar or improved terms, and that we increase the products and services our customers purchase from us. Our on-premise system purchases are generally handled on a purchase order basis, and customers are not otherwise required to purchase specific or additional quantities of products from us. Likewise, our inference solution can be purchased by developers on a pay-as-you-go model without any long-term commitments. Even when customers agree to purchase an agreed quantity of hardware products from us, we generally do not recognize revenue until shipment or delivery, depending on shipping terms, or until we have met the contractual acceptance terms. As a result, we may not generate the amount of revenue on the timeline that we expect. Moreover, our customers’ purchasing power have, in some cases, given them the ability to make greater demands on us with regard to pricing and contractual terms. We expect this trend to continue, which may adversely affect our gross margin and, should we fail to perform under these arrangements, we could also be liable for significant monetary damages. If our efforts to maintain and expand our relationships with our existing customers are not successful, our business, financial condition, results of operations, and prospects may be harmed.

Our sales cycles can be long and unpredictable, and our sales efforts require considerable time and expense. If our sales cycle lengthens or we invest substantial resources pursuing unsuccessful sales opportunities, our business and results of operations may be harmed.

Our results of operations may fluctuate, in part, because of the resource-intensive nature of our sales efforts, the length and variability of the sales cycle for our products, and the difficulty in making short-term adjustments to our operating expenses. The timing of our sales is difficult to predict. The length of our sales cycle, from initial evaluation to purchase of our product, can vary substantially from customer to customer. Our sales efforts involve educating our customers about the use, technical capabilities, and benefits of our solution. Customers often undertake a lengthy evaluation of our products, which may involve not only our solution but also those of our competitors. Potential customers have and may continue to request paid proof of concept trials prior to purchase, which can be time intensive for our sales and technical personnel. In addition, sales to large organizations and government agencies, and in highly regulated verticals, tend to take longer to conclude because such organizations typically go through a significant evaluation and lengthy negotiation process due to their size, organizational structure, and internal approval requirements. We may spend substantial time, effort, and money on sales efforts without any assurance that our efforts will produce any sales. As a result, it is difficult to predict if or exactly when we will make a sale to a potential customer, or if we can increase sales to our existing customers. New product developments or releases may lead to changes in the size, quantity and complexity of our customer relationships, making it more difficult for us to predict timing of our sales or the expected impact to our results of operations. If

our sales cycle lengthens or we invest substantial resources pursuing unsuccessful sales opportunities, our business, financial condition, and results of operations may be harmed.

Each individual sale tends to be large as a proportion of our overall sales, which has impacted our ability to accurately forecast revenue and manage cash flows. In addition, sales have, in some cases, occurred in later quarters than anticipated, or have not occurred at all, and within each quarter, it is difficult to predict when a transaction will close. Therefore, it is often difficult to determine whether we will meet our quarterly or annual expectations until near the end of such periods. Many of our expenses are relatively fixed or require time to adjust. If expectations for our business are not accurate, we may not be able to adjust our cost structure on a timely basis, and our margins and cash flows may differ from expectations.

The estimates of market opportunity and forecasts of market growth included in this prospectus may prove to be inaccurate, and even if the markets in which we compete achieve the forecasted growth, our business may not grow at similar rates, or at all.

Market opportunity estimates and growth forecasts included in this prospectus, including third-party market research and internal estimates, are subject to significant uncertainty and are based on assumptions and estimates which may not prove to be accurate. While we believe the information on which we base our market opportunity estimates and forecasts is generally reliable, such information is inherently imprecise. Furthermore, even if the markets in which we compete meet the size estimates and growth forecasts included in this prospectus, our business may not grow at similar rates, or at all. For example, we are in the early stages of commercializing our inference solution, and there can be no assurance that we will be able to capture any significant part of the inference market. Moreover, the market for high speed inference, including real-time and agent-based systems, may not exist or increase as expected. Our growth is subject to many factors, including our success in implementing our business strategy, which has many risks and uncertainties, including those described herein. If our market opportunity estimates or growth forecasts prove to be inaccurate, our future growth opportunities may be lower than we currently expect.

Issues relating to the responsible use of our technologies may result in reputational or financial harm and liability.

As with many new emerging technologies, AI presents risks and challenges and increasing ethical and legal concerns relating to its responsible use that could affect the adoption of AI, and thus our business. Concerns relating to the responsible use of new and evolving technologies in our products and services may also result in reputational or financial harm and liability and may cause us to incur costs to resolve such issues. For instance, we offer AI model services in which we develop proprietary AI models with and on behalf of our customers, as well as open-source AI models that we or our customers make generally available to the community. We may not have insight into, or control over, how our customers and other third parties use or deploy the AI models that we trained or assisted in training, or that were trained using our computing solutions, or that we otherwise make available to customers. We do not control how others, including customers, use AI models that we develop or make available. We also cannot fully control how users interact with our inference solution, including whether they may violate our terms of use or that of third-party models with which we integrate. If we enable or offer AI models that draw controversy due to their perceived or actual impact on society, including, for example, AI models that have unintended consequences, infringe intellectual property rights or rights of publicity, disseminate illegal, inaccurate, defamatory, or harmful content, or are controversial because of their impact on human rights, privacy, cybersecurity, employment or other social, economic or political issues, or if we are unable to develop effective internal policies and frameworks relating to the responsible development and use of AI models, we may experience brand or reputational harm, competitive harm, financial harm, or legal liability. Complying with multiple laws, statutes, regulations, self-regulatory frameworks, and industry standards from different jurisdictions related to AI could increase our cost of doing business, may change the way that we operate in certain jurisdictions, or may impede our ability to offer certain products and services in certain jurisdictions if we are unable to comply with applicable legal requirements. Compliance with existing and proposed government regulation of AI, including in jurisdictions such as the European Union (the “EU”), as well as under any U.S. regulation adopted in response to the Biden Administration’s October 2023 executive order on the Safe, Secure, and Trustworthy Development and Use of

Artificial Intelligence (“2023 AI Order”), may also increase the cost of related research and development and compliance, and create additional reporting or transparency requirements. In addition, unfavorable developments with evolving laws and regulations worldwide related to AI, such as those laws that may pause or inhibit continued development or adoption of AI, may limit global adoption, reduce demand for our products and services, increase our costs to provide our products and services, impede our strategy, and negatively impact our long-term expectations in this area. For example, given the adoption of the EU AI Act (the “AI Act”) in 2024, we anticipate that there will continue to be significant developing laws and regulations with respect to AI as the AI industry continues to develop. Changes in AI-related regulation may disproportionately impact and disadvantage us and require us to change our business practices, which may harm our results of operations. Our, our customers, or others’ failure to adequately address any of the foregoing concerns or regulations relating to the responsible use of AI may undermine public confidence in AI and slow adoption of our products and services or harm our reputation or business, financial condition, results of operations, and prospects.

Any failure to offer high-quality maintenance and support services for our customers may harm our relationships with our customers and, consequently, our business.

Once our solution is deployed, our customers depend on our maintenance and support teams to resolve technical and operational issues. Our ability to provide effective customer maintenance and support is dependent on our ability to attract, train, and retain qualified personnel with machine learning, software programming, IT, and complex hardware maintenance experience. In particular, our hardware systems are very complex and difficult to repair in the field, with many components that are challenging and expensive to upgrade. When we sell hardware units for installation on premise at our customers’ facilities, we may in certain circumstances provide spare units that may be used during repair downtime. Where we also provide cluster management services for hardware that we have sold, our customer agreements typically provide for service credits in the event we do not meet certain uptime requirements.

Additionally, growth in our customer base puts additional pressure on our customer maintenance and support teams. If we are unable to provide efficient, high-quality customer maintenance and support services or if we are unable to hire sufficient additional maintenance and support personnel in a timely or efficient manner, we may incur significant warranty, support, and repair, replacement, or service credit costs, and our business and our relationships with our customers may be harmed.

If we are not able to maintain and enhance our reputation and brand recognition, our business, financial condition, results of operations, and prospects may be harmed.

We believe that maintaining and enhancing our reputation and brand recognition is critical to our relationships with existing customers and our ability to attract new customers. The promotion of our brand may require us to make substantial investments and we anticipate that, as our market becomes increasingly competitive, these marketing initiatives may become increasingly difficult and expensive. Our marketing activities may not be successful or yield increased revenue, and to the extent that these activities yield increased revenue, the increased revenue may not offset the expenses we incur, and our results of operations may be harmed. In addition, any factor that diminishes our reputation or that of our management, including failing to meet the expectations of our customers and public or investor perception, may make it substantially more difficult for us to attract new customers. Similarly, because our customers often act as references for us with prospective new customers, any existing customer that questions the quality of our work or that of our employees could impair our ability to secure additional new customers. If we do not successfully maintain and enhance our reputation and brand recognition with our customers, our business may not grow and we may lose these relationships, which may harm our business, financial condition, results of operations, and prospects.

Events outside of our control, adverse economic conditions, or economic uncertainty may harm our business, financial condition, results of operations, and prospects.

Disruptive and/or unpredictable global events beyond our control (such as geopolitical conflict or terrorism, including the ongoing conflict between Russia and Ukraine and in the Middle East, global pandemics or health

crises, natural disasters, or labor unrest) may increase the severity of political and economic volatility around the world, and could result in a protracted localized or widespread decrease in demand for our products. This decrease in demand, depending on its scope and duration, could significantly impact and harm our business, financial condition, results of operations, and prospects over the short and long-term.

Additionally, many of our suppliers, including our sole source of wafers, TSMC, are located in areas of high geopolitical tension, such as Taiwan and South Korea, or in areas prone to natural disasters, such as earthquakes and typhoons. Any substantial disruption in TSMC’s supply of wafers to us, or in the other supply services that we utilize, as a result of a natural disaster, climate change, water shortages, political unrest, military conflict, geopolitical turmoil, trade tensions, government orders, medical epidemics, economic instability, equipment failure or other causes, may harm our customer relationships, business, financial condition, results of operations, and prospects. Further, a significant majority of our revenue is generated from one customer headquartered in the United Arab Emirates. Recent political instability and protests in the Middle East have caused significant disruptions to many industries. Conflict, political instability, and social unrest in, policy changes with, or negative public sentiment toward, the Middle East may significantly harm our business, financial condition, results of operations, and prospects.

Further, adverse macroeconomic conditions, including a general slowdown in the United States or global economy or in a particular region or industry, inflation, recession, changes to fiscal and monetary policy, tighter credit, higher interest rates, new or increased trade tariffs, or an increase in trade tensions with United States trading partners, may harm our business, financial condition, results of operations, and prospects. A deterioration in economic conditions may cause our customers to reduce, delay or forego technology spending, including spending on our products. Significant inflation may increase our costs and expenses, which we may choose not, or not be able to, pass on to our customers. In addition, uncertainty about, or a decline in, global or regional economic conditions may have a significant impact on our suppliers, some of which are located outside the United States. If customers or prospective customers elect not to purchase our products or services as a result of a weak economy or rising inflation and increased costs or otherwise, or our suppliers are unable to continue supplying our products, our business, financial condition, results of operations, and prospects may be harmed. Potential effects include financial instability, inability to obtain credit to finance operations and insolvency.

Risks Related to Operations

Our business and operations have experienced significant growth, and if we do not effectively manage our growth, or are unable to improve our systems and processes, our business, financial condition, results of operations, and prospects will be harmed.

We have grown rapidly since we were founded in 2016. For example, our total headcount has grown to approximately 400 people as of September 26, 2024, and we have employees located in the United States, Canada, India, and other countries. Our customer base has also grown and we expect it to continue to grow. The rapid growth and expansion of our business places a continuous and significant strain on our management, operational, and financial resources.

To manage future growth effectively, we will need to expand and improve our operating, financial, IT and other administrative systems, controls, and procedures, and continue to manage headcount expansion and capital in an efficient manner. We may not be able to successfully implement requisite improvements to these systems, processes, controls, and procedures in a timely or efficient manner. Any failure to improve our systems and processes, or their failure to operate in the intended manner, whether as a result of the significant growth of our business or otherwise, may result in our inability to manage the growth of our business or to accurately forecast our revenue, expenses and earnings, or to prevent certain losses or replace anticipated revenue that we do not receive as a result of delays arising from these factors. Moreover, the failure of our systems and processes could undermine our ability to provide accurate, timely and reliable reports on our results of operations and financial condition and could adversely impact the effectiveness of our internal controls over financial reporting. In addition, our systems and processes may not prevent or detect all errors, omissions, or fraud. Accordingly, our results of operations in future reporting periods may be below the expectations of investors.

We will also face increased compliance costs associated with growth and the expansion of business, particularly as we expand into new countries, if the number of our government customers increases, and as we become a public company. We have encountered and expect to continue to encounter additional risks and difficulties frequently experienced by growing companies in rapidly changing industries, including effectively managing increasing expenses, and allocating valuable financial and other resources as we continue to grow our business. If we do not manage these risks successfully, our business, financial condition, results of operations, and prospects will be harmed.

We are subject to the risks associated with conducting business operations outside the United States, which may harm our business.

In addition to our U.S. operations, we also have international operations, primarily in research and development and operations, and foreign subsidiaries in Canada and India. We also on occasion provide services at our customers’ facilities, including those not located in the United States, and engage in sales and marketing efforts in many foreign jurisdictions. International activities are subject to the uncertainties associated with international business operations, including global laws and regulations, economic sanctions, tax regulations, import and export regulations, duties, tariffs, and other trade restrictions and barriers, changes in trade policies, anti-corruption laws, foreign governmental regulations, potential vulnerability of and reduced protection for intellectual property, and our ability to acquire and retain local employees, any of which may harm our business, financial condition, results of operations, and prospects. Our business, financial condition, results of operations, and prospects may also be harmed in the event of political conflicts, economic crises, wars, or other changes in international relations affecting countries where our subsidiaries, manufacturers, suppliers, and customers are located.

A deterioration in relations between the United States and any country in which we have significant operations or sales, or the implementation of government regulations in such a country, may result in the adoption or expansion of trade restrictions, including economic sanctions and export license requirements, that may harm our business.

We plan to expand our international operations, including to jurisdictions where we have limited operating experience and may be subject to increased business and economic risks that may harm our financial condition and results of operations.

We plan to continue the international expansion of our business operations. We may enter new international markets where we have limited or no experience in marketing, selling, and deploying our products and services. If we fail to deploy or manage our operations in these countries successfully, our business and operations may suffer. In addition, we are subject to a variety of risks inherent in doing business internationally, including:

•political, social and/or economic instability, including as a result of the ongoing conflict between Russia and Ukraine and in the Middle East;

•risks related to governmental regulations in foreign jurisdictions and unexpected changes in regulatory requirements and enforcement;

•existing trade laws and regulations, including those related to exports and deemed exports of U.S. technology;

•changes in trade relationships, including the imposition of new trade restrictions and sanctions, trade embargoes, trade protection measures, import or export requirements, entity lists, tariffs, quotas, and other trade barriers and restrictions, including those related to the ongoing trade disputes between China and the United States;

•fluctuations in currency exchange rates;

•higher levels of credit risk and payment fraud;

•enhanced difficulties of integrating any foreign acquisitions;

•burdens of complying with a variety of foreign laws;

•reduced protection for intellectual property rights in some countries and practical difficulties in enforcing intellectual property and other rights outside the United States;

•difficulties in staffing and managing global operations and the increased travel, infrastructure and legal compliance costs associated with multiple international locations and subsidiaries;

•new and different sources of competition;

•different regulations and practices with respect to employee/employer relationships, existence of workers’ councils and labor unions, and other challenges caused by distance, language, and cultural differences, making it harder to do business in certain international jurisdictions;

•compliance with statutory equity requirements; and

•management of tax consequences.

If we are unable to manage the complexity of our global operations successfully, our results of operations and financial condition may be harmed.

Our business and ongoing expansion depends on attracting and retaining qualified personnel, especially our engineering and technical personnel.

Our success depends in large part on our ability to attract and retain highly qualified personnel. We strive to employ talented engineering and technical personnel, as well as effective sales, marketing, finance, and support employees, all of which are in high demand. Maintaining our brand and reputation, as well as a diverse and inclusive work environment that enables all our employees to thrive, is important to our ability to recruit and retain employees. There is intense competition for highly qualified technologists in the AI industry, particularly in the San Francisco Bay Area, where our headquarters are located, as well as in Toronto, Canada and in Bangalore, India, where we also have offices. We use various measures, including offering competitive salaries and benefits and an equity incentive program, to attract and retain the personnel we require to operate and grow our business effectively. However, these measures may not be sufficient, and our employees may decide not to continue working for us and leave us with little or no notice. Many of our competitors are significantly larger than we are, with greater financial resources and publicly traded stock, and may be able to offer more attractive compensation packages than we can, particularly with regard to equity compensation. As a result, it may be difficult for us to continue to retain and motivate these employees, and the value of their holdings could affect their decisions about whether or not they continue to work for us. Our ability to attract, retain and motivate employees may be adversely affected by declines in the price of our Class A common stock. If we issue significant equity to attract employees or to retain our existing employees, we would incur substantial additional stock-based compensation expense and the ownership of our existing stockholders would be further diluted. In addition, some of our employees are employed with us on temporary work visas, and any change in U.S. or Canadian immigration laws affecting their status or such visas may make it difficult to retain such individuals or hire new personnel. The loss of even a few qualified employees, or an inability to attract, retain and motivate additional highly skilled employees could harm our ability to develop and sell our products, as well as our results of operations, and impair our ability to expand and grow our business.

As our company grows and evolves, we may need to implement more complex organizational management structures, adapt our corporate culture and work environments, streamline our organization, or adjust the size and structure of our workforce to scale for the future and execute our long-term growth plan. If we fail to attract new personnel, or to retain and motivate our current personnel, our business, financial condition, results of operations, and prospects could be harmed.

The loss of one or more members of our senior management team could harm our business.

We currently depend on the continued services and contributions of our senior management team, particularly the services of our co-founders, including our Chief Executive Officer and President, Andrew Feldman. The members of our senior management team and co-founders are at-will employees, which means that each person could resign or could be terminated for any reason at any time. Members of our senior management team, and in particular Mr. Feldman, are critical to the management of our company and instrumental in the development of our technology and our strategic direction, and should one or more of such persons stop working for us for any reason, it is unlikely that we would be able to immediately find a suitable replacement. We also do not maintain any key person life insurance policies. The loss of the services of any of them, other members of senior management, or members of our senior technology personnel, could disrupt our operations, hamper our ability to implement our business strategy, and harm our business, financial condition, results of operations, and prospects.

We may seek to expand our business through the acquisition of complementary businesses or technologies. Any future acquisitions and investments may disrupt our business and harm our financial condition and results of operations.

We may seek to expand our business and the products we offer, or our employee base through the acquisition of complementary businesses and technologies. We also may enter into relationships with other businesses to expand our platform, which could involve preferred or exclusive licenses, additional channels of distribution, discount pricing, or investments in other companies. Negotiating these transactions can be time-consuming, difficult, and expensive, and our ability to close these transactions may be subject to approvals that are beyond our control, including approvals related to foreign investments and acquisitions. In addition, we have not previously acquired another business. Even if we are able to identify a suitable acquisition or investment target, we may not be able to complete the acquisition on commercially reasonable terms or at all. We may expend a significant amount of time and incur significant out-of-pocket costs, as well as divert management attention, in connection with an acquisition that is not ultimately completed. If an acquired business fails to meet our expectations, our business, financial condition, results of operations, and prospects may be harmed.

In the event we do consummate an acquisition, we face a number of risks as a result, including:

•diversion of management time and focus;

•coordination of research and development and sales and marketing functions;

•retention and motivation of key employees from the acquired company;

•cultural challenges associated with integrating employees from the acquired company;

•integration of the acquired company’s accounting, management information, human resources, and other administrative systems and processes;

•difficulty achieving the anticipated synergies of the transaction;

•liability for activities of the acquired company before the acquisition, including intellectual property infringement, violation, or misappropriation claims, violations of laws, commercial disputes, tax liabilities and other known and unknown liabilities; and

•litigation or other claims in connection with the acquired company, including claims from terminated employees, users, former stockholders or other third parties.

Our failure to address these risks or other problems encountered in connection with acquisitions could cause us to fail to realize the anticipated benefits of these acquisitions, cause us to incur unanticipated liabilities and harm our business generally. Future acquisitions could also result in dilutive issuances of our equity securities, the incurrence

of substantial amounts of debt, which could have terms that impose significant restrictions on our business, contingent liabilities, amortization expenses, incremental operating expenses, or the impairment of goodwill, any of which may harm our business, financial condition, results of operations, and prospects.

Our business is subject to the risks of earthquakes, fire, power outages, floods, and other natural disasters and catastrophic events, and to interruption by man-made problems such as war and terrorism.

A significant natural disaster or other catastrophic event, such as an earthquake, fire, flood, power outage, telecommunications failure, cyber-attack, war, terrorist attack, sabotage, other intentional acts of vandalism or misconduct, geopolitical event, pandemic, or other public health crisis, or other catastrophic occurrence may harm our business, financial condition, results of operations, and prospects. Our principal corporate offices and a significant number of our employees, as well as data centers in which we have deployed our products, are located in the San Francisco Bay Area, a region known for seismic activity. Geopolitical changes between China and Taiwan could also disrupt the operations of our Taiwan-based third-party wafer foundry and other suppliers, negatively impact delivery of products, and harm our business, financial condition, results of operations, and prospects. Furthermore, escalation of geopolitical tensions, including as a result of escalations in the ongoing conflict between Russia and Ukraine, the conflict in the Middle East, could have a broader impact that expands into other markets where we do business, which may harm our business, vendors, partners, customers, or the economy as a whole. Despite any precautions we may take, the occurrence of a natural disaster or other unanticipated problems could result in lengthy interruptions in our services or disruptions in our activities or the activities of our suppliers, manufacturers, partners, customers, or the economy as a whole. All of the aforementioned risks may be further increased if our disaster recovery plans prove to be inadequate. We do not carry business interruption insurance sufficient to compensate us for the potentially significant losses, including the potential harm to our business that may result from interruptions in our ability to provide our products and services. Any such natural disaster or man-made problem may harm our business, financial condition, results of operations, and prospects.

Risks Related to IT Systems, Cybersecurity, and Intellectual Property

Our business is dependent upon the proper functioning of our internal business processes and IT systems and modification or interruption of such systems may disrupt our business, processes, and internal controls.

We rely upon a number of internal business processes and IT systems to support key business functions, including our supply chain and inventory management systems, and the efficient operation of these processes and systems is critical to our business. Our business processes and IT systems need to be sufficiently scalable to support the growth of our business and may require modifications or upgrades that expose us to a number of operational risks. As such, our IT systems will continually evolve and adapt in order to meet our business needs. These changes may be costly and disruptive to our operations and could impose substantial demands on management time. These changes may also require changes in our IT systems, modification of internal control procedures and significant training of employees and third parties, as well as other resources. We continuously work on simplifying our IT systems and applications through consolidation and standardization efforts. There can be no assurance that our business and operations will not experience any disruption in connection with this transition. Our IT systems, and those of our third-party IT providers or business partners, may also be vulnerable to damage or disruption caused by circumstances beyond our control including catastrophic events, power anomalies or outages, natural disasters, viruses or malware, cyber-attacks, insider threat attacks, unauthorized system or data modifications, data breaches and computer system or network failures, exposing us to significant cost, reputational harm and disruption or damage to our business. In addition, as our IT environment continues to evolve, we are embracing new ways of communicating and sharing data internally and externally with customers and partners using methods such as mobility and the cloud that can promote business efficiency. However, these practices can also result in a more distributed IT environment, making it more difficult for us to maintain visibility and control over internal and external users, and meet scalability and administrative requirements. If our security controls cannot keep pace with the speed of these changes or if we are not able to meet regulatory and compliance requirements, or if we are unable to address any of the concerns described above, our business financial condition, results of operations, and prospects may be harmed.

Any failure of our IT systems or those of one or more of our IT service providers, business partners, vendors, suppliers, or other third-party service providers, or any other failure by such third parties to provide services to us may negatively impact our relationships with customers and harm our business.

Our business depends on various IT systems and outsourced IT services. We rely on third-party IT service providers, business partners, vendors, suppliers, and cloud-based service providers to provide critical IT system, corporate infrastructure, and other services and are, by necessity, dependent on them to adequately address cybersecurity threats to, and other vulnerabilities, defects, or deficiencies of or in their own systems. This includes infrastructure such as electronic communications, finance, marketing, and recruiting platforms and services such as IT network development and network monitoring, and third-party data center hosting of our systems for our internal and customer use. We do not own or control the operation of the third-party facilities or equipment used to provide such services. Our third-party vendors and service providers have no obligation to renew their agreements with us on commercially reasonable terms or at all. If we are unable to renew these agreements on commercially reasonable terms, including with respect to service levels and cost, or at all, we may be required to transition to a new provider, and we may incur significant costs and possible service interruption in connection with doing so. In addition, such service providers could decide to close their facilities or change or suspend their service offerings without adequate notice to us. Moreover, any financial difficulties, such as bankruptcy, faced by such vendors or cloud-based service providers, the nature and extent of which are difficult to predict, may harm on our business. Since we cannot easily switch vendors and cloud-based service providers, any disruption with respect to our current providers would impact our operations and our business may be harmed. Furthermore, our disaster recovery systems and those of such third parties may not function as intended or may fail to adequately protect our business information in the event of a significant business interruption, Any termination, failure, or other disruption of any of such systems or services of our third-party IT providers, business partners, vendors, suppliers, and cloud-based service providers could lead to operating inefficiencies or disruptions, which could harm our business, financial condition, results of operations, and prospects.

Product, IT system security, network, and data protection breaches, as well as cyber-attacks, incidents, or other unauthorized access to, or disclosure or other processing of, our proprietary, confidential, or sensitive information, including personal information, may disrupt our operations, reduce our expected revenue, increase our expenses, and harm our business and reputation.

In the ordinary course of our business, we and our third-party providers collect, store, and otherwise process confidential and sensitive information, including personal information about individuals such as our employees and customers as well as proprietary business information and intellectual property. We also process training and inference data provided to us by our customers and users, which may include personal information. This data, including the personal information, is processed on our IT systems as well as those provided by certain third-party providers upon whom we rely for critical services such as cloud-based infrastructure, encryption and authentication technology, employee email and other functions.

We and our providers face various evolving cybersecurity risks that threaten the confidentiality, integrity, and availability of our IT systems and data, including sensitive, proprietary, and personal information, that we process or that is processed on our behalf. These risks include physical or electronic break-ins, security or cybersecurity breaches, incidents and disruptions, computer malware, social-engineering attacks/phishing, ransomware, denial-of-service attacks, employee theft, misuse, or other malfeasance by insiders, human or technological error (such as software bugs, server malfunctions, hardware, or software failures), loss of data or other IT assets, and other cyber-attacks by threat actors.

Individuals, groups of hackers, and sophisticated organizations, including nation-states and nation-state-supported actors, terrorists, criminals, competitors, and other threat actors, have engaged and are expected to continue to engage in cyber-attacks. The techniques employed in such attacks (such as the use of emerging AI technologies), which we may not recognize until launched against a target or which may be difficult to discover for an extended period, change frequently and are becoming increasingly sophisticated, making it more difficult to successfully detect, defend against them or implement adequate preventative measures. We may also experience cybersecurity breaches that may remain undetected for an extended period. Even if identified, we may be unable to

adequately investigate or remediate incidents or breaches due to attackers increasingly using tools and techniques that are designed to circumvent controls, to avoid detection, and to remove or obfuscate forensic evidence. Such cyber-attacks and other cybersecurity breaches, incidents, or disruptions may continue to evolve in frequency, sophistication, and volume, and may be difficult to detect for long periods of time. Any of the foregoing breaches, incidents, or disruptions may compromise our networks, IT systems, or applications, or the data collected or processed on such systems, causing interruptions, delays, loss, or other operational malfunctions, which in turn could harm our business, financial condition, results of operations, and prospects.

Certain aspects of effective cybersecurity are dependent upon our employees, contractors, or other third-party service providers safeguarding our sensitive information and adhering to our security policies and access control mechanisms, and we may face cybersecurity threats due to error or intentional misconduct by such employees, contractors, or other third-party service providers. Remote and hybrid working arrangements at our company (and at many third-party providers) also increase cybersecurity risks due to the challenges associated with managing remote computing assets and security vulnerabilities that are present in many non-corporate and home networks. Additionally, due to geopolitical conflicts and during times of war or other major conflicts, we and the third parties we rely upon may be vulnerable to a heightened risk of cyber-attacks that could materially disrupt our ability to provide services and products.

Like many other companies, we and our third party providers may experience security incidents, cyber-attacks, or other unauthorized access to, or disclosure or other processing of, our proprietary, confidential or sensitive information, including arising from a failure to properly handle IT systems and the data that is processed through them, including personal information, or to adhere to our security policies and access control mechanisms and, although no such events have had a material adverse effect on our business to date, there can be no assurance that we will not have a materially adverse incident in the future. To defend against security incidents, we must continuously engineer more secure products and enhance security and reliability features. We must also continue to develop our security measures, including training programs and security awareness initiatives, as well as vendor management processes, designed to ensure that we and our suppliers have appropriate security measures in place, and continue to meet the evolving security requirements of our customers, applicable industry standards, and government regulations. While we invest in training programs and security awareness initiatives and take steps to detect and remediate vulnerabilities, we may not always be able to prevent threats or detect and mitigate all vulnerabilities in our security controls, systems, or software, including third-party software we have installed, as such threats and techniques change frequently and may not be detected until after a security incident has occurred.

Further, we cannot guarantee that third parties and infrastructure in our supply chain or our partners’ supply chains have not been compromised or that they do not contain exploitable vulnerabilities, defects, or bugs that could result in a breach of or disruption to our IT systems, including our products and services, or the third-party IT systems that support our services. We may also incorporate third-party data into our AI algorithms or use open-source datasets to train our algorithms. These datasets may be flawed, insufficient, or contain certain biased information, contain information (including intellectual property and confidential, proprietary, or personal information) for which the third party did not have appropriate rights, and may otherwise be vulnerable to security incidents, or negatively affect safety, security, and other functioning of our AI compute solutions. We may have limited insight into the data privacy or security practices of third-party suppliers, including with respect to our AI models. Our ability to monitor these third parties’ information security practices is limited, and they may not have adequate information security or legal compliance measures in place or sufficient rights in the underlying data to make them available. In addition, if one of our third-party suppliers suffers a security incident, our response may be limited or more difficult because we may not have direct access to their systems, logs and other information related to the security incident. Finally, we may experience delays in developing and deploying remedial measures designed to address identified vulnerabilities on our IT systems or those of our third-party providers. These vulnerabilities could, if exploited, result in a security incident.

Actual or perceived breaches of our security measures or unapproved access to our dissemination of proprietary, confidential, or sensitive information, including personal information, about or by us or third parties, could expose us and the parties affected to a risk of loss, or misuse of this information, potentially resulting in litigation (including class actions) and subsequent liability, regulatory inquiries or actions including potential penalties and fines,

additional reporting requirements or other oversight, restrictions on processing data, indemnification obligations, being required to provide credit monitoring or identity-theft prevention services, diversion of funds, diversion of management attention, financial loss, loss of data, material disruptions in our systems and operations, supply chain, and ability to produce, sell and distribute our products and services, damage to our brand and reputation or erosion confidence in the effectiveness of our security measures, or significant incident response, system restoration or remediation, and future compliance costs or other harms, which may harm our business, financial condition, results of operations, and prospects. Applicable data privacy and security obligations may also require us to notify relevant stakeholders, including affected individuals, customers, regulators, and investors, of security incidents, and investigations into and mandatory disclosures with respect to such incidents could be costly and lead to negative publicity.

While our insurance policies include liability coverage for certain of these matters, subject to retention amounts that could be substantial, if we experience a significant security breach, incident or disruption, we could be subject to liability or other damages that exceed our insurance coverage and we cannot be certain that such insurance policies will continue to be available to us on economically reasonable terms, or at all, or that any insurer will not deny coverage as to any future claim. The successful assertion of one or more large claims against us that exceed available insurance coverage, or the occurrence of changes in our insurance policies, including premium increases or the imposition of large deductible or co-insurance requirements, could adversely affect our business, financial condition, results of operations, and prospects.

Failure to obtain, maintain, protect, or enforce our intellectual property rights could harm our brand, business, and results of operations.

We regard the protection of our intellectual property as critical to our success. We strive to protect our intellectual property rights by relying on a combination of patent, trademark, trade secret, unfair competition and other related laws in the United States and internationally as well as confidentiality procedures and contractual provisions to protect and establish our rights in our intellectual property, including our proprietary technologies and know-how. We spend significant resources to monitor and protect our intellectual property rights, including monitoring the unauthorized use of our products, but even with significant expenditures, we may not be able to protect the intellectual property rights that are valuable to our business. In particular, we are unable to predict or assure that:

•our intellectual property rights will not lapse or be invalidated, circumvented, challenged, or, in the case of third-party intellectual property rights licensed to us, be licensed to others;

•our intellectual property rights will provide competitive advantages to us;

•rights previously granted by third parties to intellectual property licensed or assigned to us, including portfolio cross-licenses, will not hamper our ability to assert our intellectual property rights or hinder the settlement of currently pending or future disputes;

•any of our pending or future patent, copyright, or trademark applications will be issued or have the coverage originally sought;

•we will be able to enforce our intellectual property rights in certain jurisdictions where competition is intense or where legal protection may be weak; or

•we have sufficient intellectual property rights to protect or continue to offer our products or services or operate our business.

We pursue the registration of our patents, trademarks, service marks, and domain names in the United States and in certain foreign jurisdictions. We cannot guarantee that any pending or future patent applications will be issued to have the coverage originally sought, and even if the pending patent applications are granted, the rights granted to us may not be meaningful or provide us with any commercial advantage. Additionally, our patents could

be opposed, contested, narrowed, circumvented, challenged, abandoned, or designed around by our competitors or be declared invalid or unenforceable in judicial or administrative proceedings. The patent prosecution process is expensive, time-consuming, and complex, and we have not in the past, and may not in the future be able to file, prosecute, maintain, enforce, or license all necessary or desirable patent applications at a reasonable cost or in a timely manner. Since we may not have sufficient resources or capital to pursue patent registration for our patentable technology, our competitors could gain a competitive advantage if we fail to adequately protect such technologies as trade secrets such that a competitor could develop similar technologies and pursue and obtain patent registration for those technologies that would exclude us from continuing to use the patented technology, even if the technology was initially proprietary to us. It is also possible that we will fail to identify patentable aspects of our research and development output in time to obtain patent protection. Failure to timely seek patent protection on products or technologies generally precludes us from seeking future patent protection on these products or technologies. Even if we do timely seek patent protection, the coverage claimed in a patent application can be significantly reduced before a patent is issued, and its scope can be reinterpreted after issuance, and as a result we can give no assurance that any patents that we have issued or may have issued in the future will protect all significant aspects or components of our current and future products or services, will provide us with any competitive advantage, or will not be challenged, invalidated or circumvented in the future. Furthermore, we may not be able to obtain or maintain patent applications and patents due to the subject matter claimed in such patent applications and patents being in disclosures in the public domain. In addition, when patents expire, we lose the protection and competitive advantages they originally provided to us.

Additionally, we believe that our success also depends on the technical expertise we have developed in designing, testing, and manufacturing products, and we rely on confidential and proprietary information to develop and maintain our competitive position. As a result, we also typically enter into confidentiality and invention assignment agreements with our employees, contractors, and business partners in order to limit access to, and disclosure and use of, our proprietary information. However, we cannot guarantee we have entered into such agreements with each party that has or may have had access to our trade secrets, confidential or proprietary information, or technology, including our AI products and solutions. Even if entered into, these contractual arrangements and the other steps we have taken to protect our intellectual property may not prevent the misappropriation, infringement, violation, dilution, or disclosure of our confidential or proprietary information or technology, including our AI solutions, trade secrets, or intellectual property rights, or deter independent development of similar or competing technologies by others.

Obtaining and maintaining effective intellectual property rights, including the costs of defending our rights is expensive. We have obtained a number of issued patents, and are seeking additional patent protection, and to register our trademarks and domain names in the United States and in certain foreign jurisdictions. These processes are expensive and may not be successful in all jurisdictions or for every such application, and we may not pursue such protections in all jurisdictions that may be relevant. Further, effective intellectual property protections may not be available in every country in which we offer our products or services, and even where present, the laws of such countries may not recognize intellectual property rights or protect them to the same extent as their equivalents in the United States. Additionally, any changes in, or unexpected interpretations of, intellectual property laws may compromise our ability to enforce our trade secret and intellectual property rights. Any of the foregoing could make it difficult for us to stop the infringement, misappropriation, dilution or other violation of our intellectual property or marketing of competing products or services, and any failure to obtain or maintain adequate protection of our trade secrets or other intellectual property rights may harm our competitive position and may harm our business, financial condition, results of operations, and prospects.

Litigation may be necessary to enforce our intellectual property rights, protect our proprietary rights, or determine the validity and scope of proprietary rights claimed by others. Any actual or threatened litigation of this nature, regardless of outcome or merit, could result in substantial costs and diversion of management and technical resources, any of which may harm our business, financial condition, results of operations, and prospects. In addition, we believe that the protection of our trademark rights is an important factor in product recognition, protecting our brand and maintaining goodwill and if we do not adequately protect our rights in our trademarks from infringement, any goodwill that we have developed in those trademarks could be lost or impaired, which may harm our brand and our business. There may be potential trade name or trademark infringement claims brought by owners of other

trademarks that are similar to our trademarks, and as a result, we may incur significant costs in enforcing our trademarks against those who attempt to imitate the “Cerebras” brand and other valuable trademarks and service marks. If we fail to maintain, protect, and enhance our intellectual property rights, our brand, business, and results of operations may be harmed.

Further, we may acquire companies with intellectual property that is subject to certain licensing obligations or restrictions. These licensing obligations may extend to our own intellectual property following any such potential acquisition and may limit our ability to assert, protect, or otherwise use our intellectual property rights. From time to time, we may pursue litigation to assert our intellectual property rights, including, in some cases, against our customers and suppliers, where we believe they have infringed, misappropriated, or otherwise violated any of our intellectual property rights. Conversely, third parties have and may in the future pursue intellectual property litigation against us. Claims of any of the foregoing could also harm our relationships with our manufacturers and customers and might deter future manufacturers and customers from doing business with us. Furthermore, an adverse decision in any such legal action may result in material expense and limit our ability to assert our intellectual property rights and limit the value of our products and services, which may otherwise harm our business, financial condition, results of operations, and prospects.

Our ability to design and introduce new products in a timely manner includes the use of certain third-party intellectual property.

In the design and development of new and enhanced products and services, including AI computing solutions, we rely on certain third-party intellectual property, such as development and testing tools for certain hardware and software. Further, we have, on occasion, leveraged third parties for software development. Furthermore, certain of our product features may rely on intellectual property acquired from third parties that incorporate into our hardware or software. The design requirements necessary to meet customer demand for more features and greater functionality from semiconductor products may exceed the capabilities of the third-party intellectual property or development or testing tools available to us. If the third-party intellectual property that we use becomes unavailable, is not available with required functionality or performance in the time frame or price point needed for our new products or fails to produce designs that meet customer demands, or laws are adopted that affect our use of third party intellectual property in certain regions or products, our business, financial condition, results of operations, and prospects may be harmed.

We may face claims of intellectual property infringement, misappropriation, dilution, or other violations, which could be time-consuming or costly to defend or settle, result in the loss of significant rights or harm our relationships with our customers or reputation in the industry.

Third parties have in the past, and may in the future, assert against us their patent and other intellectual property rights to technologies or information that are used in or are important to our business, which may be time consuming and costly to defend or settle. We have in the past, and may in the future, particularly as a public company with an increased profile and visibility, receive communications from others alleging our infringement, misappropriation, dilution, or other violation of intellectual property rights. In addition, in the event that we recruit employees or contractors from other companies, including certain potential competitors, and these employees or contractors are involved in the development of products that are similar to the products they assisted in developing for their former employers, we may become subject to claims that such employees or contractors have used or disclosed trade secrets or other proprietary information in an unauthorized manner. We may also in the future be subject to claims by our third-party suppliers, employees, or contractors asserting an ownership right in our issued patents, pending patent applications or other intellectual property, including our AI solutions, as a result of the work they performed on our behalf, or claims for indemnification by our customers who are subject to infringement or other claims by third parties.

While we do not license for profit, sell access to, or otherwise derive revenue directly from the use of AI models, we have trained AI models on publicly available datasets, similar to many other developers of AI models, and released certain of such models to the community under certain open-source licenses. We also provide AI model services to our customers, where we leverage our expertise to help customers train their models with architectures,

parameter sizes, and data sets and types of their choosing, which may include publicly available or proprietary data sets or a combination of both. The act of such training necessarily involves transmission and use of certain data on our systems. Like other developers of AI models who are subject to litigation and other disputes arising from the training, fine-tuning, use, or development of AI models, we too may face lawsuits alleging that we reproduced, copied, displayed, distributed, or made derivative works of, or otherwise misused copyrighted materials to train our or our customers’ AI models without the authorization of the relevant copyright owners, or otherwise infringed third-party proprietary rights in training data, including rights of publicity. However, this remains an unsettled area of U.S. law and U.S. courts are currently weighing a number of lawsuits involving claims that the reproduction of data for training AI models, or the use of AI models trained on copyrighted data, infringes the rights of copyright holders. In addition, we have and may continue to fine-tune certain third-party AI models. While we believe we are in compliance with the applicable license terms of such models, the interpretation of such licenses may vary and have not been tested in litigation, and we may be subject to claims that we have violated the terms of such licenses.

Claims that our products or processes infringe, misappropriate, dilute, or otherwise violate third-party intellectual property rights, regardless of their merit or resolution, could be time-consuming or costly to defend or settle and could divert the efforts and attention of our management and technical personnel. Infringement claims also could harm our relationships with our customers and might deter future customers from doing business with us. We do not know whether we would prevail in these proceedings given the complex technical issues and inherent uncertainties in intellectual property litigation. If any pending or future proceedings result in an adverse outcome, we could be required to:

•cease the manufacture, use, or sale of the infringing products or processes;

•pay substantial damages for infringement, misappropriation, dilution, or other violation, including enhanced damages for any willful infringement;

•expend significant resources to develop non-infringing products or processes, which may not be successful;

•license certain components or data from the third-party claiming infringement, which license may not be available on commercially reasonable terms, or at all;

•cross-license our products to a competitor to resolve an infringement claim, which could weaken our ability to compete with that competitor; or

•pay substantial damages to our customers or end-users to discontinue their use of or to replace infringing product or process sold to them with non-infringing products or processes, if available.

Additionally, even if successful in any such proceedings, our rights in our products or services and other intellectual property may be invalidated, encumbered, narrowed, or otherwise diminished. Moreover, there could be public announcements of the results of hearings, motions or other interim proceedings or developments, and if securities analysts or investors perceive these results to be negative, it could have a substantial adverse effect on the price of our Class A common stock. Any of the foregoing results may harm our business, financial condition, results of operations, and prospects. Litigation against our customers as a result of third-party claims of intellectual property infringement could trigger indemnification obligations under some of our agreements, which could result in substantial expense to us, and which may materially harm our business, financial condition, results of operations, and prospects.

Certain of our intellectual property has been and may be developed under research agreements with U.S. government entities, and may be subject to federal regulations that limit our exclusive rights in certain circumstances.

Certain of our intellectual property that generally pertain to applications outside of our offerings in the AI computing market has been and may be developed under contracts with U.S. government entities. As a result, the U.S. government may have certain rights to intellectual property that we use in our current or future products

pursuant to the Bayh-Dole Act of 1980, as amended (the “Bayh-Dole Act”) or as otherwise required by our contractual arrangements. Under the Bayh-Dole Act, U.S. government rights in certain “subject inventions” developed under such contracts include a nonexclusive, non-transferable, and irrevocable worldwide license to use inventions for any governmental purpose. In addition, the U.S. government has the right to require us, or an assignee or exclusive licensee to such inventions, to grant licenses to these inventions to the U.S. government or a third party if the U.S. government determines that: (i) adequate steps have not been taken to commercialize the invention; (ii) government action is necessary to meet public health or safety needs; (iii) government action is necessary to meet requirements for public use under federal regulations; or (iv) the right to use or sell such inventions is exclusively licensed to an entity within the United States and substantially manufactured outside the United States without the U.S. government’s prior approval. We may lose exclusivity to our intellectual property rights if we fail to comply with reporting obligations regarding subject inventions, fail to file for patent protection within specified time limits, or fail to comply with other relevant Bayh-Dole Act restrictions. If any of our intellectual property becomes subject to the rights or remedies available to the U.S. government or third parties pursuant to the Bayh-Dole Act or related contractual arrangements, the value of our intellectual property may be impaired and our business may be harmed.

Our use of third-party open-source software may pose risks to our proprietary software and services in a manner that may harm our business.

Certain of our software, as well as that of our vendors or partners, may use or be derived from “open-source” software that is generally made available to the public by its authors or other third parties. Some open-source software licenses require end-users, who use, distribute or make available across a network software and services that include open-source software, to make publicly available or to license at no cost all or part of such software (which in some circumstances may include valuable proprietary code, such derivative works of the open-source software) under the terms of the particular open-source license. These obligations may require us to make source code for the derivative works available to the public or license such derivative works under a particular type of license rather than the more limited access rights we customarily grant our customers and their users. This type of licensing may subject us to disclosure of valuable, proprietary software code.

While our policies and processes are intended to enable us to monitor and comply with the licenses of third-party open-source software and protect our valuable proprietary source code, we may inadvertently use third-party open-source software in a manner that exposes us not only to the risk of a forced disclosure of our own proprietary software, but also to claims of non-compliance with the terms of third-party licenses, including claims of infringement or for breach of contract. We cannot be sure that all open-source software is identified, reviewed, or submitted for approval prior to use in our operations or platform. Also, there exists today an increasing number of types of open-source software licenses, and those licenses may not yet have faced legal challenges in courts that could result in guidance to users in their efforts to avoid legal issues. If we were to receive a claim of non-compliance with the terms of any of these licenses, not only would the potential exposure of our own source code be very harmful to us, but we may be required to invest substantial time and resources to re-engineer some of our software or license alternative software on terms unfavorable to us. Any of the foregoing may disrupt and harm our intellectual property, business, financial condition, results of operations, and prospects.

Additionally, the use of certain open-source software can lead to greater risks than use of third-party commercial software, as open-source licensors generally do not provide warranties or controls on the functionality or origins of software or other contractual protections regarding infringement claims or the quality of the licensed code, including with respect to security and architectural vulnerabilities. There is typically no support available for open-source software and such software is ordinarily provided on an “as-is” basis, and we cannot be sure that the authors of such open-source software will implement or push updates to address security risks or will not abandon further development and maintenance. Many of the risks associated with the use of open-source software, such as the lack of warranties or assurances of title or performance, cannot be eliminated, and could, if not properly addressed, negatively affect our business. Use of open-source software may also present additional security risks because the public availability of such software may make it easier for hackers and other third parties to determine how to compromise our services. Further, our use of any AI solutions that use or incorporate any open-source

software may heighten any of the foregoing risks. Any of these risks could be difficult to eliminate or manage, and, if not addressed, may harm our business, financial condition, results of operations, and prospects.

Risks Related to Legal and Regulatory Matters

Sales to government entities and highly regulated organizations are subject to a number of challenges and risks.

We have sold in the past, and may sell in the future, products to governmental agencies or entities and customers in highly regulated industries, such as healthcare and financial services. Selling to such entities can be highly competitive, expensive, and time consuming, often requiring significant upfront time and expense without any assurance that these efforts will generate a sale. In addition, government demand and payment for our products are affected by changes in administration, public sector budgetary cycles and funding authorizations, and government contracting requirements may change from time to time, any of which can limit our ability to sell into the government sector. In certain foreign jurisdictions, our ability to win business may be constrained by political or other factors unrelated to our competitive position in the market.

Further, government and other highly regulated entities can have more complex IT and data environments, and often have longer implementation or deployment cycles than others. They have and may continue to demand contract terms that differ from our standard arrangements and may be less favorable than terms agreed with other private sector customers, and may expect greater payment flexibility. Government contracts may contain provisions that give the government substantial rights and remedies, many of which are not typically found in commercial contracts, including provisions relating to intellectual property “march-in” rights, preferential pricing, refund rights, obligation modifications, U.S. manufacturing requirements, export control, and termination or non-renewal due to funding availability.

Government contracting requirements may change and in doing so restrict our ability to sell into the government sector until we have obtained any required government certifications. Further, to contract with certain government agencies, some of our employees may be required to have security clearances. Obtaining and maintaining such security clearances is a lengthy process. If our employees are unable to obtain or maintain such clearances, or we cannot recruit employees with such clearances, it would harm our ability to sell to, or work with, such government agencies, which may harm our business, results of operations, and financial condition. Government demand and payment for our products and services are affected by public sector budgetary cycles and funding authorizations, with funding reductions or delays adversely affecting public sector demand for our solutions.

As a government contractor or subcontractor, we must comply with laws, regulations, and contractual provisions relating to the formation, administration, and performance of government contracts, all of which may impose additional costs on our business. Governments routinely investigate and audit government contractors’ administrative processes, and any unfavorable audit could result in the government refusing to continue purchasing our products and services. In addition, as a result of actual or perceived noncompliance with government contracting laws, regulations, or contractual provisions, we may be subject to non-ordinary course audits and internal investigations which may prove costly to our business financially, divert management attention or limit our ability to continue selling our products and services to our government customers. Failure to comply with these or other applicable regulations and requirements could lead to claims for damages, downward contract price adjustments or refund obligations, civil or criminal penalties, and termination of contracts and suspension or debarment from government contracting for a period of time with government agencies. Any such damages, penalties, disruption of, or limitation in our ability to do business with a government would harm our reputation, business, financial condition, results of operations, and prospects.

Our global operations expose us to numerous legal and regulatory requirements and failure to comply with such requirements, including unexpected changes to such requirements, may harm our results of operations.

We service our customers around the world. We are subject to numerous, and sometimes conflicting, legal regimes of the United States and foreign national, state, and provincial authorities on matters as diverse as anti-corruption, trade restrictions, tariffs, taxation, sanctions, anti-competition, intellectual property, data security, and

privacy. U.S. laws may be different in significant respects from the laws of countries where we operate or we or our customers may enter, forcing businesses to choose between compliance with conflicting legal regimes. We also may seek to expand operations in emerging market jurisdictions where legal systems are less developed or familiar to us.

In addition, there can be no assurance that the laws or administrative practices relating to taxation (including the current position as to income and withholding taxes), foreign exchange, export controls, economic sanctions, or otherwise in the jurisdictions where we have operations will not change. Changes in tax laws in some jurisdictions may also have a retroactive effect and we may be found to have paid less tax than required in such jurisdictions. Compliance with diverse legal requirements is costly, time consuming and requires significant resources. Violations of one or more of these regulations in the conduct of our business could result in significant fines, criminal sanctions against us or our officers, prohibitions on doing business and damage to our reputation. Violations of these regulations in connection with the performance of our obligations to our customers also could result in liability for significant monetary damages, fines or criminal prosecution, unfavorable publicity and other reputational damage, and allegations by our customers that we have not performed our contractual obligations. Due to the varying degrees of development of the legal systems of the countries in which we operate, local laws might be insufficient to protect our rights.

Our global operations and collection, storage, use and other processing of proprietary, confidential, and sensitive information, including personal information, expose us to numerous data privacy and security laws, regulations, contractual requirements, and other obligations relating to data privacy and security, and the actual or perceived failure to comply with such obligations, including unexpected changes to such obligations, may harm our business, financial condition, results of operations, and prospects.

The processing of personal information, including the personal information of our employees and customers, makes us, or may make us, subject to a complex patchwork of evolving data privacy and security laws that are not always interpreted uniformly. Additionally, we may be bound by contractual requirements applicable to our collection, storage, transmission, use and other processing of proprietary, confidential, and sensitive information, including personal information, and may be bound or asserted to be bound by, or voluntarily comply with, self-regulatory or other industry standards relating to the processing of such information. These laws, rules, regulations, industry standards, contractual requirements and other obligations are constantly evolving, and we expect that we will continue to become subject to new proposed laws, rules, regulations, industry standards, contractual requirements and other obligations in the United States and other jurisdictions where we operate. This evolution, among other things, may create uncertainty in our business; affect us or our collaborators’, service providers’ and contractors’ ability to operate in certain jurisdictions or to collect, store, transfer, use and share personal information; necessitate the acceptance of more onerous obligations in our contracts; result in liability; or impose additional costs on us; necessitate changes to our IT systems, and practices and to those of any third parties that process personal information on our behalf, or require us to change our business model. There is no guarantee that regulators or consumers will agree with our approach to compliance and any failure, or perceived failure, to comply with applicable data privacy or security laws or regulations may harm our business, financial condition, results of operations, and prospects.

In the United States, numerous state and federal laws, regulations, standards, and other legal obligations, including consumer protection laws and regulations, which govern the collection, dissemination, use, access to, confidentiality, security, and other processing of personal information, including health-related information, apply to our operations or the operations of our customers, third-party service providers, or partners. For example, the Health Insurance Portability and Accountability Act of 1996, and regulations promulgated thereunder (“HIPAA”) imposes privacy, security, and breach notification obligations on certain healthcare providers, health plans, and healthcare clearinghouses, known as covered entities, as well as their business associates that perform certain services that involve creating, receiving, maintaining, or transmitting individually identifiable health information for or on behalf of such covered entities, and their covered subcontractors. Among other requirements, HIPAA requires business associates to develop and maintain policies with respect to the protection of, use and disclosure of protected health information (“PHI”), including the adoption of administrative, physical, and technical safeguards to protect such information, certain notification requirements in the event of a breach of unsecured PHI, and requirements to report breaches of unsecured PHI to covered entities within 60 days of discovery of the breach by the business associate or

its agents. Depending on the facts and circumstances, we could be subject to significant civil, criminal, and administrative fines and penalties and/or additional reporting and oversight obligations if found to be in violation of HIPAA.

We are also subject to the rules and regulations promulgated under the authority of the Federal Trade Commission (“FTC”), which, together with many state Attorneys General, has the authority to regulate and enforce against unfair or deceptive acts or practices in or affecting commerce, including acts and practices with respect to privacy, data protection and cybersecurity. For example, according to the FTC, failing to take appropriate steps to keep consumers’ personal information secure can constitute unfair acts or practices in or affecting commerce in violation of Section 5(a) of the Federal Trade Commission Act. The FTC expects a company’s data security measures to be reasonable and appropriate in light of the sensitivity and volume of consumer information it holds, the size and complexity of its business, and the cost of available tools to improve security and reduce vulnerabilities. Congress also has considered, and continues to consider, many proposals for comprehensive national privacy, data protection, and cybersecurity legislation, including with respect to AI, to which we may become subject if enacted.

Certain U.S. states have also adopted comparable data privacy and security laws and regulations, which govern the privacy, processing, and protection of personal information. Such laws and regulations will be subject to interpretation by various courts and other governmental authorities, thus creating potentially complex compliance issues for us and our future customers and strategic partners. For example, the California Consumer Privacy Act, as amended by the California Privacy Rights Act (collectively, the “CCPA”), provides for enhanced privacy rights for California residents and requires covered businesses that process the personal information of California residents to, among other things: (i) provide certain disclosures to California residents regarding the business’s collection, use, and disclosure of their personal information; (ii) receive and respond to requests from California residents to access, delete, and correct their personal information, or to opt out of certain disclosures of their personal information; and (iii) enter into specific contractual provisions with service providers that process California resident personal information on the business’s behalf. The CCPA is enforced by the California Attorney General and the California Privacy Protection Agency (“CPPA”), and provides for civil penalties for certain violations, as well as a private right of action for certain data breaches that may increase the likelihood of and risks associated with data breach litigation. In addition, numerous other states have enacted, or are in the process of enacting or considering, comprehensive state-level privacy, data protection and cybersecurity laws, rules and regulations that share similarities with the CCPA, which creates the potential for a patchwork of overlapping but different domestic privacy laws. In addition, all 50 states have laws that require the provision of notification for breaches of personal information to affected individuals, state officers or others. Non-compliance with HIPAA, FTC rules and regulations, the CCPA, or other U.S. privacy laws may result in enforcement actions, litigation, or other disputes and expose us to additional liability, which could harm our business, financial condition, results of operations, and prospects.

We are also required or may be required to comply with foreign data privacy and security laws in jurisdictions in which we have offices or conduct business. For example, in Europe, the EU General Data Protection Regulation and applicable national supplementing laws (collectively, the “GDPR”) impose strict requirements for processing the personal data of individuals within the European Economic Area (“EEA”) or for activities within the EEA. Following the withdrawal of the UK from the EU, we may also be subject to the UK General Data Protection Regulations and Data Protection Act 2018 (collectively, the “UK GDPR”). The GDPR and UK GDPR are wide-ranging in scope and impose numerous additional requirements on companies that process personal data, including imposing special requirements in respect of the processing of personal data, requiring that consent of individuals to whom the personal data relates is obtained in certain circumstances, requiring additional disclosures to individuals regarding information processing activities, requiring that safeguards are implemented to protect the security and confidentiality of personal data, creating mandatory data breach notification requirements in certain circumstances and requiring that certain measures (including contractual requirements) are put in place when engaging third-party processors. The GDPR and UK GDPR also provide individuals with various rights in respect of their personal data, including rights of access, erasure, portability, rectification, restriction, and objection. Failure to comply with the GDPR and the UK GDPR can result in significant fines and other liability. European data protection authorities have shown a willingness to impose significant fines and issue orders preventing the processing of personal information on non-compliant businesses and have imposed fines for GDPR violations up to, in some cases, hundreds of millions of Euros. While the UK GDPR currently imposes substantially the same obligations as the GDPR, any changes to

the GDPR going forward need to be specifically incorporated by the UK government. Moreover, the UK government has announced plans to reform the UK GDPR in ways that, if formalized, are likely to deviate from the GDPR, all of which creates a risk of divergent parallel regimes and related uncertainty, along with the potential for increased compliance costs and risks for affected businesses.

Among other requirements, the GDPR regulates transfers of personal data subject to the GDPR or UK GDPR to so-called third countries outside the EEA and the UK that have not been determined by the relevant data protection authorities to provide an adequate level of protection to such personal data, including the United States, and the efficacy and longevity of current transfer mechanisms between the EEA, and the United States remains uncertain. Case law from the Court of Justice of the EU (“CJEU”) indicates that reliance on the standard contractual clauses—a standard form of contract approved by the European Commission as an adequate personal data transfer mechanism—alone may not necessarily be sufficient in all circumstances and that transfers must be assessed on a case-by-case basis. In July 2023, the European Commission adopted an adequacy decision in relation to the new EU-U.S. Data Privacy Framework (“DPF”) rendering the DPF effective as a GDPR transfer mechanism for personal data transferred from the EEA to the U.S. by U.S. entities self-certified under the DPF. In October 2023, the UK Extension to the DPF came into effect, as approved by the UK government, as a data transfer mechanism from the UK to U.S. entities self-certified under the DPF. However, the DPF adequacy decisions do not foreclose, and are likely to face, future legal challenges and the ongoing legal uncertainty with respect to international data transfers may increase our costs and our ability to efficiently process personal data from the EEA or the UK. In addition to the ongoing legal uncertainty with respect to data transfers from the EEA or the UK, additional costs may need to be incurred in order to implement necessary safeguards to comply with the GDPR and the UK GDPR, and potential new rules and restrictions on the flow of data across borders could increase the cost and complexity of conducting business in some markets. If our policies and practices or those of our third-party vendors, service providers, contractors or consultants are, or are perceived to be, insufficient, or if our customers or others have concerns regarding our transfer of personal data from the EEA or the UK to the United States, we could be subject to enforcement actions or investigations, including by individual EU or UK data protection authorities, or lawsuits by private parties. Other jurisdictions outside the EU and the UK are similarly introducing or enhancing privacy, data protection and cybersecurity laws, rules, and regulations, which could increase our compliance costs and the risks associated with noncompliance. We cannot yet fully determine the impact these or future laws, rules, and regulations may have on our business or operations. These laws, rules and regulations may be inconsistent from one jurisdiction to another, subject to differing interpretations and may be interpreted to conflict with our practices.

While we have implemented various measures to help ensure that our policies, processes, and systems are in compliance with our legal obligations with respect to our collection, storage, use and other processing of proprietary, confidential and sensitive information, including personal information, any inability, or perceived inability, to adequately address privacy concerns or comply with applicable laws, even if unfounded, may result in significant regulatory and third-party liability, increased costs, disruption of our business and operations, and a loss of client confidence and other reputational damage. Furthermore, as new privacy-related laws and regulations are implemented, the time and resources needed for us to seek compliance with such laws and regulations continues to increase.

The AI industry is subject to complex, evolving regulatory, statutory, and other requirements that may be difficult and expensive to comply with and that could negatively impact our business.

The regulatory framework for our products and services, including our AI computing solutions and AI model services, is rapidly evolving as many federal, state, and foreign government bodies and agencies have introduced or are currently considering additional laws and regulations related to AI. Additionally, existing laws and regulations may be interpreted in ways that would affect our or our customers’ operations, or our ability to offer our AI products and services in the markets in which we operate. As a result, implementation standards and enforcement practices are likely to remain uncertain for the foreseeable future, and we cannot yet determine the impact future laws, regulations, standards, or market perception of their requirements may have on our business, and we may not be able to adequately anticipate or respond to these evolving laws or regulations. In addition, because AI-related technologies are themselves highly complex and rapidly developing, it is not possible to predict all of the legal or regulatory risks that may arise relating to our use of such technologies. New laws, guidance or decisions in this area

could provide a new regulatory framework that may require us to adjust and make changes to our operations that may decrease our operational efficiency, resulting in an increase to operating costs and/or hindering our ability to improve our products and services.

Already, certain existing legal regimes (including, those related to data privacy and cybersecurity) regulate certain aspects of AI models and automated decision-making, and new laws regulating AI technologies are expected to enter into force in the United States and the EU in 2024. For example, in the United States, the 2023 AI Order sets out principles intended to guide the design, development, deployment and use of AI and signals the increase in U.S. governmental involvement and regulation over AI. The 2023 AI Order also instructed several other agencies to promulgate additional regulations within specific timeframes from the date of the 2023 AI Order regarding the use and development of AI. Already agencies such as the Department of Commerce, the FTC, and the Department of Health and Human Services have issued proposed rules governing the use and development of AI. Legislation related to AI has also been introduced at the federal level and is advancing at the state level. For example, Utah recently passed the AI Policy Act, which took effect in May 2024, imposing certain disclosure requirements on the use of AI, and Colorado enacted the Colorado AI Act, which will take effect in February 2026. In addition, the CPPA is currently in the process of finalizing regulations under the CCPA regarding the use of automated decision-making, and the California state legislature is considering several bills related to AI that would impose additional requirements on users, developers, and providers of AI systems. Any of such regulations, or any similar regulations, may impact the development, use, and commercialization of AI in the future.

Further, in Europe, the AI Act, which establishes a comprehensive, risk-based governance framework for AI in the EU market applies to, amongst other entities, providers, importers, and distributors of AI systems or general-purpose AI models that are placed on the EU market or put into service or used in the EU. The AI Act entered into force in August 2024, with the majority of the AI Act’s substantive requirements coming into effect in 2026. The AI Act establishes a risk-based governance framework for regulating high-risk AI systems and categorizes AI systems based on the risks associated with such AI systems’ intended purposes as creating “unacceptable,” “high,” or “limited” risks. The AI Act also includes various requirements for providers, importers, distributors, and users of AI systems, including with respect to transparency, conformity assessments and monitoring, risk assessments, human oversight, security and accuracy, general-purpose AI models, and foundation models, and introduces significant fines for noncompliance. While the AI Act has yet to be enforced, there is a risk that our current or future products and services may be subject to heightened obligations under the AI Act (including, for example, by being categorized as “high risk” AI systems under the AI Act), requiring us to comply with the applicable requirements of the AI Act, which may impose additional costs on us, increase our risk of liability or adversely affect our business. Even if our products and services are not categorized as “unacceptable” or “high” risk, we may be subject to additional transparency and other obligations for providers, distributors, or importers of AI systems, which may require us to expend resources to comply with such obligations. In addition, in September 2022, the European Commission proposed two Directives seeking to establish a harmonized civil liability regime for the use of AI technologies in the EU (“AI Liability Directives”). There are also specific obligations regarding the use of automated decision-making under the GDPR. The AI Act, AI Liability Directives, and GDPR may have a material impact on the way AI is regulated, and developing interpretation and applications of the foregoing, together with developing guidance and/or decisions in this area, may affect our planned business activities involving the development and/or use of AI. Additionally, our customers may become subject to such upcoming AI regulations, which could cause a delay or impediment to the commercialization of AI technologies and could lead to a decrease in demand for our customers’ AI systems.

It is likely that further new laws and regulations will be adopted in the United States and in other non-U.S. jurisdictions, or that existing laws and regulations, including competition and antitrust laws, may be interpreted in ways that would limit ours or our customers’ ability to use AI or in a manner that negatively affects the performance of our products, services, and business. We may need to expend resources to adjust our AI products or services in certain jurisdictions if the laws, regulations, or decisions are not consistent across jurisdictions. Further, the cost to comply with such laws, regulations, or decisions and/or guidance interpreting existing laws could be significant and would increase our operating expenses (such as by imposing additional reporting obligations). Such an increase in operating expenses, as well as any actual or perceived failure to comply with such laws and regulations, could adversely affect our business, financial condition, results of operations, and prospects. The regulatory environment

surrounding the implementation of AI technologies may adversely affect our ability to produce and export our products and services and as a result may cause harm to our reputation, business, financial condition, results of operations, and prospects.

We may be subject to litigation, investigations, or other actions, which may lead us to incur significant costs and harm our business and our stockholders.

We are, and may become, party to lawsuits and claims arising in the normal course of business, which may include putative class action suits or other lawsuits, investigations or other claims relating to intellectual property, open-source software, customer matters, our marketing and sales practices, contracts, employment matters, regulatory compliance, or other aspects of our business.

Many companies in the semiconductor industry own large numbers of patents, copyrights, trademarks, domain names, and trade secrets, and frequently enter into litigation based on allegations of infringement, misappropriation, or other violations of intellectual property or other rights. As we face increasing competition and gain a higher profile, the possibility of intellectual property rights claims against us grows.

Defending any lawsuit, even when comprised of unmeritorious claims, is costly and can impose a significant burden on, and divert the attention of, management and employees, and harm our reputation. As litigation is inherently unpredictable, we cannot assure you that any potential claims or disputes will not harm our business, financial condition, results of operations, and prospects. Any claims or litigation, even if fully indemnified or insured, may make it more difficult to effectively compete or to obtain adequate insurance in the future. Any litigation to which we are a party may result in an onerous or unfavorable judgment that may not be reversed upon appeal, or in the payment of substantial monetary damages or fines, or we may decide to settle lawsuits on similarly unfavorable terms, which may harm our business, financial condition, results of operations, and prospects. In the case of an unfavorable outcome in intellectual property case, we could also be required to:

•pay substantial damages for past, present, and future use of the infringing technology;

•cease use of an infringing product (or component), which may involve redesigning a product or component part so that it does not infringe;

•expend significant resources to develop non-infringing technology;

•license technology from the third-party claiming infringement, which license may not be available on commercially reasonable terms, or at all;

•enter into cross-licenses with our competitors, which could weaken our overall intellectual property portfolio and our ability to compete in particular product categories;

•indemnify our customers;

•pay substantial damages to our direct or end customers to discontinue use or replace infringing technology with non-infringing technology; or

•relinquish intellectual property rights associated with one or more of our patent claims if such claims are held invalid or otherwise unenforceable.

Any of the foregoing results may harm our business, financial condition, results of operations, and prospects.

We may be subject to warranty claims and product liability.

From time to time, we may be subject to warranty or product liability claims arising from defects or perceived defects in our products or in third-party components that we integrate into our products, which may lead to

significant expenses. If a customer’s equipment fails in use, the customer may incur significant expenses, as well as lost revenue. The customer may claim that a defect in our product caused the equipment failure and assert a claim against us to recover monetary damages, including indirect and consequential damages. The process of identifying a defective or potentially defective product in complex systems may be lengthy and require significant resources, and we may incur significant replacement costs and contract damage claims from our customers. In certain situations, we may consider incurring the costs or expenses related to a recall of one of our products in order to avoid the potential claims that may be raised should customer suffer a failure due to a design or manufacturing process defect. Any such liabilities may greatly exceed any revenue we receive from the relevant products. Costs, payments, or damages incurred or paid by us in connection with warranty and product liability claims could exceed our product liability insurance coverage, or warranty reserves, and could harm our business, financial condition, results of operations, and prospects.

Regulations related to conflict minerals may cause us to incur additional expenses and may limit the supply and increase the costs of certain metals used in the manufacturing of our products.

We are subject to requirements under the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, requiring us to conduct due diligence on and disclose whether certain conflict minerals originating from certain countries and geographic regions are necessary for the manufacture or functionality of our products. The implementation of these requirements could adversely affect the sourcing, availability, and pricing of the materials used in the manufacture of components used in our products. In addition, we will incur additional costs to comply with the potential disclosure requirements, including costs related to conducting diligence procedures to determine the sources of minerals that may be used or necessary to the production of our products and, if applicable, potential changes to products, processes, or sources of supply as a consequence of such due diligence activities. It is also possible that we may face reputational harm if we determine that any of our products contain minerals not determined to be free of conflict minerals or if we are unable to alter our products, processes, or sources of supply to avoid such materials.

Risks Related to Financial and Accounting Matters

We identified material weaknesses in our internal control over financial reporting. If we are unable to remediate these material weaknesses, or if we identify additional material weaknesses in the future or otherwise fail to maintain an effective system of internal controls, we may not be able to accurately or timely report our financial condition or results of operations, which may adversely affect investor confidence in us and, as a result, the value of our Class A common stock.

Prior to the completion of this offering, we have been a private company since our inception and, as such, we have not had the internal control and financial reporting requirements that are required of a publicly traded company. In connection with the preparation of our financial statements, we identified certain material weaknesses in our internal control over financial reporting for the years ended December 31, 2023 and 2022. A material weakness is a deficiency, or combination of deficiencies, in internal control over financial reporting such that there is a reasonable possibility that a material misstatement of our financial statements will not be prevented or detected on a timely basis.

The material weaknesses that we identified relate to (i) inadequate or missing resources who possess an appropriate level of expertise to timely review account reconciliations and identify, select, and apply U.S. generally accepted accounting principles (“GAAP”) pertaining to several financial statement areas, including revenue recognition, inventory, and equity administration and (ii) the failure to maintain adequate IT general controls, including ineffective segregation of duties.

In response to the identified material weaknesses, we have begun adding additional resources, formalizing processes, and implementing new controls. We have hired additional accounting and finance personnel with expertise we believe to be appropriate to strengthen our overall controls over the review of account reconciliations, the application of GAAP, and the IT environment. We intend to continue to take steps to remediate these material weaknesses. The material weaknesses will not be considered remediated until management designs and implements

effective controls that operate for a sufficient period of time and management has concluded, through testing, that these controls are effective. Our management will monitor the effectiveness of our remediation plans and will make changes determined to be appropriate.

We can give no assurance that the measures we have taken and plan to take in the future will remediate the material weaknesses identified or that any additional material weaknesses or restatements of financial results will not arise in the future due to a failure to implement and maintain adequate internal control over financial reporting or circumvention of these controls. If the steps we take do not correct these material weaknesses in a timely manner, we will be unable to conclude that we maintain effective internal control over financial reporting. Accordingly, there could continue to be a reasonable possibility that a material misstatement of our financial statements would not be prevented or detected on a timely basis. We have limited experience with implementing the systems and controls that will be necessary to operate as a public company. If these new systems or controls and the associated process changes do not give rise to the benefits that we expect or do not operate as intended, it may harm our financial reporting systems and processes, our ability to produce timely and accurate financial reports or the effectiveness of internal control over financial reporting.

If we fail to remediate our existing material weaknesses or identify new material weaknesses in our internal control over financial reporting, if we are unable to comply with the disclosure and attestation requirements of Section 404 of the Sarbanes-Oxley Act in a timely manner, if we are unable to conclude that our internal control over financial reporting is effective, or if our independent registered public accounting firm is unable to conclude that our internal control over financial reporting is effective when we are no longer an emerging growth company, investors may lose confidence in the accuracy and completeness of our financial reports and the price of our Class A common stock could be negatively affected. As a result, we could also become subject to investigations by the Nasdaq Stock Market LLC (“Nasdaq”), the SEC, or other regulatory authorities, and become subject to litigation from stockholders, which could harm our reputation and financial condition or divert financial and management resources from our regular business activities.

In addition, even if we are successful in strengthening our controls and procedures, in the future those controls and procedures may not be adequate to prevent or identify irregularities or errors or to facilitate the fair presentation of our financial statements.

If we fail to maintain an effective system of disclosure controls and internal control over financial reporting, our ability to produce timely and accurate financial statements or comply with applicable regulations may be harmed.

As a public company, we will be subject to the reporting requirements of the Exchange Act of 1934, as amended (the “Exchange Act”), the Sarbanes-Oxley Act, and the stock exchange listing requirements. We expect that the requirements of these rules and regulations will continue to increase our legal, accounting, and financial compliance costs, make some activities more difficult, time-consuming, and costly, and place significant strain on our personnel, systems, and resources.

The Sarbanes-Oxley Act requires, among other things, that we maintain effective disclosure controls and procedures and internal control over financial reporting. We are continuing to develop and refine our disclosure controls and other procedures that are designed to ensure that information required to be disclosed by us in the reports that we will file with the SEC is recorded, processed, summarized, and reported within the time periods specified in SEC rules and forms and that information required to be disclosed in reports under the Exchange Act is accumulated and communicated to our principal executive and financial officers. We are also continuing to improve our internal control over financial reporting, which includes hiring additional accounting and financial personnel to implement such processes and controls.

We have identified material weaknesses in our internal control over financial reporting for the years ended December 31, 2023 and 2022 and cannot assure you that there will not be material weaknesses or significant deficiencies in our internal controls in the future. In order to maintain and improve the effectiveness of our disclosure controls and procedures and internal control over financial reporting, we have expended, and anticipate that we will continue to expend, significant resources, including accounting-related costs, new internal processes and

procedures, and significant management oversight. If any of these new or improved controls and systems do not perform as expected, we may experience further deficiencies in our controls.

Our current controls and any new controls that we develop may become inadequate because of changes in conditions in our business. Further, to the extent we acquire other businesses, the acquired company may not have a sufficiently robust system of controls and we may discover deficiencies. Any failure to develop or maintain effective controls or any difficulties encountered in their implementation or improvement may harm our results of operations or cause us to fail to meet our reporting obligations and may result in a restatement of our financial statements for prior periods. Any failure to implement and maintain effective internal control over financial reporting also may adversely affect the results of periodic management evaluations and annual independent registered public accounting firm attestation reports regarding the effectiveness of our internal control over financial reporting that we will eventually be required to include in our periodic reports that will be filed with the SEC. Ineffective disclosure controls and procedures and internal control over financial reporting could also cause investors to lose confidence in our reported financial and other information, which would likely cause the price of our Class A common stock to decline. In addition, if we are unable to continue to meet these requirements, we may not be able to remain listed on a stock exchange. We are not currently required to comply with the SEC rules that implement Section 404 of the Sarbanes-Oxley Act and are therefore not required to make a formal assessment of the effectiveness of our internal control over financial reporting for that purpose. As a public company, we will be required to provide an annual management report on the effectiveness of our internal control over financial reporting commencing with our second annual report on Form 10-K.

Upon becoming a public company, and particularly after we are no longer an “emerging growth company,” we expect our independent registered public accounting firm will be required to formally attest to the effectiveness of our internal control over financial reporting. At such time, our independent registered public accounting firm may issue a report that is adverse in the event it is not satisfied with the level at which our internal control over financial reporting is documented, designed, or operating. Any failure to maintain effective disclosure controls and internal control over financial reporting may harm our business, financial condition, results of operations, and prospects, and may cause the price of our Class A common stock to decline.

We may have a limited ability to use some or all of our net operating loss carryforwards in the future.

As of December 31, 2023, we had net operating loss (“NOL”) carryforwards for federal and state income tax purposes of $333.8 million and $395.5 million, respectively. Our NOL carryforward for federal purposes has an indefinite carryforward period but is limited to offset 80% of taxable income in the year utilized, except for $20.6 million that will begin to expire in 2036. Our NOL carryforwards for state purposes have various carryover periods and will begin to expire as early as 2036. As a result of prior operating losses, we have significant NOL carryforwards for federal income tax purposes. Our ability to utilize our NOLs to reduce taxable income in future years could become subject to significant limitations under Section 382 of the U.S. Internal Revenue Code of 1986, as amended (the “Code”) if we undergo an “ownership change” within the meaning of Section 382, or the value of such NOLs could be reduced in the event that the relevant rules under the Code were to be revised. We would undergo an ownership change if, among other things, the stockholders who own, directly or indirectly, 5% or more of our common stock, or are otherwise treated as “5% shareholders” under Section 382 of the Code and the regulations promulgated thereunder, increase their aggregate percentage ownership of our stock by more than 50 percentage points over the lowest percentage of the stock owned by these stockholders at any time during the testing period, which is generally the three-year period preceding the potential ownership change. Similar rules may apply under state tax laws.

There is also a risk that due to regulatory changes, such as suspensions on the use of NOLs or other unforeseen reasons, our existing NOLs could expire or otherwise be unavailable to reduce future income tax liabilities, including for state tax purposes. For these reasons, we may not be able to utilize a material portion of the NOLs reflected on our balance sheet, even if we attain profitability, which may potentially result in increased future tax liability to us and may harm our results of operations and financial condition.

Unanticipated changes in our effective tax rate and additional tax liabilities may impact our results of operations.

We are subject to taxes in the United States and certain foreign jurisdictions. Due to economic and political conditions, tax rates in various jurisdictions, including the United States, may be subject to change. For example, the U.S. government may enact significant changes to the taxation of business entities, including, among others, a permanent increase in the corporate income tax rate, an increase in the tax applicable to the global low-taxed income and the imposition of minimum taxes or surtaxes on certain types of income. Our future effective tax rates could be affected by changes in the mix of earnings in countries with differing statutory tax rates, changes in the valuation of deferred tax assets and liabilities and changes in tax laws or their interpretation.

We may also be subject to additional tax liabilities and penalties due to changes in non-income based taxes resulting from changes in federal, state, or foreign tax laws, changes in taxing jurisdictions’ administrative interpretations, decisions, policies and positions, results of tax examinations, settlements or judicial decisions, changes in accounting principles, changes to the business operations, including acquisitions, as well as the evaluation of new information that results in a change to a tax position taken in a prior period. We are currently unable to predict whether such changes will occur and, if such changes occur, the ultimate impact on our tax liabilities. Any resulting increase in our tax obligation or cash taxes paid may harm our cash flows and results of operations.

We expect to require additional capital to support business growth, and this capital might not be available when needed on favorable terms or at all.

We intend to continue to make investments to support our business growth and may require additional funds to respond to business challenges and opportunities, including the need to develop new products or services, enhance our existing products or services, enhance our operating infrastructure, expand internationally, and acquire complementary businesses and technologies. In order to achieve these objectives, we expect to require additional capital resources in the future, and may determine to raise additional funds. If we raise additional funds through future issuances of equity or convertible debt securities, our existing stockholders could suffer significant dilution, and any new equity securities we issue could have rights, preferences, and privileges superior to those of holders of our Class A common stock. In addition, the incurrence of indebtedness would increase our fixed obligations, and may include covenants or other restrictions that impede our ability to manage our operations. We may not be able to obtain additional financing on terms favorable to us when needed, or at all. Our inability to obtain adequate financing or financing on terms satisfactory to us, when we require it, could significantly limit our ability to continue supporting our business growth and responding to business challenges and opportunities.

Some of our suppliers provide us with a line of credit to meet the needs of our normal business requirements. Most of our suppliers set dollar limits on the trade credit they will afford us at any given time. If our suppliers were to cease to sell to us on trade credit terms or were to substantially lower the credit limits they have set on our open accounts, we would need to accelerate our payments to those suppliers, creating additional demands on our cash resources, or we would need to find other sources for those goods. Further, some of our customers advance us funds pursuant to their purchase order, which we are required to hold in trust to be used only for the purposes specified in such purchase order. Under certain circumstances, we may be required to refund the portion of the advanced funds that has not yet been used for the specified purposes on demand, and we may not have enough available cash or be able to obtain financing at the time we are required to repay the portion of the advanced funds. The customer may also take title to the components purchased using the advanced funds. Additionally, we may have to pay taxes on the customer advances before we have recognized any revenue from the components purchased with the advanced funds. Our inability to repay the advanced funds when required, or the requirement to pay taxes prior to recognizing revenue may harm our business, financial condition, results of operations, and prospects.

Our results of operations may be harmed by changes in financial accounting standards or by the application of existing or future accounting standards to our business as it evolves.

Our reported results of operations are impacted by the accounting standards promulgated by the SEC and accounting standards bodies and the methods, estimates, and judgments that we use in applying our accounting

policies. A change in accounting standards may have a significant effect on our reported results of operations and may even affect the reporting of transactions completed before the announcement or effectiveness of a change. The frequency of accounting standards changes could accelerate, including conversion to unified international accounting standards. Any future changes to accounting standards may cause our results of operations to fluctuate.

As we enhance, expand, and diversify our business, products, and services, the application of existing or future financial accounting standards may harm our results of operations or financial condition.

Risks Related to this Offering and Ownership of Our Class A Common Stock

The price of our Class A common stock may be volatile and may decline regardless of our operating performance, and you may lose all or part of your investments.

The price of our Class A common stock may fluctuate significantly in response to numerous factors, many of which are beyond our control, including:

•overall performance of the equity markets and/or publicly listed semiconductor companies.

•actual or anticipated fluctuations in our financial and operating metrics;

•an adverse development in our strategic partnership with G42 or a material reduction in purchases by G42, or the anticipation of such events;

•changes in the financial projections we provide to the public or our failure to meet these projections;

•failure of securities analysts to initiate or maintain coverage of us, changes in financial estimates by any securities analysts who follow our company, or our failure to meet the estimates or the expectations of investors or analysts;

•the economy as a whole and market conditions in our industry;

•rumors and market speculation, and operating results and forecasts, involving us or other companies in our industry;

•announcements by us or our competitors of significant innovations, acquisitions, strategic partnerships, joint ventures, or capital commitments;

•new laws or regulations or new interpretations of existing laws or regulations applicable to our business;

•lawsuits threatened or filed against us;

•recruitment or departure of key personnel;

•changes in the U.S. regulatory environment impacting jurisdictions with which we can transact;

•other events or factors, including those resulting from war, incidents of terrorism, or responses to these events;

•conversions of shares of non-voting Class N common stock into shares of Class A common stock; and

•anticipated sales of our common stock, including pursuant to the “Early Release” provision in lock-up or market standoff agreements described elsewhere in this prospectus, or upon the expiration of such lock-up or market standoff agreements or in connection with the G42 Primary Purchase or the G42 Option.

In addition, extreme price and volume fluctuations in the stock markets have affected and continue to affect many semiconductor and technology companies’ stock prices. Often, their stock prices have fluctuated in ways unrelated or disproportionate to the companies’ operating performance. In the past, stockholders have filed securities class action litigation following periods of market volatility. If we were to become involved in securities litigation, it could subject us to substantial costs, divert resources and the attention of management from our business and harm our business. Moreover, because of these fluctuations, comparing our results of operations on a period-to-period basis may not be meaningful. You should not rely on our past results as an indication of our future performance. This variability and unpredictability could also result in our failing to meet the expectations of industry or financial analysts or investors for any period. If our net revenue or results of operations fall below the expectations of analysts or investors or below any forecasts we may provide to the market, or if the forecasts we provide to the market are below the expectations of analysts or investors, the price of our Class A common stock could decline substantially. Such a stock price decline could occur even when we have met any previously publicly stated net revenue or earnings forecasts that we may provide.

No public market for our common stock currently exists and an active liquid market may not develop or be sustained following this offering.

No public market for our common stock currently exists. An active public trading market for our Class A common stock may not develop following the closing of this offering or, if developed, it may not be sustained. The lack of an active market may impair your ability to sell your shares at the time you wish to sell them or at a price that you consider reasonable. The lack of an active market may also reduce the fair value of your shares. An inactive market may also impair our ability to raise capital to continue to fund operations by selling securities and may impair our ability to acquire other companies or technologies by using our securities as consideration.

Future sales of our Class A common stock in the public market could cause the price of our common stock to decline.

Sales of a substantial number of shares of our Class A common stock in the public market, particularly sales by our directors, executive officers, and principal stockholders, or the perception that these sales might occur, could cause the price of our Class A common stock to decline and could impair our ability to raise capital through the sale of additional equity securities.

We and all of our directors and executive officers, the selling stockholders, and the holders of substantially all of our shares of Class A common stock outstanding and securities exercisable for or convertible into shares of our Class A common stock, have entered into lock-up agreements with the underwriters and/or agreements with market standoff provisions that restrict our and their ability to sell or transfer shares of our capital stock and securities convertible into or exercisable or exchangeable for shares of our capital stock, for a period ending on the earlier of (i) 9:30 a.m. Eastern Time on the second trading day immediately following our public release of earnings for the year ending December 31, 2024 and (ii) 180 days from the date of this prospectus (the “Lock-up Period”). Such agreements provide for early release of a portion of our stockholder’s securities if certain stock-price conditions are satisfied within a certain time period, and are generally subject to certain exceptions. For example, certain exceptions to the lock-up agreements and market standoff agreements provide that certain shares of our Class A common stock will be eligible for sale in the open market during the Lock-up Period in sell-to-cover transactions in order to satisfy tax withholding obligations in connection with the settlement of RSUs. We expect RSUs to vest and settle during the Lock-up Period. Assuming a % tax withholding rate, we would expect approximately shares of Class A common stock to be eligible for sale in the open market in connection with the satisfaction of such tax withholding obligations. The number of shares actually sold in sell-to-cover transactions will depend on the tax withholding rate applicable to the relevant stockholder. It is also possible that we may decide to net settle all or a portion of such RSUs instead. In addition, Citigroup Global Markets Inc. and Barclays Capital Inc., on behalf of the underwriters, may release any of the securities subject to these lock-up agreements and market standoff agreements at any time, subject to the applicable notice requirements. See the sections titled “Shares Eligible for Future Sale” and “Underwriting” for a discussion of the Early Release and other exceptions that may allow for sales or transfers during the Lock-up Period. If not earlier released, all of the securities subject to lock-up or market standoff agreements will become eligible for sale upon expiration of the Lock-up

Period. Shares held by directors, executive officers, and other affiliates will be subject to volume limitations under Rule 144 under the Securities Act of 1933, as amended (the “Securities Act”).

In addition, as of June 30, 2024, after giving effect to the Option Exercise and the RSU Net Settlement, we had options outstanding that, if fully exercised, would result in the issuance of shares of our Class A common stock and shares of our Class A common stock issuable upon vesting of outstanding RSUs. We intend to file one or more registration statements on Form S-8 under the Securities Act to register the shares of our Class A common stock subject to outstanding stock options and RSUs, as of the date of this prospectus and shares that will be issuable pursuant to future awards granted under our equity incentive plans. Once we register these shares, they can be freely sold in the public market upon issuance, subject to applicable vesting requirements, compliance by affiliates with Rule 144, and other restrictions provided under the terms of the applicable plan and/or the award agreements entered into with participants.

Following this offering, the holders of up to shares of our Class A common stock will have rights, subject to some conditions, to require us to file registration statements for the public resale of shares of our Class A common stock or to include such shares in registration statements that we may file for us or other stockholders. The 22,851,296 shares of our Class N common stock reserved for future purchase pursuant to the G42 Primary Purchase and any shares of Class N common stock that may be issued pursuant to the G42 Option will have identical registration rights. Any registration statement we file to register additional shares, whether as a result of registration rights or otherwise, could cause the price of our Class A common stock to decline or be volatile.

If you purchase shares of our Class A common stock in this offering, you will incur immediate and substantial dilution.

The initial public offering price will be substantially higher than the pro forma net tangible book value per share of our Class A common stock immediately following this offering based on the total value of our tangible assets less our total liabilities. Therefore, if you purchased our Class A common stock in this offering, at the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, you would experience an immediate dilution of $ per share, the difference between the price per share you pay for our Class A common stock and our pro forma net tangible book value per share as of June 30, 2024, after giving effect to the issuance by us of shares of our Class A common stock in this offering, the Preferred Stock Conversion, the Option Exercise, and the RSU Net Settlement. In addition, you may also experience additional dilution if options, RSUs, stock appreciation rights (“SARs”), or other rights to purchase our common stock that are outstanding or that we may issue in the future are exercised, vest, or are converted or we issue additional shares of our common stock at prices lower than our net tangible book value at such time.

Future securities issuances could result in significant dilution to our stockholders, may subject us to regulatory and other uncertainty, and impair the price of our Class A common stock.

Future issuances of shares of our common stock, or the perception that these sales may occur, could depress the price of our Class A common stock and result in dilution to existing holders of our Class A common stock. To the extent outstanding options are exercised or RSUs vest and settle, or additional options, RSUs, SARs, or other stock-based awards are granted, there will be further dilution. The amount of dilution could be substantial depending upon the size of the issuances or exercises. Furthermore, we may issue additional equity securities that could have rights senior to those of our Class A common stock. As a result, purchasers of our Class A common stock bear the risk that future issuances of debt or equity securities may reduce the value of our Class A common stock and further dilute their ownership interest.

For example, under the Preferred Stock Purchase Agreement, we expect to issue an aggregate of 22,851,296 shares of our Class N common stock to G42 in connection with the G42 Primary Purchase and may issue additional shares of our Class N common stock to G42 pursuant to the G42 Option. See the section titled “Certain Relationships and Related Party Transactions” for additional information. In addition, if the Committee on Foreign Investment in the United States (“CFIUS”) indicates that any of these purchases of securities under the Preferred Stock Purchase Agreement with G42 are subject to its jurisdiction, and in the course of reviewing the investment,

CFIUS or the President of the United States takes any action that prevents implementation of the transactions contemplated by the Preferred Stock Purchase Agreement, we and G42 have agreed to use reasonable best efforts to agree in good faith on a suitable economic solution that reflects the economic structure (i.e., pay-ins and pay-outs) to which G42 would have otherwise been entitled had G42 been able to purchase shares. The parties have not discussed, or agreed upon, what form such a solution might take and may not agree on that solution. The additional dilution from these future issuances and the regulatory and other uncertainty applicable to the transactions may cause the price of our Class A common stock to decline.

We have broad discretion in the use of the net proceeds from this offering and may not use them effectively.

We will have broad discretion in the application of the net proceeds to us from this offering, including for any of the purposes described in the section titled “Use of Proceeds,” and you will not have the opportunity as part of your investment decision to assess whether the net proceeds are being used appropriately. Because of the number and variability of factors that will determine our use of the net proceeds from this offering, our ultimate use may vary substantially from our currently intended use. Investors will need to rely on the judgment of our management with respect to the use of proceeds. Pending use, we may invest the net proceeds from this offering in short-term, investment-grade, interest-bearing securities, such as money market funds, corporate notes and bonds, certificates of deposit, commercial paper, and guaranteed obligations of the U.S. government that may not generate a high yield for our stockholders. If we do not use the net proceeds that we receive in this offering effectively, our business, financial condition, results of operations, and prospects could be harmed, and the price of our Class A common stock could decline.

We will incur increased costs as a result of operating as a public company, and our management will be required to devote substantial time to support compliance with our public company responsibilities and corporate governance practices.

As a public company, we will incur significant finance, legal, accounting, and other expenses, including director and officer liability insurance, that we did not incur as a private company, and which we expect to further increase after we are no longer an “emerging growth company.” The Sarbanes-Oxley Act, the Dodd-Frank Wall Street Reform and Consumer Protection Act, stock exchange listing requirements, and other applicable securities rules and regulations impose various requirements on public companies in the United States. Our management and other personnel are expected to devote a substantial amount of time to support compliance with these requirements. Moreover, these rules and regulations will increase our legal and financial compliance costs and will make some activities more time-consuming and costly. We cannot predict or estimate the amount of additional costs we will incur as a public company or the specific timing of such costs.

We currently have no plans to pay dividends on our common stock.

We have never declared or paid any cash dividends on shares of our capital stock. We currently intend to retain all available funds and any future earnings for use in the operation of our business and do not anticipate paying any dividends in the foreseeable future. Any determination to pay dividends in the future will be at the discretion of our board of directors and will depend on many factors, including our financial condition, results of operations, earnings, capital requirements, business expansion opportunities, level of indebtedness, statutory and contractual restrictions applying to the payment of dividends, and other considerations that our board of directors deem relevant.

If securities or industry analysts do not publish research or publish inaccurate or unfavorable research about our business, our stock price and trading volume could decline.

The trading market for our Class A common stock will be influenced by the research and reports that industry or financial analysts publish about us or our business. We do not control these analysts or the content and opinions included in their reports. As a new public company, we may be slow to attract research coverage, and the analysts who publish information about our Class A common stock will have had relatively little experience with our company, which could affect their ability to accurately forecast our results and make it more likely that we fail to meet their estimates. In the event we obtain industry or financial analyst coverage, if any of the analysts who cover

us issues an inaccurate or unfavorable opinion regarding our stock price, our stock price may decline. In addition, the stock prices of many companies in the technology industry have declined significantly after those companies have failed to meet, or exceed, the financial guidance publicly announced by the companies or the expectations of analysts. If our results of operations fail to meet, or exceed, our announced guidance or the expectations of analysts or public investors, analysts could downgrade our Class A common stock or publish unfavorable research about us, and the price of our Class A common stock would likely decline as a result of such failure to meet our guidance or analyst expectations. If one or more of these analysts cease coverage of our Class A common stock or fail to publish reports on us regularly, our visibility in the financial markets could decrease, which in turn could cause our stock price or trading volume to decline.

For as long as we are an emerging growth company, we will not be required to comply with certain reporting requirements, including those relating to accounting standards and disclosure about our executive compensation, that apply to other public companies.

We are an “emerging growth company,” as defined in the JOBS Act, and we may take advantage of certain exemptions from various reporting requirements that are applicable to other public companies that are not “emerging growth companies,” including the auditor attestation requirements of Section 404, reduced disclosure obligations regarding executive compensation in our periodic reports and proxy statements, and exemptions from the requirements of holding a nonbinding advisory vote on executive compensation and stockholder approval of any golden parachute payments not previously approved. Pursuant to Section 107 of the JOBS Act, as an emerging growth company, we have elected to use the extended transition period for complying with new or revised accounting standards until those standards would otherwise apply to private companies. As a result, our consolidated financial statements may not be comparable to the financial statements of issuers who are required to comply with the effective dates for new or revised accounting standards that are applicable to public companies, which may make our Class A common stock less attractive to investors. In addition, if we cease to be an emerging growth company, we will no longer be able to use the extended transition period for complying with new or revised accounting standards.

We will remain an emerging growth company until the first to occur of: (1) the last day of the year following the fifth anniversary of this offering; (2) the last day of the first year in which our annual gross revenue is $1.235 billion or more; (3) the date on which we have, during the previous rolling three-year period, issued more than $1.0 billion in non-convertible debt securities; and (4) the date we qualify as a “large accelerated filer,” with at least $700 million of equity securities held by non-affiliates.

We cannot predict if investors will find our Class A common stock less attractive if we choose to rely on these exemptions. For example, if we do not adopt a new or revised accounting standard, our future results of operations may not be as comparable to the results of operations of certain other companies in our industry that adopted such standards. If some investors find our Class A common stock less attractive as a result, there may be a less active trading market for our Class A common stock, and our stock price may be more volatile.

Anti-takeover provisions contained in our amended and restated certificate of incorporation and amended and restated bylaws, as well as provisions of Delaware law, could impair a takeover attempt.

Our amended and restated certificate of incorporation and amended and restated bylaws, which will become effective immediately prior to the completion of this offering, contain, and Delaware law contains, provisions which could have the effect of rendering more difficult, delaying, or preventing an acquisition deemed undesirable by our board of directors. These provisions will provide for the following:

•a classified board of directors with staggered three-year terms, which may delay the ability of stockholders to change the membership of a majority of our board of directors;

•no cumulative voting in the election of directors, which limits the ability of minority stockholders to elect director candidates;

•the exclusive right of our board of directors to establish the size of the board of directors and to appoint a director to fill a vacancy, however occurring, including by expanding the board of directors;

•the ability of our board of directors to authorize the issuance of shares of preferred stock and to determine the price and other terms of those shares, including voting or other rights or preferences, without stockholder approval, which could be used to significantly dilute the ownership of a hostile acquiror;

•the ability of our board of directors to alter our amended and restated bylaws without obtaining stockholder approval;

•supermajority voting requirement to amend certain provisions in our amended and restated certificate of incorporation and amended and restated bylaws;

•a prohibition on stockholder action by written consent, which forces stockholder action to be taken at an annual or special meeting of our stockholders;

•the requirement that a special meeting of stockholders may be called only by our Chief Executive Officer or a majority of our board of directors then in office, which may delay the ability of our stockholders to force consideration of a proposal or to take action, including the removal of directors;

•advance notice procedures that stockholders must comply with in order to nominate candidates to our board of directors or to propose matters to be acted upon at a stockholders’ meeting, which may discourage or deter a potential acquiror from conducting a solicitation of proxies to elect the acquiror’s own slate of directors or otherwise attempting to obtain control of us; and

•the limitation of liability of, and provision of indemnification to, our directors and officers.

These provisions, alone or together, could delay or prevent hostile takeovers and changes in control or changes in our management.

As a Delaware corporation, we are also subject to provisions of Delaware law, including Section 203 of the General Corporation Law of the State of Delaware (the “Delaware General Corporation Law”), which prevents some stockholders holding more than 15% of our outstanding common stock from engaging in certain business combinations without approval of the holders of substantially all of our outstanding common stock. For a description of our capital stock, see the section titled “Description of Capital Stock.”

Any provision of our amended and restated certificate of incorporation, amended and restated bylaws, or Delaware law that has the effect of delaying or deterring a change in control could limit the opportunity for our stockholders to receive a premium for their shares of our Class A common stock, and could also affect the price that some investors are willing to pay for our Class A common stock.

Claims for indemnification by our directors and officers may reduce our available funds to satisfy successful third-party claims against us and may reduce the amount of money available to us.

Our amended and restated certificate of incorporation and amended and restated bylaws, which will become effective immediately prior to the completion of this offering, provide that we will indemnify our directors and officers, in each case to the fullest extent permitted by Delaware law.

In addition, as permitted by Section 145 of the Delaware General Corporation Law, our amended and restated bylaws to be effective immediately prior to the completion of this offering, and our indemnification agreements that we have entered or intend to enter into with our directors and officers provide that:

•we will indemnify our directors and officers for serving us in those capacities or for serving other business enterprises at our request to the fullest extent permitted by Delaware law. Delaware law provides that a

corporation may indemnify such person if such person acted in good faith and in a manner such person reasonably believed to be in or not opposed to the best interests of the registrant and, with respect to any criminal proceeding, had no reasonable cause to believe such person’s conduct was unlawful;

•we may, in our discretion, indemnify employees and agents in those circumstances where indemnification is permitted by applicable law;

•we are required to advance expenses, as incurred, to our directors and officers in connection with defending a proceeding, except that such directors or officers will undertake to repay such advances if it is ultimately determined that such person is not entitled to indemnification;

•the rights conferred in our amended and restated bylaws are not exclusive, and we are authorized to enter into indemnification agreements with our directors, officers, employees, and agents and to obtain insurance to indemnify such persons; and

•we may not retroactively amend our amended and restated bylaw provisions to reduce our indemnification obligations to directors, officers, employees, and agents.

While we maintain a directors’ and officers’ insurance policy, such insurance may not be adequate to cover all liabilities that we may incur, which may reduce our available funds to satisfy third-party claims and may harm our business and financial position.

Our amended and restated certificate of incorporation and amended and restated bylaws will provide for an exclusive forum in the Court of Chancery of the State of Delaware for certain disputes between us and our stockholders, and that the federal district courts of the United States will be the exclusive forum for the resolution of any complaint asserting a cause of action under the Securities Act.

Our amended and restated certificate of incorporation and amended and restated bylaws, which will become effective immediately prior to the completion of this offering, will provide, that: (i) unless we consent in writing to the selection of an alternative forum, the Court of Chancery of the State of Delaware (or, if such court does not have subject matter jurisdiction thereof, the federal district court of the State of Delaware) will, to the fullest extent permitted by law, be the sole and exclusive forum for: (A) any derivative action or proceeding brought on behalf of the company, (B) any action asserting a claim for or based on a breach of a fiduciary duty owed by any of our current or former directors, officers, other employees, agents, or stockholders to the company or our stockholders, including without limitation a claim alleging the aiding and abetting of such a breach of fiduciary duty, (C) any action asserting a claim against the company or any of our current or former directors, officers, employees, agents, or stockholders arising pursuant to any provision of the Delaware General Corporation Law or our certificate of incorporation or bylaws or as to which the Delaware General Corporation Law confers jurisdiction on the Court of Chancery of the State of Delaware, or (D) any action asserting a claim related to or involving the company that is governed by the internal affairs doctrine; (ii) unless we consent in writing to the selection of an alternative forum, the federal district courts of the United States will, to the fullest extent permitted by law, be the sole and exclusive forum for the resolution of any complaint asserting a cause of action arising under the Securities Act, and the rules and regulations promulgated thereunder; (iii) any person or entity purchasing or otherwise acquiring or holding any interest in shares of capital stock of the company will be deemed to have notice of and consented to these provisions; and (iv) failure to enforce the foregoing provisions would cause us irreparable harm, and we will be entitled to equitable relief, including injunctive relief and specific performance, to enforce the foregoing provisions. Nothing in our current certificate of incorporation or bylaws or our amended and restated certificate of incorporation or amended and restated bylaws precludes stockholders that assert claims under the Exchange Act, from bringing such claims in federal court to the extent that the Exchange Act confers exclusive federal jurisdiction over such claims, subject to applicable law.

The choice of forum provisions may limit a stockholder’s ability to bring a claim in a judicial forum that it finds favorable for disputes with us or any of our current or former directors, officers, other employees, agents, or

stockholders, which may discourage such claims against us or any of our current or former directors, officers, other employees, agents, or stockholders and result in increased costs for investors to bring a claim.

SPECIAL NOTE REGARDING FORWARD-LOOKING STATEMENTS

This prospectus contains forward-looking statements about us and our industry that involve substantial risks and uncertainties. All statements other than statements of historical facts contained in this prospectus, including statements regarding our strategy, future financial condition, future operations, projected costs, prospects, plans, objectives of management and expected market growth, are forward-looking statements. In some cases, you can identify forward-looking statements because they contain words such as “may,” “will,” “shall,” “should,” “expects,” “plans,” “anticipates,” “could,” “intends,” “target,” “projects,” “contemplates,” “believes,” “estimates,” “predicts,” “potential,” “goal,” “objective,” “seeks,” or “continue,” or the negative of these words or other similar terms or expressions that concern our expectations, strategy, plans, or intentions. Forward-looking statements contained in this prospectus include, but are not limited to, statements about:

•our future financial performance, including our expectations regarding our revenue, expenses, and other results of operations;

•our ability to acquire new customers and grow our customer base;

•our ability to successfully retain existing customers, including G42 and other significant customers, and expand sales within our existing customer base;

•our expectations with respect to the performance of our products;

•our ability to successfully maintain our relationships with our third-party suppliers and manufacturers;

•launching new products and adding new product capabilities;

•future investments in developing and enhancing our business;

•our expectations regarding our ability to expand;

•design, manufacturing, or product defects;

•our ability to effectively manage our growth;

•future investments in our business, our anticipated capital expenditures, and our estimates regarding our capital requirements;

•the estimated size of our addressable market opportunity;

•investments in our sales and marketing efforts;

•our ability to compete effectively with existing competitors and new market entrants;

•our reliance on our senior management team and our ability to identify, recruit, and retain skilled personnel;

•our ability to obtain, maintain, protect, and enforce our intellectual property rights and any costs associated therewith;

•our ability to comply with laws and regulations that currently apply or become applicable to our business both in the United States and internationally;

•economic and industry trends and other macroeconomic factors, such as fluctuating interest rates and rising inflation;

•the impact of geopolitical changes or tensions, political conflicts, and other global financial, economic, and political events and wars on our industry, business, financial condition, results of operations, and prospects and any global pandemics or health crises;

•the timing and completion of the G42 Primary Purchase;

•our expected use of proceeds from this offering; and

•other risks and uncertainties described in this prospectus, including those under the section titled “Risk Factors.”

We caution you that the foregoing list does not contain all of the forward-looking statements made in this prospectus.

You should not rely upon forward-looking statements as predictions of future events. We have based the forward-looking statements contained in this prospectus primarily on our current expectations, estimates, forecasts, and projections about future events and trends that we believe may affect our business, financial condition, results of operations, and prospects. Although we believe that we have a reasonable basis for each forward-looking statement contained in this prospectus, we cannot guarantee that the future results, levels of activity, performance, or events and circumstances reflected in the forward-looking statements will be achieved or occur at all. The outcome of the events described in these forward-looking statements is subject to risks, uncertainties and other factors described in the section titled “Risk Factors” and elsewhere in this prospectus. Moreover, we operate in a very competitive and rapidly changing environment. New risks and uncertainties emerge from time to time, and it is not possible for us to predict all risks and uncertainties that could have an impact on the forward-looking statements contained in this prospectus. The results, events, and circumstances reflected in the forward-looking statements may not be achieved or occur, and actual results, events or circumstances could differ materially from those described in the forward-looking statements.

The forward-looking statements made in this prospectus relate only to events as of the date on which the statements are made. We undertake no obligation to update any forward-looking statements made in this prospectus to reflect events or circumstances after the date of this prospectus or to reflect new information or the occurrence of unanticipated events, except as required by law. We may not actually achieve the plans, intentions, or expectations disclosed in our forward-looking statements, and you should not place undue reliance on our forward-looking statements. Our forward-looking statements do not reflect the potential impact of any future acquisitions, mergers, dispositions, joint ventures, or investments we may make.

In addition, statements that “we believe” and similar statements reflect our beliefs and opinions on the relevant subject. These statements are based upon information available to us as of the date of this prospectus, and while we believe such information forms a reasonable basis for such statements, such information may be limited or incomplete, and our statements should not be read to indicate that we have conducted an exhaustive inquiry into, or review of, all potentially available relevant information. These statements are inherently uncertain, and you are cautioned not to unduly rely upon these statements.

You should read this prospectus and the documents that we reference in this prospectus and have filed as exhibits to the registration statement, of which this prospectus is a part, completely and with the understanding that our actual future results may be materially different from what we expect. We qualify all of the forward-looking statements in this prospectus with these cautionary statements.

MARKET AND INDUSTRY DATA

This prospectus contains estimates, projections, and other information concerning our industry and our business, as well as data regarding market research, estimates, and forecasts prepared by our management. Information that is based on estimates, forecasts, projections, market research, or similar methodologies is inherently subject to uncertainties, and actual events or circumstances may differ materially from events and circumstances that are assumed in this information. The industry in which we operate is subject to a high degree of uncertainty and risk due to a variety of factors, including those described in the section titled “Risk Factors.” Unless otherwise expressly stated, we obtained this industry, business, market, and other data from reports, research surveys, studies, and similar data prepared by market research firms and other third parties, industry and general publications, government data, and similar sources. Forecasts and other forward-looking information with respect to industry, business, market, and other data are subject to the same qualifications and additional uncertainties regarding the other forward-looking statements in this prospectus. See the section titled “Special Note Regarding Forward-Looking Statements” for additional information.

Among others, we refer to estimates compiled by the following industry sources:

•Artificial Analysis, Inc.;

•Bloomberg Intelligence;

•Gartner, Inc.* (“Gartner®”), Press Release, Gartner Poll Finds 55% of Organizations are in Piloting or Production Mode with Generative AI, October 2023;

•International Data Corporation (“IDC”), Generative Artificial Intelligence: A New Chapter for Enterprise Business Applications, March 2023 (#US50471523);

•McKinsey & Company (“McKinsey”), The economic potential of generative AI: The next productivity frontier, June 2023; and

•McKinsey, The state of AI in GCC countries—and how to overcome adoption challenges, May 2023.

* The Gartner content described herein (the “Gartner Content”) represent(s) research opinion or viewpoints published, as part of a syndicated subscription service, by Gartner, Inc. (“Gartner”), and are not representations of fact. Gartner Content speaks as of its original publication date (and not as of the date of this prospectus), and the opinions expressed in the Gartner Content are subject to change without notice.

USE OF PROCEEDS

We estimate that we will receive net proceeds from this offering of approximately $ (or $ if the underwriters exercise their option to purchase additional shares of Class A common stock in full), based on an assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. We will not receive any proceeds from the sale of Class A common stock by the selling stockholders.

Each $1.00 increase or decrease in the assumed initial public offering price per share of $ , which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would increase or decrease, as applicable, the net proceeds to us from this offering by approximately $ , assuming that the number of shares of Class A common stock offered by us, as set forth on the cover page of this prospectus, remains the same and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. Similarly, each increase or decrease of 1.0 million shares in the number of shares of Class A common stock offered by us would increase or decrease, as applicable, the net proceeds to us from this offering by approximately $ , assuming that the initial public offering price per share remains at $ , which is the midpoint of the estimated price range set forth on the cover page of this prospectus, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us.

The principal purposes of this offering are to obtain additional capital to fund our operations, create a public market for our Class A common stock, facilitate an orderly distribution of shares for the selling stockholders, facilitate our future access to the public equity markets, and increase awareness of our company among potential partners. We currently intend to use the net proceeds from this offering, together with our existing cash, cash equivalents, and investments, for general corporate purposes, including working capital, operating expenses and capital expenditures. We may also use a portion of the net proceeds to in-license, acquire, or invest in complementary technologies, assets, businesses, or intellectual property. We periodically evaluate strategic opportunities; however, we have no current commitments to enter into any such acquisitions or make any such investments.

We intend to use a portion of the net proceeds from this offering to satisfy tax withholding and remittance obligations related to the RSU Net Settlement. Based on the assumed initial public offering price of $ per share of Class A common stock, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, an estimated shares underlying RSUs vesting in connection with our initial public offering, and an assumed % tax withholding rate, we would use approximately $ to satisfy our tax withholding and remittance obligations related to the vesting of such RSUs. A $1.00 increase or decrease, as applicable, in the assumed initial public offering price of $ per share of Class A common stock, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, assuming no change to the applicable tax rate, would increase or decrease, as applicable, the amount we would be required to pay to satisfy these tax withholding and remittance obligations by approximately $ .

In addition, it is possible that in the future, we will decide to “net settle” additional RSUs upon the applicable vesting date, meaning that we will withhold a portion of the vested shares on the applicable vesting date and use some of the net proceeds from this offering to satisfy tax withholding and remittance obligations related to the vesting and settlement of such awards.

The expected use of net proceeds from this offering represents our intentions based upon our present plans and business conditions. We cannot predict with certainty all of the particular uses for the proceeds of this offering or the amounts that we will actually spend on the uses set forth above. Accordingly, our management will have broad discretion in applying the net proceeds of this offering. The timing and amount of our actual expenditures will be based on many factors, including cash flows from operations and the anticipated growth of our business.

Pending their use, we intend to invest the net proceeds from this offering in a variety of capital-preservation investments, including short- and intermediate-term investments, interest-bearing investments, investment-grade securities, government securities, and money market funds.

DIVIDEND POLICY

We have never declared or paid any cash dividends on our capital stock. We currently intend to retain any future earnings for use in the operation of our business and do not anticipate declaring or paying any cash dividends in the foreseeable future. Any future determination to declare cash dividends will be made at the discretion of our board of directors, subject to applicable laws and will depend on a number of factors, including our results of operations, financial condition, capital requirements, contractual restrictions, general business conditions, and other factors our board of directors may deem relevant.

CAPITALIZATION

The following table sets forth our cash and cash equivalents and total capitalization as of June 30, 2024:

•on an actual basis;

•on a pro forma basis to give effect to the following immediately prior to the completion of this offering: (i) the filing and effectiveness of our amended and restated certificate of incorporation; (ii) the Preferred Stock Conversion; (iii) the Option Exercise; (iv) the RSU Net Settlement; (v) the increase in accrued expenses and other current liabilities and an equivalent decrease in additional paid-in capital of $ in connection with the estimated tax withholding and remittance obligations related to the RSU Net Settlement; and (vi) stock-based compensation expense of approximately $ that we will recognize upon the completion of this offering related to RSUs subject to service-based and liquidity-based vesting conditions for which the service-based vesting condition was satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering; and

•on a pro forma as adjusted basis to give effect to: (i) the pro forma adjustments set forth above; (ii) the issuance and sale of shares of Class A common stock by us in this offering at an assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us; (iii) the receipt by us of gross proceeds of approximately $ in connection with the Option Exercise; and (iv) the use of a portion of the net proceeds from this offering to satisfy the estimated tax withholding and remittance obligations related to the RSU Net Settlement.

The pro forma as adjusted information discussed below is illustrative only and will be adjusted based on the actual initial public offering price and other terms of this offering determined at pricing. This table should be read in conjunction with the sections titled “Summary Consolidated Financial Data” and “Management’s Discussion and Analysis of Financial Condition and Results of Operations” and our unaudited condensed consolidated financial statements and related notes included elsewhere in this prospectus.

As of June 30, 2024

Actual

Pro Forma

As Adjusted(1)

(in thousands, except share and per share amounts)

Cash and cash equivalents

90,931

Redeemable convertible preferred stock, par value $0.00001 per share; 104,386,199 shares authorized, 77,032,857 shares issued and outstanding, actual; no shares authorized, issued, or outstanding, pro forma and pro forma as adjusted

722,780

Stockholders’ deficit:

Preferred stock, par value $0.00001 per share; no shares authorized, issued, or outstanding, actual; 100,000,000 shares authorized, no shares issued or outstanding, pro forma and pro forma as adjusted

Common stock, par value $0.00001 per share; 203,087,000 shares authorized, 51,177,237 shares issued and outstanding, actual; no shares authorized, issued, or outstanding, pro forma and pro forma as adjusted

Class A common stock, par value $0.00001 per share; no shares authorized, issued, or outstanding, actual; 2,000,000,000 shares authorized and shares issued and outstanding, pro forma; 2,000,000,000 shares authorized and shares issued and outstanding, pro forma as adjusted


	As of June 30, 2024
	Actual		Pro Forma	Pro Forma As Adjusted(1)
	(in thousands, except share and per share amounts)
Class N common stock, par value $0.00001 per share; no shares authorized, issued, or outstanding, actual; 100,000,000 shares authorized, no shares issued or outstanding, pro forma and pro forma as adjusted
Additional paid-in capital	148,386
Treasury stock	(88)
Accumulated other comprehensive income	550
Accumulated deficit	(728,160)
Total stockholders’ deficit	(579,312)
Total capitalization	$	143,468	$	$

_______________

(1)The pro forma as adjusted information discussed above is illustrative only and will depend on the actual initial offering price and other terms of this offering determined at pricing. Each $1.00 increase or decrease in the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would increase or decrease, as applicable, each of pro forma as adjusted cash and cash equivalents, additional paid-in capital, total stockholders’ equity, and total capitalization by approximately $ , assuming that the number of shares of Class A common stock offered by us, as set forth on the cover page of this prospectus, remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. Similarly, each increase or decrease of 1.0 million shares in the number of shares of Class A common stock offered by us would increase or decrease, as applicable, each of pro forma as adjusted cash and cash equivalents, additional paid-in capital, total stockholders’ equity, and total capitalization by approximately $ , assuming that the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us.

If the underwriters exercise their option to purchase additional shares of our Class A common stock in full, pro forma as adjusted cash and cash equivalents, additional paid-in capital, total stockholders’ equity, total capitalization and shares of Class A common stock outstanding as of June 30, 2024 would be $ , $ , $ , $ , and shares, respectively.

The number of shares of our common stock issued and outstanding, pro forma, and pro forma as adjusted in the table above is based shares of our Class A common stock and no shares of our Class N common stock outstanding as of June 30, 2024, after giving effect to the Preferred Stock Conversion, the Option Exercise, and the RSU Net Settlement, and excludes:

• shares of our Class A common stock issuable upon the exercise of stock options granted after June 30, 2024, with a weighted-average exercise price of $ per share;

• shares of our Class A common stock reserved for future issuance under the 2024 Plan, which will become effective on the business day immediately prior to the date of effectiveness of the registration statement of which this prospectus forms a part, including new shares and the number of shares (i) that remain available for grant of future awards under the 2016 Plan at the time the 2024 Plan becomes effective, which shares will cease to be available for issuance under the 2016 Plan at such time and (ii) underlying outstanding Prior Plan Awards that expire, or are cancelled, forfeited, reacquired, or withheld; and

• shares of our Class A common stock reserved for future issuance under the ESPP, which will become effective on the business day immediately prior to the date of effectiveness of the registration statement of which this prospectus forms a part.

DILUTION

If you purchase shares of our Class A common stock in this offering, your ownership interest will be diluted to the extent of the difference between the initial public offering price per share of our Class A common stock in this offering and the pro forma as adjusted net tangible book value per share of our Class A common stock immediately after this offering.

As of June 30, 2024, our historical net tangible book value (deficit) was $ , or $ per share of our Class A common stock. Our historical net tangible book value (deficit) per share represents our total tangible assets less total liabilities and redeemable convertible preferred stock, divided by the aggregate number of shares of our Class A common stock outstanding as of June 30, 2024.

Our pro forma net tangible book value as of June 30, 2024 was $ , or $ per share of Class A common stock. Pro forma net tangible book value per share represents tangible assets, less liabilities, divided by the aggregate number of shares of Class A common stock outstanding, after giving effect to (i) the filing and effectiveness of our amended and restated certificate of incorporation; (ii) the Preferred Stock Conversion; (iii) the Option Exercise; (iv) the RSU Net Settlement; (iv) the increase in accrued expenses and other current liabilities and an equivalent decrease in additional paid-in capital of $ in connection with the estimated tax withholding and remittance obligations related to the RSU Net Settlement; and (v) stock-based compensation expense of approximately $ that we will recognize upon the completion of this offering related to RSUs subject to service-based and liquidity-based vesting conditions for which the service-based vesting condition was satisfied as of June 30, 2024 and for which the liquidity-based vesting condition will be satisfied in connection with this offering.

After giving effect to (i) the pro forma adjustments set forth above, (ii) the sale by us of shares of our Class A common stock in this offering at an assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us, (iii) the receipt by us of gross proceeds of approximately $ in connection with the Option Exercise, and (iv) the use of a portion of the net proceeds from this offering to satisfy the estimated tax withholding and remittance obligations related to the RSU Net Settlement, our pro forma as adjusted net tangible book value as of June 30, 2024 would have been $ , or $ per share. This represents an immediate increase in pro forma net tangible book value to existing stockholders of $ per share and an immediate dilution in pro forma net tangible book value to new investors of $ per share. Dilution per share represents the difference between the price per share to be paid by new investors for the shares of our Class A common stock sold in this offering and the pro forma as adjusted net tangible book value per share immediately after this offering.

The following table illustrates this dilution on a per share basis:


Assumed initial public offering price per share		$
Historical net tangible book value (deficit) per share as of June 30, 2024	$
Pro forma increase in net tangible book value per share as of June 30, 2024 attributable to the pro forma transactions described above
Pro forma net tangible book value per share as of June 30, 2024
Increase in pro forma net tangible book value per share attributable to new investors participating in this offering
Pro forma as adjusted net tangible book value per share after this offering
Dilution per share to new investors participating in this offering		$

Each $1.00 increase or decrease in the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would increase or decrease, as applicable, our pro forma as adjusted net tangible book value per share after this offering by $ per share and the dilution in pro forma per share to investors participating in this offering by $ per share, assuming that the

number of shares of Class A common stock offered by us, as set forth on the cover page of this prospectus, remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. Similarly, each increase or decrease of 1.0 million shares in the number of shares of Class A common stock offered by us would increase or decrease, as applicable, our pro forma as adjusted net tangible book value per share after this offering by $ per share and the dilution in pro forma as adjusted net tangible book value per share to investors participating in this offering by $ per share, assuming the initial public offering price of $ per share remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us.

If the underwriters exercise their option to purchase additional shares of our Class A common stock in full, the pro forma as adjusted net tangible book value per share of our Class A common stock after this offering would be $ per share, and the dilution in pro forma net tangible book value per share to investors participating in this offering would be $ per share of our Class A common stock.

The following table sets forth, on the pro forma basis described above, as of June 30, 2024, the number of shares of Class A common stock purchased from us, the total consideration paid, or to be paid, and the weighted-average price per share paid, or to be paid, by existing stockholders and by the new investors, at an assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, before deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us:


	Shares Purchased			Total Consideration			Weighted- Average Price Per Share
	Number	Percent		Amount	Percent
	(in thousands, except share, per share and percent data)
Existing stockholders		%		$	%		$
New investors
Total		100	%	$	100	%

Sales by the selling stockholders in this offering will cause the number of shares held by existing stockholders before this offering reflected in the table above to be reduced to shares, or % of the total number of shares of our Class A common stock outstanding immediately after the completion of this offering, and will increase the number of shares held by new investors to shares, or % of the total number of shares of our Class A common stock outstanding immediately after the completion of this offering.

Each $1.00 increase or decrease in the assumed initial public offering price of $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would increase or decrease, as applicable, the total consideration paid by new investors, total consideration paid by all stockholders, and the weighted-average price per share paid by all stockholders by approximately $ , $ , and $ , respectively, assuming that the number of shares of Class A common stock offered by us and the selling stockholders, as set forth on the cover page of this prospectus, remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us. Similarly, each increase or decrease of 1.0 million shares in the number of shares of Class A common stock offered by us would increase or decrease, as applicable, the total consideration paid by new investors, total consideration paid by all stockholders, and the weighted-average price per share paid by all stockholders by approximately $ , $ , and $ , respectively, assuming the assumed initial public offering price of $ per share remains the same, and after deducting estimated underwriting discounts and commissions and estimated offering expenses payable by us.

The foregoing tables assume no exercise of the underwriters’ option to purchase additional shares. If the underwriters’ exercise their option to purchase additional shares of Class A common stock in full, the number of shares of Class A common stock held by our existing stockholders will represent approximately % of the total number of shares of our Class A common stock outstanding after this offering and the number of shares held by new

investors will represent approximately % of the total number of shares of our Class A common stock outstanding after this offering.

The foregoing tables and calculations (other than the historical net tangible book value calculation) are based on shares of our Class A common stock and no shares of our Class N common stock outstanding as of June 30, 2024, after giving effect to the Preferred Stock Conversion, the Option Exercise, and the RSU Net Settlement, and excludes:

• shares of our Class A common stock issuable upon the exercise of stock options granted after June 30, 2024, with a weighted-average exercise price of $ per share;

•a variable number of shares of our Class N common stock that may be issued pursuant to the G42 Option (see the sections titled “—G42 Option” and “Certain Relationships and Related Party Transactions” for additional information);

If all of the foregoing securities, other than the shares reserved for issuance pursuant to the G42 Option, the 2024 Plan, or the ESPP, were converted, exercised, or vested in connection with this offering, the number of shares of Class A common stock held by our existing stockholders would represent approximately % of the total number of shares of our Class A common stock outstanding after this offering and the number of shares held by new investors would represent approximately % of the total number of shares of our Class A common stock

outstanding after this offering, in each case, assuming no exercise of the underwriters’ option to purchase additional shares.

To the extent we issue any additional stock options, warrants, or RSUs or any outstanding stock options or RSUs are exercised or settled, or to the extent we issue any other securities or convertible debt in the future, including in connection with the G42 Primary Purchase or the G42 Option, investors will experience further dilution.

G42 Option

Investors may experience further dilution in connection with the G42 Option. Pursuant to the Preferred Stock Purchase Agreement, if G42 or certain third parties at the direction of G42 purchase more than $500.0 million in one purchase order, and less than $5.0 billion in the aggregate, of high-performance computing clusters from us (the “G42 Option Threshold”), we will grant G42 the option to purchase additional shares of our Series F-2 redeemable convertible preferred stock (or, if such option is granted or exercised following the completion of this offering, shares of our Class N common stock), subject to the terms and conditions thereof (the “G42 Option”). Each share of our Class N common stock is convertible at any time at the option of the holder into one share of our Class A common stock. G42 has agreed to not convert its shares of Class N common stock into shares of Class A common stock before July 31, 2025. See the section titled “Description of Capital Stock—Common Stock” for additional information.

Shares may be issued pursuant to the G42 Option in one or more closings. The $1.43 billion of products and services that G42 has committed to purchase pursuant to the G42 May 2024 Agreement (as defined in “Management’s Discussion & Analysis—G42 Relationship”) does not count toward the G42 Option Threshold. The G42 Option expires if the G42 Option Threshold is not achieved by December 31, 2025.

The maximum number of shares that may be purchased pursuant to the G42 Option will be the quotient of (i) the total aggregate purchase price for the G42 Option, which will equal 10% of the value of the relevant purchaser order(s), divided by (ii) (A) if the G42 Option is granted and exercised prior to the completion of this offering, a price per share that is 17.5% below the price per share of our then most recent arms-length sale of our redeemable convertible preferred stock (excluding our Series F-1 and Series F-2 redeemable convertible preferred stock), or (B) if the G42 Option is granted or exercised following the completion of this offering, a price per share that is 17.5% below the average closing price per share of our Class A common stock over the 30-day period prior to the G42 Option Threshold being met.

Assuming the G42 Option Threshold is satisfied after this offering, the minimum and maximum number of shares that could be purchased pursuant to the G42 Option, assuming the average closing price per share of our Class A common stock over the 30-day period prior to the G42 Option Threshold being met is $ per share, which is the midpoint of the estimated price range set forth on the cover page of this prospectus, would be shares of Class N common stock (assuming the G42 Option Threshold is satisfied by product sales of $500.0 million in one purchase order) and shares of Class N common stock (assuming the G42 Option Threshold is satisfied by product sales of $5.0 billion in the aggregate), respectively. If the price per share of our Class A common stock over the 30-day period prior to the G42 Option Threshold being met is greater or lesser than $ per share, the number of shares that could be purchased pursuant to the G42 Option would decrease or increase, respectively.

MANAGEMENT’S DISCUSSION AND ANALYSIS OF FINANCIAL CONDITION AND RESULTS OF OPERATIONS

The following discussion and analysis of our financial condition and results of operations should be read in conjunction with the “Summary Consolidated Financial Data” section, our audited consolidated financial statements and related notes, and other financial information appearing elsewhere in this prospectus. In addition to historical consolidated financial information, the following discussion contains forward-looking statements that reflect our plans, estimates, and beliefs that involve significant risks and uncertainties. Our actual results could differ materially from those discussed in the forward-looking statements. Factors that could cause or contribute to those differences include those discussed below and elsewhere in this prospectus, particularly in “Risk Factors” and “Special Note Regarding Forward-Looking Statements.”

Data as of and for the years ended December 31, 2023 and 2022 has been derived from our audited consolidated financial statements included elsewhere in this prospectus. Data as of and for the six months ended June 30, 2024 and 2023 has been derived from our unaudited interim condensed consolidated financial statements included elsewhere in this prospectus. Our historical results are not necessarily indicative of the results to be expected for any period in the future, and results for any interim period should not be construed as an inference of what our results would be for any full year or future period.

Overview

Our mission is to accelerate AI by making it faster, easier to use, and more energy efficient, making AI accessible around the world.

We design processors for AI training and inference. We build AI systems to power, cool, and feed the processors data. We develop software to link these systems together into industry-leading supercomputers that are simple to use, even for the most complicated AI work, using familiar ML frameworks like PyTorch. Customers use our supercomputers to train industry-leading models. We use these supercomputers to run inference at speeds unobtainable on alternative commercial technologies. We deliver these AI capabilities to our customers on premise and via the cloud.

Our business model is designed to meet the needs of our customers. Organizations seeking control over their data and AI compute infrastructure can purchase Cerebras AI supercomputers for on-premise deployment. Those that want the flexibility of a cloud-based platform can purchase Cerebras high-performance AI compute via a consumption-based model through the Cerebras Cloud. We offer customers the flexibility to choose the solution that best aligns with their budgetary, security, and scalability requirements, and some customers choose to use both options simultaneously.

Our customers include CSPs, leading enterprises, Sovereign AI programs, national laboratories, research institutions, and other innovators at the forefront of AI and at the intersection of AI and HPC. We are working to grow our user base and collaborate with our customers to harness the power of AI to tackle their most significant challenges and drive breakthroughs across industries.

We have experienced rapid growth, with revenue of $78.7 million and $24.6 million for the years ended December 31, 2023 and 2022, respectively, representing year-over-year growth of 220%. During the six months ended June 30, 2024 and 2023, we generated $136.4 million and $8.7 million in revenue, respectively. Our net loss for the years ended December 31, 2023 and 2022 was $127.2 million and $177.7 million, respectively, representing a year-over-year reduction of 28%. Our net loss for the six months ended June 30, 2024 and 2023 was $66.6 million and $77.8 million, respectively, representing a year-over-year reduction of 14%.

Our Business Model

collection of services from data center deployment to AI expert services and AI Supercomputer operation and management, to provide our customers with the support they need to train, deploy, and accelerate GenAI time-to-value.

On Premise. We sell our AI Supercomputers to leading organizations who seek maximum control over their data and their AI infrastructure, fulfilling their needs for high-performance AI compute on premise. For on-premise use, we offer deployment services, as well as a subscription to an ongoing stream of software updates and upgrades.

Cloud-Based Computing Services. We sell Cerebras solutions primarily via our hosted cloud offering. Our cloud solutions provide customers fast and flexible access to our powerful AI acceleration hardware, with payment based on time or work (number of models). This offering gives our customers the ability to train LLMs with extraordinary speed and deploy them for inference at ultra-low latencies, all without the complexity or time needed to build and manage on-premise infrastructure.

Cerebras Inference Cloud. Our real-time inference solution is also available via a dedicated inference cloud service. Leveraging our Cerebras Inference Serving Stack, this cloud API offering allows developers to directly point their applications towards efficient and reliable model serving endpoints. On Cerebras Inference Cloud, we host both popular open-source models and proprietary customer models. For customers who do not need direct compute access and are not interested in managing their own inference serving software stack, our inference cloud offering is the quickest and simplest way to leverage our fast model inference services.

We provide a combination of these offerings to customers who may benefit from leveraging both on-premise and cloud-based options—for example, enabling them to quickly use the cloud for their largest training jobs, while enjoying the cost efficiencies of on-premise infrastructure for their baseline AI work. This flexibility allows customers to choose the solution that best aligns with their budgetary, security, and scalability requirements.

We also provide professional services to assist customers throughout the AI workflow.

G42 Relationship

We have developed a strategic relationship with Group 42 Holding Ltd (together with its affiliates, “G42”) as a partner, customer, and investor. G42, a technology holding group, is an AI and cloud-computing company that develops and deploys cutting-edge, large-scale, AI-powered solutions for governments and large corporations. Founded in Abu Dhabi and operating worldwide, G42 offers a broad range of products and services that apply AI, cloud computing, and data analytics in several verticals, including healthcare, smart cities, finance, energy, and consumer businesses. We have been informed by G42 that its portfolio companies—which contribute AI-driven solutions across various sectors—include Bayanat, a developer of surveying technology; Core42, a provider of cloud infrastructure and professional services; Inception, a developer of AI data and models; Khazna Data Centers, a provider of wholesale data centers; M42, a joint venture between G42 and Mubadala and a provider of medical solutions leveraging data-centric technologies; and Presight, a provider of turnkey solutions across multiple sectors using big data analytics and AI, which also controls AIQ, an AI pioneer in the energy sector. We have been informed by G42 that as of June 30, 2024, G42 was owned by Mubadala, a sovereign investor owned by the Government of Abu Dhabi; SilverLake, a leading global private equity firm specializing in technology; Microsoft Corporation, a multinational technology corporation; the Dalio Family Office, the private investment office of the Dalio family; RGH1 Investment SPV RSC Ltd., a member of the Royal Group of companies; and the Kai Foundation, a foundation incorporated in the Abu Dhabi Global Market.

In November 2021, G42, through an affiliate, acquired an approximate 1% ownership interest in our company as of such date by purchasing approximately 1.4 million shares of our Series F redeemable convertible preferred stock at $27.74 per share, resulting in gross proceeds to us of approximately $40.0 million.

During the year ended December 31, 2023, G42 purchased cloud-based computing services in the second quarter. In the third quarter of 2023, we entered into a commercial framework agreement and received the first purchase order for an on-premise high-performance computing cluster, as well as a multi-year contract for related data center space and operational services. We began recognizing significant revenue from G42 in the third quarter of 2023. We also received new purchase orders for two additional high-performance computing clusters and related services in the fourth quarter of 2023. We generated significant revenue from G42 in the year ended December 31, 2023, representing $65.1 million, or 83%, of our total revenue.

In April 2024, we entered into an agreement with G42 (the “G42 April 2024 Agreement”) pursuant to which G42, or a third party nominee affiliated with G42, intends to issue purchase orders for high-performance computing products and services for a minimum value of $300 million. Pursuant to the G42 April 2024 Agreement, we received a prepayment of $300.0 million from G42 in May 2024 to be used for payments to third-party vendors to manufacture high-performance computing infrastructure. If the purchase orders are not issued by G42, any portion of the prepayment not paid by us to our third-party vendors will be payable to G42 on demand and our rights to inventory purchased with the prepayment will transfer to G42.

In May 2024, we entered into a Series F-1 redeemable convertible purchase agreement (as subsequently amended and restated in September 2024, the “Preferred Stock Purchase Agreement”) with G42 and other parties. Pursuant to the Preferred Stock Purchase Agreement, G42 agreed to purchase an aggregate of approximately 22.9 million shares of our Series F-2 redeemable convertible preferred stock (or, if purchased following the completion of this offering, shares of our Class N common stock) at $14.66 per share, for anticipated gross proceeds to us of $335.0 million (the “G42 Primary Purchase”). G42 has committed to purchase these shares by April 15, 2025. Following the completion of the G42 Primary Purchase, the shares owned by G42 will represent approximately % of our outstanding capital stock following the completion of this offering (or approximately % of our outstanding capital stock if the underwriters exercise in full their option to purchase additional shares from us). Additionally, pursuant to the Preferred Stock Purchase Agreement, under certain circumstances, G42 will have the option to purchase additional shares of our Class N common stock at a 17.5% discount to the then-fair market value of our shares of Class A common stock. For additional information, see the section titled “Certain Relationships and Related Party Transactions.”

In May 2024, we also entered into an agreement with G42 (the “G42 May 2024 Agreement”) pursuant to which we agreed to certain pricing commitments with G42 through the end of 2025, and G42 agreed that it will purchase, or will cause a third party that may be affiliated or unaffiliated and under a commercial agreement with G42, to purchase, our high-performance computing systems, installation, and support services in an aggregate amount of approximately $1.43 billion by completing the prepayment of such amount before February 28, 2025 and executing binding purchase orders totaling such amount. G42 has informed us that any third party with whom it coordinates to purchase high-performance computing clusters from us would engage G42 or its affiliates to manage and operate the high-performance computing clusters for a fee. The Cerebras systems would be housed in facilities leased by G42 or its affiliates, who would oversee their operation. In September 2024, Mohamed bin Zayed University of Artificial Intelligence (“MBZUAI”), an Abu Dhabi-based university focused on research and graduate-level education in AI, agreed to purchase $350.0 million of the $1.43 billion committed pursuant to the G42 May 2024 Agreement. To the extent MBZUAI or other third parties make purchases subject to the G42 May 2024 Agreement, such purchases would reduce the amount that G42 purchases directly, but would not alter the total amount committed for purchase under the G42 May 2024 Agreement.

In July 2024, we received a purchase order from G42 for the sale of $178.7 million of our high-performance computing systems, including installation and support services (the “July 2024 PO”). Additionally, we entered into a new agreement, and modified an existing agreement, pursuant to which we agreed to provide operation and management services for certain high-performance computing systems purchased from us by G42. The fees to be paid by G42 under these agreements will be deducted from the initial $300 million prepayment we received as part of the G42 April 2024 Agreement, and the $178.7 million fee for the July 2024 PO fulfills a portion of the prepayment and purchase order commitments under the G42 May 2024 Agreement.

In September 2024, we received a $350.0 million prepayment from MBZUAI in connection with the purchase order described above. The prepayment will be used for payments to third-party vendors to manufacture high performance computing infrastructure.

Key Factors Affecting Our Performance

We believe that the growth and future success of our business depends on many factors. While these factors present significant opportunities for our business, they also pose important challenges that we will need to address in order to improve our results of operations.

Commercial relationship with G42. We have derived, and expect to continue to derive, a substantial portion of our revenue from sales to G42. Any changes in G42’s demand for our solution, whether due to competitive factors, data center availability, changes in its business strategy, budgetary constraints, or other reasons could materially impact our financial performance in the future. We have received substantial advance payments from G42 for our AI systems to support future installations. These large prepayments have allowed us to reduce our working capital needs and provide G42 access to a significant portion of our initial production volumes to support its ambitious, multi-year AI investment plans.

AI compute demand and increasingly complex models. The escalating global demand for AI training and inference compute, fueled by advancements in LLMs, GenAI, and other AI-powered applications, plays a significant role in the demand for our solution. We have designed our AI compute platform to address the challenges of accelerating large-scale AI workloads, and our customer base in research, government, CSPs, and enterprise sectors reflects the increasing recognition of our differentiated technology. We believe that the increasing complexity of AI models, which require substantial computational resources for training and inference will support the demand for our high-performance AI compute platform. We will depend on continued growth in compute demand to drive our future financial performance.

New customer adoption. Attracting new customers to our platform is a key driver of our revenue growth strategy. We have successfully grown our customer base to include CSPs, leading enterprises, Sovereign AI programs, national laboratories, and research institutions seeking to leverage the power of AI for their specific needs. An increase in new customers will impact our revenue growth and market share and also contributes to a diversified customer base, which can provide greater stability over the long term. Additionally, new customers bring fresh perspectives and use cases, which can drive innovation and product development, further enhancing our competitive position.

AI adoption and the emergence of new use cases and applications. We expect to benefit from the rapid adoption of AI across industries and the emergence of innovative AI use cases and applications. As organizations increasingly recognize the transformative potential of AI to enhance efficiency, productivity, and decision-making, we expect the demand for high-performance AI compute solutions to accelerate. The rapidly expanding range of AI use cases, from natural language processing and computer vision to drug discovery and climate modeling, among others, creates a broad market opportunity for our AI compute solution. We expect our ability to address the evolving needs of diverse industries and applications with cutting-edge AI compute technology to be instrumental in driving our growth and market leadership in the years to come.

Sovereign AI initiatives. The expansion of Sovereign AI initiatives represents a significant opportunity to drive demand for our solution. We believe that our focus on developing secure, high-performance AI infrastructure positions us well to meet this demand.

Investment in technology leadership and product development. Increasing our investment in technology leadership and product development directly impacts our competitive position and future financial performance. Our continued commitment to research and development enables us to focus on emerging market needs and expand into new applications of AI. As we develop new products and services tailored to specific industries and use cases, we can unlock additional demand and continue to broaden our customer base. Additionally, technical leadership is important to help us attract and retain top talent, which is essential for maintaining our technological edge and

driving continuous innovation. A talented workforce, in turn, contributes to the development of new solutions that can continue to fuel our revenue growth.

Supply chain and manufacturing capacity. We operate a fabless business model that utilizes third-party suppliers and manufacturers, such as third-party wafer foundries and module assembly and test service providers in a number of countries, including outside the United States. We do not generally have long-term capacity commitments with our suppliers, and we source a number of the components used in our products from sole or single-source suppliers or use a single supplier to perform certain of the processes involved in the manufacture of our products. The continued and timely supply of input materials and the availability of manufacturing capacity and packaging and testing services impact our ability to meet customer demand. Onboarding new third-party suppliers and our dependency on them to allocate sufficient manufacturing capacity to meet our needs in a cost-effective and timely manner may impact our ability to scale and support growing customer demand.

Non-GAAP Financial Measures

Non-GAAP Operating Loss

A reconciliation of our GAAP operating loss, the most directly comparable GAAP financial measure, to non-GAAP operating loss is presented below:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
GAAP operating loss	$	(133,934)	$	(178,821)	$	(41,811)	$	(81,015)
Add: Stock-based compensation expense	26,631		23,044		32,329		9,260
Non-GAAP operating loss	$	(107,303)	$	(155,777)	$	(9,482)	$	(71,755)

Non-GAAP Net Loss

A reconciliation of our GAAP net loss, the most directly comparable GAAP financial measure, to our non-GAAP net loss is presented below:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
GAAP net loss	$	(127,155)	$	(177,719)	$	(66,605)	$	(77,820)
Add: Stock-based compensation expense(1)	26,631		23,044		32,329		9,260
Add: Change in fair value of forward contract liability	—		—		30,327		—
Non-GAAP net loss	$	(100,524)	$	(154,675)	$	(3,949)	$	(68,560)

_______________

(1)Non-GAAP net loss does not include the tax effects of the stock-based compensation expense adjustment because such tax effects were not material during the periods presented.

Components of Results of Operations

Revenue

We generate revenue primarily from the sale of AI systems, support services, cloud-based computing services, and custom AI modeling services. We primarily sell directly to end customers. Contracts with our customers typically include multiple performance obligations. For contracts with more than one performance obligation, we allocate the transaction price to each separate obligation. Our payment terms vary by contract type and type of customer and generally range from 30 to 60 days from the invoice date.

Hardware Revenue

Hardware revenue consists of sales of our AI systems that can be used for both training and inference. We recognize revenue from sales of AI systems when control of the goods transfers to the customer, which generally occurs upon shipment or delivery, depending on shipping terms or upon meeting the contractual acceptance terms.

Services and Other Revenue

We generate services and other revenue primarily through sales of one- to three-year support services, cloud-based computing services, and custom AI modeling services.

We recognize revenue from sales of support services and cloud-based computing services, including hosted inference, over the service term, as the customer benefits from our services throughout the contract period.

We generate revenue from custom AI modeling services over time as services are provided or at a point-in-time upon completion and acceptance by the customer of contract deliverables, depending on the terms of the agreement.

Cost of Sales and Gross Profit

Hardware Cost of Sales

Cost of sales for hardware consists primarily of the cost of materials, such as wafers processed by third-party foundries, costs associated with packaging, assembly, shipping, logistics, quality assurance, warranty cost, cost of personnel, including salaries, stock-based compensation, and employee benefits, write-down of inventories, and facilities expenses.

Services and Other Cost of Sales

Cost of sales for services and other revenue primarily consist of data center costs, depreciation of equipment, cost of personnel, including salaries, stock-based compensation, and employee benefits, and facilities expenses.

Gross Profit and Gross Margin

Gross profit represents revenue less cost of sales. Gross margin is gross profit expressed as a percentage of revenue. Our gross profit has been, and we expect will continue to be, influenced by several factors, including sales volume and pricing of our products and services, changes in inventory costs, including wafer yield, contract manufacturing, and supplier pricing, data center costs, cost of logistics, and personnel costs.

Operating Expenses

Research and Development Expenses

Research and development expenses primarily consist of costs incurred in performing research and development activities and include salaries, stock-based compensation, employee benefits, tape-out costs, which include layout services, mask sets, prototype components, system qualification and testing incurred before releasing new system designs into production, shipping, data center costs, depreciation and amortization, professional services fees, cloud computing, and facilities expenses. We expense research and development costs as incurred.

We also expense software development costs, including costs to develop the software component of hardware to be sold, leased, or marketed to external users, before technological feasibility is reached. Technological feasibility is typically reached shortly before the release of such products.

Sales and Marketing Expenses

Sales and marketing expenses primarily consist of personnel costs, including salaries, stock-based compensation, employee benefits, public relations costs, tradeshow and other sales event costs, travel and entertainment costs, and facilities expenses.

General and Administrative Expenses

General and administrative expenses consist primarily of personnel costs, including salaries, stock-based compensation, employee benefits and bonuses related to corporate, finance, legal, information technology and human resource functions, professional services fees, audit and compliance expenses, software subscription costs, travel and entertainment costs, insurance costs, depreciation and amortization, allocation of facilities and other general corporate expenses. We expect to incur additional expenses as a result of operating as a public company, including expenses to comply with the rules and regulations applicable to companies listed on a national securities exchange, expenses related to auditing, compliance, and reporting obligations pursuant to the rules and regulations of the SEC, as well as higher expenses for general and director and officer insurance, investor relations, and professional services.

Interest Income and Other Income (Expense), Net

Interest income and other income (expense), net consists primarily of interest income, dividend income, and realized gains or losses on debt securities, foreign exchange losses, and gains and losses arising from remeasurement of forward contract liability and warrant liability to fair value at each reporting date.

Income Tax Expense

Income tax expense consists of U.S. federal and state income taxes and income taxes in certain foreign jurisdictions in which we conduct business. We maintain a full valuation allowance on our federal and state deferred tax assets as we have concluded that it is more likely than not that the deferred tax assets will not be realized. Our effective tax rate is affected by tax rates in foreign jurisdictions and the relative amounts of income we earn in those jurisdictions, as well as non-deductible expenses, such as stock-based compensation, and changes in our valuation allowance.

Results of Operations

The following tables set forth selected consolidated statements of operations data for each of the periods indicated:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
Revenue:
Hardware	$	57,114	$	15,599	$	104,269	$	1,559
Services and other	21,630		9,020		32,133		7,105
Total revenue	78,744		24,619		136,402		8,664
Cost of sales(1):
Hardware	45,559		19,195		66,442		1,980
Services and other	6,827		2,534		13,941		2,306
Total cost of sales	52,386		21,729		80,383		4,286
Gross profit	26,358		2,890		56,019		4,378
Operating expenses:
Research and development(1)	140,057		155,408		77,742		76,295
Sales and marketing(1)	9,642		9,401		7,237		4,176
General and administrative(1)	10,593		16,902		12,851		4,922
Total operating expenses	160,292		181,711		97,830		85,393
Loss from operations	(133,934)		(178,821)		(41,811)		(81,015)
Interest income	5,683		1,076		3,809		2,349
Other income (expense), net	1,228		230		(28,284)		918
Loss before income tax	(127,023)		(177,515)		(66,286)		(77,748)
Income tax expense	132		204		319		72
Net loss	$	(127,155)	$	(177,719)	$	(66,605)	$	(77,820)

_______________

(1)Includes stock-based compensation expense as follows:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
Cost of sales	$	309	$	223	$	420	$	51
Research and development	21,187		17,732		23,905		7,367
Sales and marketing	3,563		832		3,195		1,075
General and administrative	1,572		4,257		4,809		767
Total stock-based compensation expense	$	26,631	$	23,044	$	32,329	$	9,260

Stock-based compensation expense included $9.0 million, $8.6 million, $18.5 million, and $0.9 million for the years ended December 31, 2023 and 2022 and for the six months ended June 30, 2024 and 2023, respectively, related to secondary transactions in each period and a common stock repurchase from employees during the year ended December 31, 2022. See Note 12 to our audited consolidated financial statements and our unaudited interim condensed consolidated financial statements included elsewhere in this prospectus for additional details on the secondary transactions.

The following table sets forth selected consolidated statements of operations data expressed as a percentage of revenue for each of the periods indicated:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(as a percentage of revenue)
Revenue:
Hardware	72.5	%	63.4	%	76.4	%	18.0	%
Services and other	27.5		36.6		23.6		82.0
Total revenue	100.0		100.0		100.0		100.0
Cost of sales:
Hardware	57.8		78.0		48.7		22.9
Services and other	8.7		10.3		10.2		26.6
Total cost of sales	66.5		88.3		58.9		49.5
Gross profit	33.5		11.7		41.1		50.5
Operating expenses:
Research and development	177.9		631.2		57.0		880.6
Sales and marketing	12.2		38.2		5.3		48.2
General and administrative	13.5		68.7		9.4		56.8
Total operating expenses	203.6		738.1		71.7		985.6
Loss from operations	(170.1)		(726.4)		(30.6)		(935.1)
Interest income	7.2		4.4		2.8		27.1
Other income (expense), net	1.6		0.9		(20.7)		10.6
Loss before income tax	(161.3)		(721.1)		(48.5)		(897.4)
Income tax expense	0.2		0.8		0.2		0.8
Net loss	(161.5)	%	(721.9)	%	(48.7)	%	(898.2)	%

Comparison of the Six Months Ended June 30, 2024 and 2023

Revenue


	Six Months Ended June 30,
	2024		2023		$ Change		% Change
	(in thousands, except percentages)
Hardware	$	104,269	$	1,559	$	102,710	NM
Services and other	32,133		7,105		25,028		NM
Total revenue	$	136,402	$	8,664	$	127,738	1,474	%

_______________

NM – Not meaningful

Our total revenue for the six months ended June 30, 2024 increased by $127.7 million, or 1,474%, compared to the six months ended June 30, 2023. This increase was primarily due to hardware revenue, which grew by $102.7 million. The growth was attributable to a significant increase in the number of AI systems sold. Services and other revenue increased by $25.0 million primarily due to an increase in professional services and ongoing support services as a result of our growing installed base. We generated significant revenue from G42 for the six months ended June 30, 2024 and 2023, representing $119.1 million and $3.7 million, respectively, or 87% and 43%, respectively, of our total revenue. During the six months ended June 30, 2024, G42 represented $101.3 million, or 97%, and $17.8 million, or 56%, of hardware revenue and services and other revenue, respectively. During the six months ended June 30, 2023, G42 represented $3.7 million, or 52%, of services and other revenue. No hardware revenue was recognized from G42 during the six months ended June 30, 2023.

Cost of Sales


	Six Months Ended June 30,
	2024		2023		$ Change		% Change
	(in thousands, except percentages)
Hardware	$	66,442	$	1,980	$	64,462	NM
Services and other	13,941		2,306		11,635		NM
Cost of sales	$	80,383	$	4,286	$	76,097	NM

_______________

NM – Not meaningful

Cost of sales for the six months ended June 30, 2024 increased by $76.1 million compared to the six months ended June 30, 2023 primarily due to an increase in the number of AI systems sold, and growth in our professional services, and ongoing support services.

Gross Profit and Gross Margin


	Six Months Ended June 30,
	2024			2023			$ Change		% Change
	(in thousands, except percentages)
Gross profit	$	56,019		$	4,378		$	51,641	NM
Gross margin	41.1		%	50.5		%

_______________

NM – Not meaningful

Gross profit for the six months ended June 30, 2024 was $56.0 million, an increase of $51.6 million compared to $4.4 million in the six months ended June 30, 2023. Gross margin decreased from 50.5% to 41.1%, primarily due

to a change in the mix of hardware revenue and services and other revenue. Services and other revenue typically have higher gross margins than hardware revenue. During the six months ended June 30, 2023, a significant portion of our revenue was from services and other revenue. During the six months ended June 30, 2024, we extended a larger price discount for hardware sold to G42 in exchange for a high-volume fixed-price purchase order, which lowered the total gross margin, the impact of which was partially offset by lower bill of materials costs.

Operating Expenses

Research and Development


	Six Months Ended June 30,
	2024			2023			$ Change		% Change
	(in thousands, except percentages)
Research and development	$	77,742		$	76,295		$	1,447	2	%
Percentage of revenue	57		%	881		%

Research and development expenses for the six months ended June 30, 2024 remained largely consistent with the six months ended June 30, 2023. During the six months ended June 30, 2024, stock-based compensation expense increased by $16.5 million, though this was largely offset by a reduction in non-recurring engineering services.

Sales and Marketing


	Six Months Ended June 30,
	2024			2023			$ Change		% Change
	(in thousands, except percentages)
Sales and marketing	$	7,237		$	4,176		$	3,061	73	%
Percentage of revenue	5		%	48		%

Sales and marketing expenses for the six months ended June 30, 2024 increased by $3.1 million, or 73%, compared to the six months ended June 30, 2023. The increase in sales and marketing expenses was primarily due to an increase in stock-based compensation expenses and personnel-related expenses as a result of an increase in headcount.

General and Administrative


	Six Months Ended June 30,
	2024			2023			$ Change		% Change
	(in thousands, except percentages)
General and administrative	$	12,851		$	4,922		$	7,929	161	%
Percentage of revenue	9		%	57		%

General and administrative expenses for the six months ended June 30, 2024 increased by $ 7.9 million, or 161%, compared to the six months ended June 30, 2023. The increase in general and administrative expenses was primarily due to an increase in personnel-related expenses, including stock-based compensation expense and legal and other professional services to support our growth.

Interest Income and Other (Expense) Income, Net


	Six Months Ended June 30,
	2024		2023		$ Change		% Change
	(in thousands, except percentages)
Interest income	$	3,809	$	2,349	$	1,460	62	%
Other income (expense), net	$	(28,284)	$	918	$	(29,202)	NM

_______________

NM – Not meaningful

Interest income for the six months ended June 30, 2024 increased by $1.5 million, or 62%, compared to the six months ended June 30, 2023. The increase was primarily attributable to stronger yield on higher invested cash, cash equivalents, and investments. The change in other (expense) income, net was primarily attributable to a change in fair value of Series F-1 redeemable convertible preferred stock forward contract liability recorded during the six months ended June 30, 2024.

Income Tax Expense


	Six Months Ended June 30,
	2024		2023		$ Change		% Change
	(in thousands, except percentages)
Income tax expense	$	319	$	72	$	247	343	%

Income tax expense for the six months ended June 30, 2024 increased compared to the six months ended June 30, 2023 due to a change in the measurement of our deferred tax liability and an increase in state income tax.

Comparison of the Years Ended December 31, 2023 and 2022

Revenue


	Year Ended December 31,
	2023		2022		$ Change		% Change
	(in thousands, except percentages)
Hardware	$	57,114	$	15,599	$	41,515	266	%
Services and other	21,630		9,020		12,610		140	%
Total revenue	$	78,744	$	24,619	$	54,125	220	%

Total revenue for the year ended December 31, 2023 increased by $54.1 million, or 220%, compared to the year ended December 31, 2022, primarily due to the increase in hardware revenue. Hardware revenue increased by $41.5 million for the year ended December 31, 2023 compared to the year ended December 31, 2022. This increase was primarily due to an approximately four-fold increase in the number of AI systems sold. Services and other revenue increased $12.6 million for the year ended December 31, 2023 compared to the year ended December 31, 2022, primarily due to an increase in professional services as well as an increase ongoing support services as a result of our growing installed base. We generated significant revenue from G42 for the year ended December 31, 2023, representing $65.1 million, or 83%, of our total revenue. G42 represented $54.5 million, or 95%, and $10.7 million, or 49%, of hardware revenue and services and other revenue, respectively, for the year ended December 31, 2023. No revenue was recognized from G42 for the year ended December 31, 2022.

Cost of Sales


	Year Ended December 31,
	2023		2022		$ Change		% Change
	(in thousands, except percentages)
Hardware	$	45,559	$	19,195	$	26,364	137	%
Services and other	6,827		2,534		4,293		169	%
Cost of sales	$	52,386	$	21,729	$	30,657	141	%

Cost of sales for the year ended December 31, 2023 increased by $30.7 million, or 141%, compared to the year ended December 31, 2022, primarily due to an increase in the number of AI systems sold as well as increased costs from more professional and ongoing support services being delivered. These increases were partially offset by a $8.3 million decrease in write-offs and provision for excess and obsolete inventory due to transition from an older generation to a newer generation of our AI systems in 2022.

Gross Profit and Gross Margin


	Year Ended December 31,
	2023			2022			$ Change		% Change
	(in thousands, except percentages)
Gross profit	$	26,358		$	2,890		$	23,468	NM
Gross margin	33.5		%	11.7		%

_______________

NM – Not meaningful

Gross profit for the year ended December 31, 2023, was $26.4 million, an increase of $23.5 million compared to $2.9 million in the year ended December 31, 2022. Gross margin for the year ended December 31, 2023 increased to 33.5% from 11.7% for the year ended December 31, 2022. Charges to cost of sales for write-offs and provision for excess and obsolete inventory were $2.5 million and $10.8 million during the years ended December 31, 2023 and 2022, respectively, which negatively affected our gross margin during those years.

Operating Expenses

Research and Development


	Year Ended December 31,
	2023			2022			$ Change		% Change
	(in thousands, except percentages)
Research and development	$	140,057		$	155,408		$	(15,351)	(10)	%
Percentage of revenue	178		%	631		%

Research and development expenses for the year ended December 31, 2023 decreased by $15.4 million, or 10%, compared to the year ended December 31, 2022. The decrease in research and development expenses was primarily attributable to reduced compensation and benefits as a result of a reduction in employee workforce executed during the third quarter of 2022 and the first quarter of 2023.

Sales and Marketing


	Year Ended December 31,
	2023			2022			$ Change		% Change
	(in thousands, except percentages)
Sales and marketing	$	9,642		$	9,401		$	241	3	%
Percentage of revenue	12		%	38		%

Sales and marketing expenses for the year ended December 31, 2023 remained largely consistent with the year ended December 31 2022.

General and Administrative


	Year Ended December 31,
	2023			2022			$ Change		% Change
	(in thousands, except percentages)
General and administrative	$	10,593		$	16,902		$	(6,309)	(37)	%
Percentage of revenue	14		%	69		%

General and administrative expenses for the year ended December 31, 2023 decreased by $6.3 million, or 37%, compared to the year ended December 31, 2022. The decrease in general and administrative expenses was primarily attributable to a decrease in stock-based compensation expense in 2023 compared to the expense recorded for a common stock repurchase transaction in 2022, a decrease in compensation and benefits due to a reduction in employee workforce executed during the third quarter of 2022 and the first quarter of 2023, and a decrease in other general administrative expenses.

Interest Income and Other Income, Net


	Year Ended December 31,
	2023		2022		$ Change		% Change
	(in thousands, except percentages)
Interest income	$	5,683	$	1,076	$	4,607	NM
Other income, net	$	1,228	$	230	$	998	NM

_______________

NM – Not meaningful

Interest income for the year ended December 31, 2023 increased by $4.6 million compared to the year ended December 31, 2022. The increase in interest income was primarily attributable to higher yield received on invested cash, cash equivalents, and investments. Other income, net for the year ended December 31, 2023 increased by $1.0 million compared to the year ended December 31, 2022. The increase in other income, net was primarily attributable to an increase in dividend income and remeasurement of the warrant liability.

Income Tax Expense


	Year Ended December 31,
	2023		2022		$ Change		% Change
	(in thousands, except percentages)
Income tax expense	$	132	$	204	$	(72)	(35)	%

Income tax expense for the year ended December 31, 2023 decreased compared to the year ended December 31, 2022 due to a change in the measurement of our deferred tax liability.

Quarterly Results of Operations

The following table sets forth selected unaudited quarterly consolidated statements of operations data for each of the six quarters ended June 30, 2024. The information for each of these quarters has been prepared on the same basis as our audited consolidated financial statements and reflect, in the opinion of management, all adjustments, consisting of normal, recurring adjustments that are necessary for a fair statement of this information. These quarterly operating results are not necessarily indicative of the results that may be expected for a full year or any other fiscal period. This information should be read in conjunction with our audited consolidated financial statements and related notes included elsewhere in the prospectus.


	Three Months Ended
	March 31, 2023			June 30, 2023			September 30, 2023			December 31, 2023			March 31, 2024			June 30, 2024
	(in thousands)
Revenue:
Hardware	$	1,354		$	205		$	28,933		$	26,622		$	49,411		$	54,858
Services and other	1,583			5,522			6,182			8,343			17,220			14,913
Total revenue	2,937			5,727			35,115			34,965			66,631			69,771
Cost of sales(1):
Hardware	1,094			885			21,867			21,713			33,619			32,823
Services and other	488			1,819			2,254			2,266			7,873			6,068
Total cost of sales	1,582			2,704			24,121			23,979			41,492			38,891
Gross profit	1,355			3,023			10,994			10,986			25,139			30,880
Gross margin	46.1		%	52.8		%	31.3		%	31.4		%	37.7		%	44.3		%
Total operating expenses(1)	54,323			31,069			31,656			43,244			43,190			54,640
Loss from operations	(52,968)			(28,046)			(20,662)			(32,258)			(18,051)			(23,760)
Net loss	$	(51,592)		$	(26,227)		$	(18,461)		$	(30,875)		$	(15,750)		$	(50,855)

_______________

(1)Includes stock-based compensation as follows:


	March 31, 2023		June 30, 2023		September 30, 2023		December 31, 2023		March 31, 2024		June 30, 2024
	(in thousands)
Cost of Sales	$	10	$	41	$	86	$	172	$	189	$	231
Operating expenses	4,444		4,765		6,734		10,379		9,237		22,672
Total stock-based compensation	$	4,454	$	4,806	$	6,820	$	10,551	$	9,426	$	22,903

Revenue

Hardware revenue varies based on the number of AI systems delivered and has significantly increased over time. Services and other revenue generally increased due to ongoing increases in support services as a result of our growing installed base, other than in the second quarter of 2024, when we recognized lower professional services project-based revenue.

Cost of Sales

Cost of sales has primarily varied sequentially with the volume of AI systems, professional services, and ongoing support services delivered in each quarter and has significantly increased over time.

Gross Profit and Gross Margin

Gross profit has primarily increased as a result of higher revenues. Our gross margin percentage decreased beginning in the third quarter of 2023 due to a volume price discount for a significant customer order and then began to improve over the quarters presented primarily due to lower bill of materials costs.

Total Operating Expenses

Research and development expense was the largest component of our operating expenses and was approximately 80% or higher during the quarters presented. In the first quarter of 2023, we incurred a significant expense for non-recurring engineering services, partially offset by a reduction in workforce during the third quarter of 2022. Beginning in the fourth quarter of 2023, the level of operating expenses generally increased due to higher stock-based compensation charges attributable to secondary market transactions, as well as higher personnel and other costs to support the significant step up in revenue growth in the first half of 2024.

Quarterly Trends in Non-GAAP Financial Measures

Non-GAAP Operating Loss and Non-GAAP Net Loss

The following tables set forth our non-GAAP operating loss and non-GAAP net loss (income), for each of the periods presented. See the section titled “—Non-GAAP Financial Measures” for the details of how we calculate non-GAAP operating loss and non-GAAP net loss:


	Three Months Ended
	March 31, 2023		June 30, 2023		September 30, 2023		December 31, 2023		March 31, 2024		June 30, 2024
	(in thousands)
GAAP operating loss	$	(52,968)	$	(28,046)	$	(20,662)	$	(32,258)	$	(18,051)	$	(23,760)
Add: Stock-based compensation expense	4,454		4,806		6,820		10,551		9,426		22,903
Non-GAAP operating loss	$	(48,514)	$	(23,240)	$	(13,842)	$	(21,707)	$	(8,625)	$	(857)


	Three Months Ended
	March 31, 2023		June 30, 2023		September 30, 2023		December 31, 2023		March 31, 2024		June 30, 2024
	(in thousands)
GAAP net loss	$	(51,592)	$	(26,227)	$	(18,461)	$	(30,875)	$	(15,750)	$	(50,855)
Add: Stock-based compensation expense(1)	4,454		4,806		6,820		10,551		9,426		22,903
Add: Change in fair value of forward contract liability	—		—		—		—		—		30,327
Non-GAAP net (loss) income	$	(47,138)	$	(21,421)	$	(11,641)	$	(20,324)	$	(6,324)	$	2,375

_______________

(1)Non-GAAP net (loss) income does not include the tax effects of the stock-based compensation expense adjustment because such tax effects were not material during the periods presented.

Liquidity and Capital Resources

As of June 30, 2024 and December 31, 2023, our principal sources of liquidity were cash, cash equivalents, restricted cash, and investments of $472.5 million and $152.4 million. Restricted cash consisted of $263.0 million as of June 30, 2024 and was primarily due to prepayments from G42. Our cash and cash equivalents primarily consisted of cash deposited in money market or holding accounts with financial institutions. Our investments primarily included time deposits and U.S. Treasury securities that have an initial maturity of greater than three

months but less than or equal to one year as of period end. Since June 30, 2024, we have sold an aggregate of 5,798,089 shares of our Series F-1 redeemable convertible preferred stock at a purchase price of $14.66 per share, for an aggregate purchase price of $85.0 million. Pursuant to the Preferred Stock Purchase Agreement, G42 agreed to purchase an aggregate of 22,851,296 shares of our Series F-2 redeemable convertible preferred stock (or, if purchased following the completion of this offering, shares of our Class N common stock) at a purchase price of $14.66 per share, for anticipated gross proceeds to us of $335.0 million. G42 has committed to purchase these shares by April 15, 2025.

Since our inception, we have financed our operations primarily through sales of redeemable convertible preferred stock and payments from our customers, including prepayments from G42. We had no outstanding debt as of June 30, 2024 and December 31, 2023. Our principal uses of cash in recent periods have been to fund our operations, invest in research and development and to purchase investments. As of June 30, 2024 and December 31, 2023, we had an accumulated deficit of $728.2 million and $ 661.6 million, respectively.

Our short-term and long-term liquidity requirements primarily arise from research and development efforts, as well as inventory and other working capital requirements. We have used funds to make investments, capital expenditures, and common stock repurchases from time to time. Our ability to fund our liquidity requirements will depend, in part, on our future cash flows, which are determined by our future operating performance and, therefore, subject to prevailing global macroeconomic conditions and financial, business, and other factors, some of which are beyond our control.

If we are unable to obtain funding when necessary, we may be forced to delay, reduce, or eliminate some or all of our research and development programs, product portfolio expansion, or commercialization efforts. However, we believe that our existing cash and cash equivalents and investments will provide sufficient liquidity to operate our business and fund our current and assumed obligations for at least the next 12 months. We expect to require additional capital resources in the future, or may determine to raise additional funds opportunistically. If we raise additional funds through future issuances of equity or convertible debt securities, our existing stockholders could suffer significant dilution, and any new equity securities we issue could have rights, preferences, and privileges superior to those of holders of our Class A common stock. In addition, incurring any indebtedness would increase our fixed obligations, and may include covenants or other restrictions that could impede our ability to manage our operations. We continuously evaluate our liquidity and capital resources, including our access to external capital, and our future capital requirements will depend on many factors, including working capital needs, costs to hire and retain employees in a highly competitive industry, and research and development expenditures, among others. We cannot be sure that we will be able to obtain financing on reasonable terms, or at all, in the future.

The following table summarizes our cash flows for the periods presented:


	Year Ended December 31,				Six Months Ended June 30,
	2023		2022		2024		2023
	(in thousands)
Net cash provided by (used in) operating activities	$	(78,977)	$	(164,402)	$	311,813	$	(70,185)
Net cash provided by (used in) investing activities	81,701		137,393		(9,750)		55,490
Net cash provided by financing activities	5,933		9,104		17,225		2,376

Operating Activities

Net cash provided by operating activities of $311.8 million for the six months ended June 30, 2024 primarily consisted of net loss of $66.6 million, a $300.0 million increase in a customer deposits for an advance received from G42, a $29.7 million increase in deferred revenue due to increased sales of support services, non-cash addbacks for stock-based compensation expense of $32.3 million and the change in fair value of forward contract liability of $30.3 million. Net working capital usage included a $33.9 million increase in inventory primarily due to higher

order volumes, and an $11.8 million increase in prepaid expenses and other assets, partially offset by a $20.6 million increase in other liabilities.

Net cash used in operating activities of $70.2 million for the six months ended June 30, 2023 primarily included net loss of $77.8 million, addbacks for non-cash items of $9.3 million for stock-based compensation expense and $5.7 million for depreciation and amortization, and a reduction of $9.0 million due to an increase in inventory primarily due to higher order volumes.

Net cash used in operating activities of $79.0 million for the year ended December 31, 2023 included net loss of $127.2 million, a $25.2 million increase in inventory primarily due to increased sales, a $3.0 million increase in accounts receivable due to timing of customer payments, and a $1.6 million decrease in accrued compensation and benefits related to a decrease in headcount. These cash flow uses were partially offset by non-cash adjustments of $38.4 million, primarily consisting of stock-based compensation expense, depreciation and amortization, loss on disposal of property and equipment, changes in our provision for excess and obsolete inventories, amortization of discount on investments and non-cash lease expense, a $15.2 million increase in deferred revenue due to increased sales of support services and prepaid cloud computing services, a $10.3 million increase in accrued expenses, and a $9.8 million increase in accounts payable related to timing of purchases and payments.

Net cash used in operating activities of $164.4 million for the year ended December 31, 2022 included net loss of $177.7 million, and a $37.5 million increase in inventory primarily due to an increase in sales. Non-cash addbacks in operating activities were $49.7 million, primarily consisting of stock-based compensation expense, depreciation and amortization, loss on disposal of property and equipment, non-cash lease expense, and changes in our provision for excess and obsolete inventories.

Investing Activities

Net cash used in investing activities of $9.8 million for the six months ended June 30, 2024 was the result of $152.8 million in various purchases of investments, and $11.8 million in purchases of property and equipment, partially offset by $154.9 million in maturities and sales of various investments.

Net cash provided by investing activities of $55.5 million for the six months ended June 30, 2023 was primarily the result of $194.8 million in maturities and sales of various investments partially offset by $138.9 million in purchases of various investments.

Net cash provided by investing activities of $81.7 million for the year ended December 31, 2023 was the result of $353.9 million in maturities and sales of various investments partially offset by $265.6 million in various purchases of investments, and $6.6 million in purchases of property and equipment.

Net cash provided by investing activities of $137.4 million for the year ended December 31, 2022 was the result of $355.1 million in maturities and sales of various investments partially offset by $207.2 million in purchases of various investments and $10.5 million in purchases of property and equipment.

Financing Activities

Net cash provided by financing activities of $17.2 million for the six months ended June 30, 2024 was the result of proceeds from stock option exercises of $18.3 million offset in part by $1.0 million in payments of deferred offering costs.

Net cash provided by financing activities of $2.4 million for the six months ended June 30, 2023 was the result of proceeds from stock option exercises.

Net cash provided by financing activities of $5.9 million for the year ended December 31, 2023 was the result of proceeds from stock option exercises.

Net cash provided by financing activities of $9.1 million for the year ended December 31, 2022 was the result of $10.0 million from the sale of shares of common stock and $3.0 million in proceeds from stock option exercises, partially offset by $3.9 million from the repurchase of shares of common stock.

Commitments and Contractual Obligations

Operating lease commitments. As of June 30, 2024, our operating lease commitments included data centers and corporate office leases, for which we had fixed lease payment obligations of $32.6 million, with $10.3 million to be paid within 12 months of June 30, 2024, and the remainder thereafter. See “Leases” in Note 14 to our unaudited condensed consolidated financial statements included elsewhere in this prospectus for additional information.

Purchase commitments. Our purchase commitments are primarily related to software licenses and consulting services. As of June 30, 2024, future payments related to non-cancelable commitments of $5.9 million (2024), $11.0 million (2025), and $2.7 million (2026) . See “Commitments and Contingencies” in Note 15 to our unaudited condensed consolidated financial statements included elsewhere in this prospectus for additional information.

Quantitative and Qualitative Disclosures About Market Risk

Foreign Currency and Exchange Risk

The functional currency of all our entities is the local currency of that entity’s jurisdiction. As such, we expect to be exposed to both currency transaction remeasurement and translation risk. However, we engage in a small number and immaterial amount of transactions outside of the functional currency of the reporting unit, resulting in negligible exposure to foreign currency risk. We have not hedged such exposure, although we may do so in the future. Any fluctuations in exchange rates may adversely affect our financial position, results of operations and cash flows.

Interest Rate Risk

We had cash, cash equivalents, restricted cash, and investments of $472.5 million as of June 30, 2024. Cash and cash equivalents primarily consists of amounts deposited in money market instruments with financial institutions that have an original maturity of three months or less. Our investments generally consist of time deposits, U.S. Treasury securities, and corporate debt securities. We hold cash, cash equivalents, and investments for working capital purposes. The primary objective of our investment activities is to preserve principal. We do not enter into investments for trading or speculative purposes and have not used any derivative financial instruments to manage our interest rate risk exposure. Due to the short-term nature of our investments, we have not been exposed to, nor do we anticipate being exposed to, material risks due to changes in interest rates. A hypothetical 10% change in interest rates during any of the periods presented would not have had a material impact on our historical consolidated financial statements.

Critical Accounting Estimates

We prepare our financial statements in accordance with U.S. GAAP. In preparing these financial statements, we are required to make estimates and judgments that affect the amounts and balances reported and contingencies disclosed. We evaluate our estimates on an ongoing basis, specifically including those related to revenue recognition, inventory, stock-based compensation, valuation of our common stock, and income taxes, and base our estimates on historical experience and various other assumptions that we believe to be reasonable under the circumstances, the results of which form the basis for making judgments about the carrying values of assets and liabilities.

The following critical accounting estimates require the use of significant judgment and estimation in the preparation of our audited consolidated financial statements. The accounting estimates and judgment discussed in this section are those that we consider to be the most critical in the preparation of our audited consolidated financial

statements. See “Significant Accounting Policies” in Note 4 to our audited consolidated financial statements included elsewhere in this prospectus for additional information.

Revenue Recognition

We derive substantially all of our revenue through sales of hardware and embedded software and through provision of installation, integration, and acceptance services and other technical support to our customers.

Contracts with Multiple Performance Obligations

A critical estimate required in recognizing revenue relates to contracts consisting of more than one performance obligation. When a contract with a customer consists of more than one performance obligation, we must exercise judgment in determining whether each obligation within the contract is distinct as well as in determining the relative standalone selling price to allocate to each performance obligation.

Typically, the standalone selling price is the price at which we sell a promised good or service separately to a customer. The best evidence of the standalone selling price, when available, is the price we have charged for goods or services in similar circumstances and to similar customers. If the standalone selling price is not directly observable, management estimates the standalone selling price using various observable inputs including cost-plus expected margin analysis due to the limited standalone sales history.

See “Revenue Recognition” in Note 4 to our audited consolidated financial statements included elsewhere in this prospectus for additional information.

Inventory

Our inventories consist primarily of semiconductors, memory products, and other component parts purchased from subcontractors and are carried as raw materials, work-in-progress, and finished goods. We account for inventories on a weighted average cost basis and state them at the lower of cost or net realizable value. We calculate the cost of our inventories on an adjusted standard cost basis, which approximates actual cost. We calculate net realizable value as the estimated selling price of our products less reasonably predictable costs of completion, disposal, and transportation.

We charge cost of sales on our consolidated statements of operations for provisions made to write down our inventory balances when circumstances suggest a write down is necessary – for example, when the cost basis of the inventory exceeds its net realizable value, when quantities exceed expected demand, or when inventories are otherwise deemed obsolete. The valuation of inventory involves assumptions about the likely method of inventory disposition, other-than-temporary decreases in product demand, changes in technology, competition, or customer needs, and future business and macroeconomic conditions, which require use of management judgment. If the future demand for our products is less favorable than our forecasts, the value of the inventories may be required to be reduced, which could result in additional expense to us and affect our results of operations. We do not believe there is a reasonable likelihood that there will be a material change in the future estimates or assumptions that we use to calculate our inventory reserve. However, if estimates regarding customer demand are inaccurate or changes in technology affect demand for certain products in an unforeseen manner, we may be exposed to losses or gains that could be material. If, in any period, we are able to sell inventories that had been written down to a level below the ultimate realized selling price in a previous period, related revenue would be recorded with a lower or no offsetting charge to cost of sales, resulting in a net benefit to our gross margin in that period. See “Inventory” in Note 4 to our audited consolidated financial statements included elsewhere in this prospectus for additional information.

Stock-Based Compensation

We measure stock-based awards, including stock options, RSUs, and restricted stock awards (“RSAs”), granted to employees and non-employees based on the estimated fair value as of the grant date. Stock option awards with only service-based vesting conditions are granted to employees and non-employees. The fair value of stock options

are estimated using the Black-Scholes option pricing model, which requires the input of highly subjective assumptions, including the fair value of the underlying common stock, the expected term of the stock option, the expected volatility of the price of our common stock, risk-free interest rates, and the expected dividend yield of our common stock. Changes in the assumptions can materially affect the fair value and ultimately how much stock-based compensation expense is recognized. These inputs are subjective and generally require significant analysis and judgment to develop. The fair value of RSUs and RSAs are based on the price of our common stock on the date of grant.

We recognize the fair value of each award with only service-based vesting conditions on a straight-line basis over the requisite service period of the award. For certain equity awards that have both service- and liquidity-based vesting conditions, we recognize the expense over the requisite service period if it is probable that the performance conditions will be achieved. We reassess the achievement of the performance conditions at each reporting date and adjust the stock-based compensation accordingly for such awards. Stock-based compensation expense is based on the value of the portion of stock-based awards that is ultimately expected to vest. As such, our stock-based compensation is reduced for the estimated forfeitures at the date of grant and revised, if necessary, in subsequent periods if actual forfeitures differ from those estimates.

Common Stock Valuations

The fair value of common stock underlying our stock-based awards has historically been determined by our board of directors, with input from management and contemporaneous third-party valuations. We believe that our board of directors has the relevant experience and expertise to determine the fair value of our common stock. Given the absence of a public trading market of our common stock, and in accordance with the American Institute of Certified Public Accountants Practice Aid, Valuation of Privately-Held Company Equity Securities Issued as Compensation, our board of directors exercised reasonable judgment and considered numerous objective and subjective factors to determine the best estimate of the fair value of our common stock at each grant date. These factors included:

•the results of contemporaneous valuations performed at periodic intervals by a third-party valuation firm;

•the prices, rights, preferences, and privileges of our redeemable convertible preferred stock relative to those of our common stock;

•the prices of our redeemable convertible preferred stock and common stock sold to investors in arms-length transactions;

•our actual operating and financial performance and estimated trends and prospects for our future performance;

•our stage of development;

•the likelihood of achieving a liquidity event, such as an initial public offering, direct listing, or sale of our company, given prevailing market conditions;

•the lack of marketability involving securities in a private company;

•the market performance of comparable publicly traded companies; and

•U.S. and global capital market conditions.

In valuing our common stock, the fair value of our business was determined using various valuation methods, including combinations of the income approach and the market approach with input from management. The income approach involves applying an appropriate risk-adjusted discount rate to projected cash flows based on forecasted revenue and costs. The market approach estimates value based on a comparison of our company to comparable

public companies in a similar line of business. From the comparable companies, a representative market value multiple was determined, which was applied to our operating results to estimate the enterprise value of our company.

Once the enterprise value was determined under the market approach, we derived the equity value of our company and used a hybrid method that considered both an option pricing model (“OPM”) and the probability weighted expected return method (“PWERM”) to allocate that value among the various classes of securities to arrive at the fair value of our common stock. The OPM is based on the Black-Scholes-Merton option pricing model, which allows for the identification for a range of possible future outcomes, each with an associated probability. The OPM is appropriate to use when the range of possible future outcomes is difficult to predict and thus creates highly speculative forecasts. PWERM involves a forward-looking analysis of the possible future outcomes of an enterprise, including an initial public offering as well as non-initial public offering market-based outcomes. After the equity value is determined and allocated to the various classes of securities, a discount for lack of marketability (“DLOM”) is applied to arrive at the fair value of our common stock. A DLOM is applied based on the theory that as an owner of private company stock, the stockholder has limited opportunities to sell this stock, and any such sale would involve significant transaction costs, thereby reducing overall fair market value.

In addition, we also considered any secondary transactions involving our capital stock. In our evaluation of those transactions, we considered the facts and circumstances of each transaction to determine the extent to which they represented a fair value exchange. Factors considered included transaction volume, timing, whether the transactions occurred among unrelated parties, and whether the transactions involved investors with access to our financial information.

Application of these approaches involves the use of estimates, judgment, and assumptions that are highly complex and subjective, such as those regarding our expected future revenue, expenses, and future cash flows, discount rates, market multiples, the selection of comparable companies, and the probability of possible future events. Changes in any or all of these estimates and assumptions or the relationships between those assumptions impact our valuations as of each valuation date and may have a material impact on the valuation of our common stock.

For valuations after the completion of this offering, our board of directors will determine the fair value of each share of underlying common stock based on the closing price of our common stock as reported on the grant date. Future expense amounts for any particular period could be affected by changes in our assumptions or market conditions.

Income Taxes

We are subject to income taxes in the United States and in foreign jurisdictions. To determine taxable income for financial statement reporting purposes, we make certain estimates and judgments. These estimates and judgments are also applied when calculating certain tax liabilities and assessing the recoverability of deferred tax assets, each of which reflect the expected future tax consequences of events that have been included in the audited consolidated financial statements.

We apply the asset and liability method in accounting for deferred tax balances. Under this method, we determine deferred tax balances on the basis of differences between the financial statement and tax bases of assets and liabilities by using enacted tax rates in effect for the year in which the differences are expected to reverse. We recognize the effect of a change in tax rates on deferred tax balances in income in the period that includes the enactment date.

We recognize deferred tax assets only to the extent we expect they are more likely than not to be realized and consider all available positive and negative evidence in making such a determination, including future reversals of existing taxable temporary differences, projected future taxable income, tax planning strategies, and results of recent operations. If we determine that a deferred tax asset is recorded in an amount that exceeds the amount we ultimately expect to realize, we will make an adjustment in the form of a valuation allowance, which could reduce the

100

provision for income taxes. In determining the need to establish or maintain a valuation allowance, we consider (i) carryback of net operating losses to prior years; (ii) future reversals of existing taxable temporary differences; (iii) viable and prudent tax planning strategies; and (iv) future taxable income exclusive of reversing temporary differences and carryforwards.

We record uncertain tax positions using a two-step process. We first determine whether it is more likely than not that the tax positions will be sustained on the basis of the technical merits of the position. For those tax positions that meet the more-likely-than-not recognition threshold, the largest amount of tax benefit that is more than 50% likely to be realized upon ultimate settlement with the related tax authority is realized.

JOBS Act Accounting Election

We are an emerging growth company, as defined in the JOBS Act, and, for so long as we continue to be an emerging growth company, we may take advantage of certain exemptions from various reporting requirements that would have been applicable were we a public company that was not an emerging growth company. Such exemptions include, but are not limited to, the exemption to comply with the auditor attestation requirements of Section 404 of the Sarbanes-Oxley Act, reduced disclosure obligations regarding executive compensation in our periodic reports and proxy statements, the exemption from holding a non-binding advisory vote on executive compensation, and the exemption from stockholder approval of any golden parachute payments not previously approved. In addition, pursuant to Section 107 of the JOBS Act, as an emerging growth company, we have elected to take advantage of the extended transition period for complying with new or revised accounting standards until those standards would otherwise apply to private companies. If we cease to be an emerging growth company, we will no longer be able to take advantage of these exemptions or the extended transition period for complying with new or revised accounting standards.

Recent Accounting Pronouncements

See Note 3 to our audited consolidated financial statements included elsewhere in this prospectus for additional information.

101

BUSINESS

Our Mission

Company Overview

102

Industry Background

Over the past 40 years, the computer industry has followed a clear pattern: as major new computational workloads with distinct characteristics emerged, so too have new compute architectures. For example, the general-purpose needs of personal computing led to the x86 CPU. The low-energy needs of mobile devices resulted in the widespread adoption of ARM. Advancements in graphics demanded greater rendering parallelism, resulting in the creation of the GPU. With each new compute paradigm, technologists first attempted to adapt existing compute architectures to these workloads. But in each case, a new purpose-built architecture was ultimately needed to unlock the potential of the new paradigm.

We believe this pattern is continuing with the rise of AI – the next major compute workload, with its own unique computational demands. In 2023, IDC estimated that the worldwide economic impact of GenAI would be close to $10 trillion by the end of 2033. This growth has been accelerated by the emergence of GenAI, a new class of powerful AI models that can create new content and reason across broad domains and multiple data types. These breakthrough capabilities translate to tremendous potential value creation and have driven rapid GenAI adoption. For example, ChatGPT took only five days to reach one million users (a feat that took Instagram 10 weeks and Twitter two years), and similarly, Meta’s open-source Llama 2 model attracted hundreds of thousands of AI developers within days of its release.

The Computational Demands of Training and Inference

The explosion of GenAI in the past 18 months has been likened to the “iPhone” moment, marking a pivotal shift in the industry. GenAI is already consuming compute resources at an unprecedented rate, faster than any other workload in history. As more powerful models are created and new use cases for GenAI are brought to market, we believe that demand for powerful and efficient AI compute solutions will continue to grow rapidly. The recent growth in the U.S. data center market supports this belief. Over the course of 2023, demand for U.S. third-party data centers has grown by nearly 50%, with AI driving significant growth. Major companies like Microsoft, Google, and Amazon are driving 500+ MW deals, whereas in 2015, data center deals were commonly in the 5 MW range.

103

Model Optimization Is Important, but As You Optimize, Models Continue to Evolve

For training, the compute required is a function of a model’s size (number of parameters) and the amount of data used (number of tokens). The most capable GenAI models today have trillions of parameters and are trained on trillions of tokens. These models demonstrate superior accuracy and capabilities compared to small models due to their greater capacity to discern nuanced patterns from the data. As the industry has pushed to achieve greater AI capabilities, the size of GenAI models has grown, and we expect this trend to continue.

The increase in model and data sizes has led to a dramatic surge in computational demand. As shown below, requirements to train GenAI models have grown 40,000-fold in just five years. Today, training for GenAI requires enormous GPU clusters, sophisticated engineering teams, and months of time for a single run. A training run can cost over a hundred million dollars, and improving the model and keeping it fresh with new data requires additional fine-tuning and regular re-training.

104

Illustration of Growing Compute Requirements Over Time (ExaFLOPS to Train LLMs)

For inference, the required compute is a function of model size, user demand, and time spent on inference. Larger models use more computational resources, and each user request also increases the compute need. Dedicating more compute resources to inference also generally results in better output quality, as more compute enables deeper chains of thought and reasoning. These factors contribute to significant and ongoing operational costs.

We expect the demand for inference to grow, especially as larger and more capable models become more widely adopted. There is currently a direct tradeoff between model capability and responsiveness of user experience because the largest and most capable models demand more inference compute and therefore run more slowly during inference. As inference speeds improve, larger GenAI models can be deployed in a responsive manner, expanding their use in real-time applications and agent-based systems. Faster inference can enable multiple inference requests to be made to the same or different GenAI models, with each request building upon the results of the previous request, all within the time it previously took to execute a single inference request.

Advancements in model reasoning capabilities are equally important. For instance, OpenAI’s o1 model, released in September 2024, uses chain of thought—a reasoning process that leverages multiple intermediate inference steps—to solve complex coding, math, and science problems that were unsolvable by earlier models. We expect that increasing compute resources used for inference by orders of magnitude to enable deeper chains of thought will continue to yield significant improvements in output quality for problems that require complex reasoning. These reasoning improvements enable models to tackle more sophisticated tasks, broadening their range of use cases. We believe that as both inference speed and reasoning capabilities advance, GenAI models will support more demanding applications, thereby expanding the market opportunity for inference.

As of June 2024, we estimate that 40% of the AI data center market is attributable to inference, and we expect this to grow as more AI applications and products are brought to market.

Existing Compute Architectures Are Fundamentally Limited for GenAI Training and Inference

GPUs, though better than CPUs for AI workloads, face fundamental limitations when processing the unique characteristics of large GenAI models. In graphics, the calculation for each pixel is often independent of every other pixel. This means that for large graphics workloads, many GPUs can easily execute on separate parts of the rendering problem, without a need for high levels of interdependent communication.

The AI workload is different. GenAI models are complex, interconnected compute graphs that require the constant communication of intermediate calculations to train. This requires a high amount of data movement to and from memory and across cores. Similarly, during inference, these models generate outputs sequentially – each

105

dependent on the previous output – requiring the full model to be constantly moved in and out of memory to produce successive outputs, and again requiring massive data movement between cores and memory. GPUs face inherent scalability and complexity challenges when faced with the distinct, communication-heavy requirements of GenAI workloads.

For Training – Individual GPUs Are Too Small, and Scaling to Many GPUs is Highly Inefficient

Large GenAI models far exceed the memory and processing limits of a single GPU. For example, to train GPT-3, which OpenAI has said required ~3.14 x 10^23 floating point operations, it would take a single NVIDIA H100 more than eight years of running at peak theoretical performance to train the model. Recent models like GPT-4 and Gemini are over 10 times larger in parameter size than GPT-3. Consequently, training a large GenAI model on GPUs in a tractable amount of time requires breaking up the model and calculations, and distributing the pieces across hundreds or thousands of GPUs. These hundreds or thousands of GPUs then need to constantly communicate with each other across a network, creating extreme communication bottlenecks and power inefficiencies.

For many organizations, distributed programming is one of the highest barriers to entry for leveraging GenAI.

Large GenAI Models Must Be Divided and Coordinated Across Thousands of GPUs, Leading to Communication Bottlenecks and Developer Complexity

For Inference – GPU Efficiency is Low and Limited by Memory Bandwidth

106

relatively low memory bandwidth, meaning that the rate at which they can move information from off-chip HBM to on-chip SRAM, where the computation is done, is severely limited. This leads to low performance as GPUs cores are idle while waiting for data – they can run at less than 5% utilization on interactive generative inference tasks. Low utilization and limited memory bandwidth impact the responsiveness and throughput of GPU-based systems and hinders real-time applications for larger models. This can limit the capability and adoption of emerging inference applications which are especially latency-sensitive, such as code generation and multi-turn AI agents that need to string together multiple calls to different GenAI models. This inefficiency also necessitates larger GPU deployments and dramatically drives up the cost of inference.

GPU companies have tried to address these challenges, but the issues of small core count, memory size, and memory bandwidth are fundamental hardware limitations that persist. Interconnect technologies like InfiniBand, PCIe, and NVLink are limited by their physical interfaces, and moving data across them is thousands of times slower and more power-hungry than keeping and moving the data on silicon. Software libraries intended to simplify distributed computing still require developers to manage complex parallelism strategies and extensive codebases. Realizing the physical challenges of repurposing small chips for a large compute problem, the GPU industry has recently announced new multi-chip packaging techniques, but these also yield only minimal expansions of GPU chip size.

Our Solution

We believe Cerebras has built the world’s fastest commercially available AI training and inference solution. Our dedicated AI hardware and software platform is powered by the Cerebras Wafer-Scale Engine – a processor the size of an entire silicon wafer that brings more on-chip compute, memory, and bandwidth resources together than any other commercially available processor in the semiconductor industry. A single WSE replaces a cluster of GPUs, reducing the time-consuming, power-hungry movement of data, removing the need for complex distributed programming, and providing exceptional compute speed. Compared to the leading 8-way GPU system of the same generation, many of our customers have achieved over 10 times faster training time-to-solution. Our inference offering is over 10 times faster than GPU-based solutions from top CSPs, as benchmarked on leading open-source models. A single CS-3 system delivers three times more compute per watt than the leading 8-way GPU system.

Our solution consists of the following elements:

Cerebras Wafer-Scale Engine (WSE). At the heart of our solution is the Cerebras Wafer-Scale Engine, the largest chip ever sold. Our third generation WSE, the WSE-3, is 57 times larger than the leading commercially available GPU (NVIDIA H100) and has 52 times more compute cores, totaling 900,000. It features 880 times more on-chip memory (44 gigabytes SRAM) and 7,000 times more memory bandwidth (21 petabytes per second) than the leading commercially available GPU. Developing the WSE required overcoming decades-long-standing industry challenges, including inter-die connectivity, yield optimization, efficient packaging, power management, and advanced cooling.

107

Cerebras Wafer-Scale Engine 3 Versus NVIDIA H100 Size Comparison

Image is illustrative; scale is approximate.

Cerebras Wafer-Scale Engine 3 Versus NVIDIA H100 Capability Advantage

The immense scale of the WSE delivers significant acceleration, efficiency, and simplicity for AI training and inference.

108

For Training. Each Cerebras WSE has enough compute and on-chip memory to run even the largest, multi-trillion parameter GenAI models on a single chip, thus avoiding the complexities of chip-to-chip data movement. This is unlike GPUs, which require users to fragment large models across multiple processors and deal with complex inter-GPU communication.

To further speed up training time-to-solution, Cerebras users can simply add more WSEs to the problem. Since each WSE can fit the whole model, multiple WSEs can accelerate model training simply by having each chip independently work on a subset of the training data. Because the model is never split across the WSEs, no complex model distribution or carefully orchestrated communication is needed. The elegance of this architecture is designed to allow users to effortlessly increase training speeds with near-linear performance scaling as more WSEs are added.

For Inference. Wafer-scale integration keeps all critical data on-chip and close to compute cores, resulting in 7,000 times more memory bandwidth than the leading GPU solution. This allows the WSE-3 to deliver over 10 times lower latency for real-time GenAI inference, which means an inference response that takes 10 seconds to generate on GPU-based platforms from top CSPs takes only one second to complete using the Cerebras solution. The WSE-3 can accomplish this speedup at vastly lower power consumption, as retrieving one bit of data from on-chip SRAM on 5nm silicon requires only 1% the energy that is needed to do the same from off-chip HBM.

Cerebras Inference Is the Fastest Inference Solution on Llama 3.1 8B

109

Cerebras Inference Is the Fastest Inference Solution on Llama 3.1 70B

Cerebras System (CS). The CS AI computer system houses the WSE and delivers innovative power and cooling to the chip. Our third generation CS (“CS-3”) delivers three times more compute per unit power than the leading 8-way GPU system (NVIDIA DGX H100). This compact AI powerhouse is designed to easily integrate into standard data centers – it occupies less than half a standard rack (16RU/27 inches tall), and connects into the network via standards-based 100G Ethernet.

The CS-3 System, with Innovative Power and Cooling, is Designed to Fit in a Standard Data Center Rack

110

Cerebras AI Supercomputer. The Cerebras AI Supercomputer is designed to streamline scaling up to 2,048 CS-3 systems for maximum AI acceleration, with more efficiency and simplicity than scaling up to large GPU clusters. Unlike with GPUs, where users must break up and distribute their model across many chips, thereby introducing the need to communicate across those compute elements, each WSE can run the model in its entirety. In turn, a Cerebras AI Supercomputer only needs to spin up another copy of the full model on each additional CS system to process the training dataset more quickly. This enables a near-linear performance increase as CS systems are added to a problem, takes only seconds to configure, and does not incur the overhead or complexity of heavy inter-chip, inter-system communication.

Our scalable execution model is designed to simplify the development workflow for large GenAI training. We designed our AI Supercomputer to enable users to elastically scale workload computing resources up to 256 exaFLOPS (2,048 CS-3 systems) just by changing a single number in their code, allowing users to program as if for a single powerful device. Training a GPT-3 sized model on Cerebras, for example, uses 97% fewer lines of code compared to on clusters of GPUs, greatly accelerating the speed of AI model developments for larger-scale models. The combined power and scaling simplicity of the Cerebras AI Supercomputer provides industry-leading training and inference speeds for even the largest and most complex GenAI models, while obviating the need to invest thousands of programmer hours into distributed programming work.

Zero-Complexity Scaling on Cerebras Versus GPUs: Training a GPT-3 Sized Model With Cerebras Requires 97% Fewer Lines of Code Compared to on Clusters of GPUs

111

A Cerebras Supercomputer Deployment at One of Our Colocation Facilities

Cerebras Software Platform (CSoft). Our proprietary software platform, CSoft, is core to our solution and provides intuitive usability and improved developer productivity.

CSoft seamlessly integrates with industry-standard ML frameworks like PyTorch and with popular software development tools, allowing developers to easily migrate to the Cerebras platform. CSoft automatically saves model outputs in a standard “checkpoint” format, enabling users to resume training or inference on model work started on other hardware platforms. This checkpoint compatibility extends to open-source models available in the popular HuggingFace repository.

CSoft is co-designed with the wafer-scale hardware architecture to provide programming efficiency and simplicity. It allows ML users to accelerate training and inference on models of any size, scaled across any configuration of the Cerebras AI Supercomputer, just by changing one number in a configuration file, simulating a single-device programming experience without the complexities of distributed programming. This drastically reduces operational overhead and speeds up developer iteration time and business impact.

Cerebras Inference Serving Stack. The Cerebras Inference Stack is built on top of CSoft and is designed to allow customers to easily deploy even the largest GenAI models at industry-leading inference speeds. With the Cerebras Inference API, developers are able to easily point their applications to popular or custom models, just by changing their API key. This simple API, which closely matches the interface design of the popular OpenAI API, is designed to facilitate rapid developer adoption and ease of use. We believe there will be stickiness to the Cerebras platform as customers experience industry-leading inference performance for their interactive applications. Our serving software automatically handles system-level optimizations for our inference solution and is designed to enable low latency and high cost effectiveness.

112

AI Model Services. Our AI model services further amplify speed to solution. Our team of AI practitioners helps customers design research experiments, train models, and optimize processes designed to achieve the fastest time-to-solution. These services complement our advanced hardware and software platform, providing an end-to-end solution for rapid and efficient AI development and deployment. We believe we are one of a select number of companies in the world that has trained high-quality, large GenAI models on massively parallel compute clusters. We have contributed state-of-the-art models into the open-source community and have published widely on AI methodology and practice. Our team’s work spans across model architectures, parameter sizes, and data types – including text, time series data, code, biological and other sequence data, electronic healthcare records, medical imagery, radio frequency, and radar data. Cerebras AI researchers have trained LLMs in English, Arabic, Spanish, Japanese, and Catalan.

We excel at helping customers translate AI potential to business impact. Our AI experts guide customers in developing custom AI strategies, designing and building AI models, and applying cutting edge AI techniques to achieve high-quality results, thereby delivering downstream business impact. We augment our customers’ existing teams with specialized AI capabilities, and the models we build together routinely beat the existing state of the art, providing customers with a distinct competitive advantage.

Summary of key customer benefits include:

Enables over 10 times faster training time-to-solution compared to leading 8-way GPU systems of the same generation, as reported by many of our customers. This dramatically accelerates AI model time-to-solution, enabling businesses to test ideas faster, iterate more quickly, and bring next-generation GenAI-powered products and services to market, faster and more cost effectively.

Delivers over 10 times faster GenAI inference compared to GPU-based solutions from top CSPs, as benchmarked on leading open-source models. The WSE’s massive on-chip SRAM capacity keeps the vast majority of memory-to-compute communication on-silicon, thereby avoiding the memory bandwidth bottleneck faced by GPU-based solutions. Our resulting ultra-low latency delivers industry-leading inference speeds and real-time responses back to users, even on large, cutting-edge GenAI models. Ten times more speed compared to GPU-based solutions from top CSPs also allows developers to make ten times more inference calls in the same amount of time. This supports a new level of model capability delivered by techniques like multi-step inference and agentic AI flows, which leverage more inference calls to produce higher reasoning capability for more complex tasks in domains such as coding, math, and science applications.

End-to-end solution. Cerebras offers a unified platform purpose-built to accelerate fundamental compute characteristics of both AI training and inference. While many other emerging AI chips companies have chosen to specialize in only one phase to allow for optimization tradeoffs across the axes of compute, memory, memory bandwidth, and simplicity of use, Cerebras excels across these key dimensions, made possible by wafer-scale integration. This allows customers to swiftly transition from training to fine-tuning to deploying high-quality GenAI models on the same platform, eliminating the need for investing in and maintaining separate hardware infrastructure.

Zero distribution complexity. The biggest challenge in training on GPU clusters is the complexity of distributed programming across multiple devices and systems. This often requires large engineering teams weeks or months of optimization, and results in limited performance scaling. Using Cerebras, users can effortlessly run a GenAI model of any size. It takes no additional code to achieve automatic near-linear performance scaling across the nodes of a Cerebras Supercomputer, unlike the 20,000+ lines of distributed programming code required to scale large models across similar memory and compute resources on large GPU clusters.

Low migration cost. Our proprietary CSoft platform integrates seamlessly with familiar ML frameworks and tools, like PyTorch, eliminating the need for AI teams to learn new languages or adapt to new development environments and easing the transition from other hardware platforms.

Power and operational efficiency. Cerebras outperforms GPU systems in power efficiency due to both hardware and architectural advantages. Wafer-scale integration allows the majority of AI workload communication

113

to remain on-chip, significantly reducing data movement distance and power consumption; moving a bit of data on the WSE-3 takes less than 1% of the energy needed to do the same over off-chip GPU interconnects. This drives significant operational cost savings (CS-3 draws one-third of the power of the leading 8-way GPU system) and streamlined management for organizations deploying AI at scale.

Expert-led model training and AI integration services. We offer expert-led foundation model training, fine-tuning, and retrieval-augmented generation services to customers. Our team provides guidance on cutting-edge AI methods that work on top of our hardware, assisting customers to derive the maximum value from their AI investment.

Examples of Customer and Industry Impact

Our customers find fundamental value in the simplicity of the Cerebras solution. It unlocks breakthrough business and scientific use cases by removing limitations in development time, programming complexity, and runtime speed. For example, a leading pharmaceutical company trained an epigenomic language model using Cerebras and noted that the training speedup afforded by the Cerebras system enabled them to explore architecture variations, tokenization schemes and hyperparameter settings in a way that would have been prohibitively time and resource intensive on a typical GPU cluster.

As we continue to expand our go-to-market capabilities and more customers recognize the benefits of our solution we expect our pipeline to grow. The tangible benefits of our products are evident in the strong gains in performance and time-to-value experienced by our customers across various industries:

A global AI technology group (G42) reduced AI model convergence time for a 13 billion parameter Arabic language model from 68 days on a GPU cluster to just four days using Cerebras, improving training time by 17 times while also achieving higher model quality. They furthered this work by subsequently training a state-of-the-art bilingual Arabic-English 30 billion parameter model on 1.6 trillion tokens, now being served on the Microsoft Azure AI Model Catalog.

A leading pharmaceutical company reduced training time on an epigenomic language model from 24 days to just 2.5 days using Cerebras, achieving a 90% reduction in time-to-solution. This resulted in a new model that was the first of its kind to combine both DNA data and epigenetic state data of chromosomes for 127 different cell types. This model exceeded the state-of-the-art on multiple benchmarks and is being used by the customer to facilitate further study of gene regulation and function.

A major U.S. healthcare provider leveraged Cerebras AI services and compute to develop and train an industry-leading medical model in eight weeks, beating open-source medical model benchmarks, and upskilling their in-house AI team through partnership with our AI expertise.

A U.S. national lab trained a genome-scale foundation model, predicting the evolution of emergent variants of SARS-CoV-2 virus, 19 times more quickly on a single Cerebras system compared to the leading 8-GPU system of the time (152 times faster than a single GPU).

The same national lab trained an AI model using Cerebras that could not converge on any number of GPUs. This model won the 2022 Gordon Bell Special Prize for High Performance Computing-Based COVID-19 Research.

Another U.S. national lab showed 179 times faster performance on a molecular dynamics simulation using a single Cerebras system, compared to the entirety of Oak Ridge National Laboratory’s Frontier, the largest

114

supercomputer in the world, with more than 37,000 GPUs and reportedly costing approximately $600 million to build.

A U.S. government agency is leveraging our real-time inference capabilities to create the world’s first, large-scale, virtual radio frequency simulation environment for developing, training, and testing advanced radio frequency systems.

Our Business Model

On Premise. Our direct-to-customer sales force sells our AI Supercomputers to leading organizations who seek maximum control over their data and their AI infrastructure, fulfilling their needs for high-performance AI compute on premise. For on-premise use, we offer deployment services, as well as a subscription to an ongoing stream of software updates and upgrades. Our on-premise AI Supercomputers support both training and inference. They can be configured either for both workloads, or to be further optimized for only one, depending on our customers’ needs.

Cloud-Based Computing Services. We sell Cerebras solutions via our cloud offering as well as via the Condor Galaxy Cloud owned by Group 42 Holding Ltd (together with its affiliates, “G42”), our partner CSP. Our cloud solutions provide customers fast and flexible access to our powerful AI acceleration hardware, with payment based on time or work (number of models). This offering gives our customers the ability to train LLMs with extraordinary speed and deploy them for inference at ultra-low latencies, all without the complexity or time needed to build and manage on-premise infrastructure.

We also provide professional services to assist customers throughout the AI workflow. From developing strategy, designing and building models, to deploying and maintaining the final models either on-premise or in the cloud, we help customers achieve optimal outcomes.

Our Market Opportunity

We participate in a large and growing AI market. Enterprises, research organizations, and governments alike are developing AI initiatives to rapidly evolve and drive efficiencies across their entities. As estimated by McKinsey in 2023, GenAI use cases may add up to $4.4 trillion to the global economy on an annual basis. Our full suite of AI computing solutions addresses use cases for training, inference, software, and expert services. We estimate the TAM

115

for our AI computing platform to be approximately $131 billion in 2024, growing to $453 billion in 2027, a 51% CAGR.

This TAM is comprised of the following core markets:

AI Compute for Training. Demand for GenAI solutions has grown significantly across virtually all use cases. In a recent Gartner, Inc. poll of more than 1,400 executive leaders, 45% reported that they are in piloting mode with GenAI, and another 10% have put GenAI solutions into production (Source: Gartner®). Significant capital is being dedicated to evaluating GenAI solutions and the impact they could have on everything from productivity, efficiency, to customer support. As businesses continue to evaluate and deploy solutions, we believe the market for training new models will continue to grow. Based on market estimates in Bloomberg Intelligence research, our estimate of the TAM for AI Training Infrastructure is $72 billion in 2024, growing to $192 billion in 2027, a 39% CAGR. This segment of the market includes AI Infrastructure-as-a-Service.

AI Compute for Inference. While GenAI training is essential to developing models that are powerful and accurate, we believe inference is the next phase of the ongoing wave of AI disruption. As more enterprises develop models and start to deploy their models in applications at scale, the need for high performance and efficient inference is becoming critical to fully realize the commercial potential of ML. The demand for inference compute is driven by the number of users and frequency that inference is used. We believe that the inference opportunity is enormous, as the market is in the early phases in its adoption cycle. Our estimate of the TAM for AI Inference is $43 billion in 2024, growing at an estimated 63% CAGR to $186 billion in 2027.

Software and Expert Services. Based on market estimates in Bloomberg Intelligence research, our estimate of the TAM for GenAI Software and Services is $16 billion in 2024, growing to $75 billion in 2027, a 67% CAGR.

Customers and Ecosystem

Our customers include leading enterprises, Sovereign AI initiatives, cloud service providers, government agencies, and research institutions at the forefront of AI and at the intersection of AI and HPC. The majority of our revenue for 2023 is from the sale and deployment of our AI systems, but we expect significant traction in onboarding new customers onto our cloud offering. The scalability and flexibility of our solutions enables customers to leverage both hardware and cloud solutions concurrently, allowing them to adapt their approach as their needs grow. Our customers also leverage our AI model services to achieve state-of-the-art AI results. We sell our products directly to our customers and in some cases, we rely on fulfillment partners for deployment and logistical purposes.

Additionally, we share our innovations and research on open-source platforms like HuggingFace, where our published models and datasets have garnered significant traction. An example is our open-source BTLM Model, which was downloaded over one million times. This broad engagement keeps us at the cutting edge of AI technology and fosters a collaborative relationship with the ML community.

Competitive Strengths

We engineer our purpose-built AI systems to be faster and more capable than any other solution available in the market today. We believe our design is capable of meeting today’s needs and is scalable to address tomorrow’s challenges. Our competitive strengths include:

The world’s first and only wafer-scale chip in the market. Achieving wafer-scale success has been a goal of the semiconductor industry for decades, and we are the first company to overcome and solve the fundamental technical challenges required to productize a wafer-scale chip, including inter-die connectivity, yield, packaging, power, and cooling. Our wafer-scale chip architecture combines a massive amount of fast, on-chip memory, directly

116

next to vast compute resources on the same piece of silicon. It is capable of running the largest GenAI models on a single wafer-scale chip instead of tens, hundreds, or thousands of competitor devices. This eliminates the need for distributed computing while running large GenAI models, enabling AI developers to use up to 97% less code when working with large models on our platform compared to on clusters of GPUs and greatly accelerating the speed of AI model development for larger-scale models. The details of the hardware are completely abstracted away from the AI developer, allowing them to focus on model development and deployment, which they are empowered to do faster than on any other system.

Full system solution that is easy to deploy and efficient to operate. Built with our wafer-scale chip at its core, we design full systems that seamlessly fit into data center racks. Unlike many other players in this industry, we both design the chip and the system around it. By designing the processor and the system together, we are able to address thermal and power delivery challenges by optimizing the full system with our proprietary power delivery and cooling technology. This allows us to keep the system operating efficiently, optimizing the energy consumption and underlying operating costs for our customers.

Comprehensive software suite that leads to ease of adoption and shortens time to deployment. Developed over eight years by our world-class team, the Cerebras Software platform allows for seamless programmability of AI models through our integration with PyTorch, the industry-standard ML framework for AI developers. Our tools allow users to bring models that have been trained on other hardware onto our platform for training or for inference. Likewise, users can train models on our platform and deploy those models for inference elsewhere. This increases ease of adoption and shortens time-to-deployment.

Our AI platform addresses both training and inferencing markets. We are strongly positioned to provide an AI platform that is powerful, efficient and easier to deploy for both training and inferencing workloads. For training, our AI platform enables customers to effortlessly and swiftly use the most advanced GenAI models available on the market. It allows any single user to effectively develop and refine models with multi-billion-to-trillion parameters, without the need for specialized software frameworks or help from distributed computing experts. For inferencing, given our advantages in memory capacity, our AI platform delivers industry-lowest latencies and high generation throughput. This helps our customers to unlock cutting edge performance for ultra-low-latency use cases leveraging GenAI, such as real-time, interactive user applications and intelligent AI agents that often require processing data across multiple GenAI models and APIs.

Scalable architecture. We have developed our solutions to support models and data sets of large and varying sizes. Our current generation CS-3 is designed to support models with up to 24 trillion parameters, much larger than even today’s state-of-the-art GenAI models. Our platform is designed to seamlessly scale from 1 to 2,048 CS-3 systems, forming an AI supercomputer that is even further differentiated by our proprietary interconnect and memory technology. This enables customers to seamlessly increase compute resources from small-scale experiments to large-scale deployments. Our future-ready design enables us to support the most demanding AI applications and are designed to ensure that our customers’ investments remain forward-compatible as new advancements in AI emerge. We expect this benefit to be crucial for long-term customer satisfaction and retention.

AI model services help customers translate AI potential to business impact. We provide customers with AI model services to help them develop solutions that are customized to meet their needs and help them realize the full value of their AI investments. These services include model selection, data preparation, training, and solutions integration.

World-class AI talent with a proven track record of innovation and execution. We have assembled a world-class team of industry leaders in integrated circuit design, processor architecture, power delivery, cooling, system engineering and software. Over the last five years, we have introduced three generations of our WSE, each time achieving two times the performance of its predecessor, and bringing new IC, power, and cooling technologies to bear. This track record of innovation and execution demonstrates our singular focus on solving the critical performance challenges of AI compute, and systematically improving our product to achieve more performant AI compute. Our research and development organization was over 80% of our total headcount as of December 31, 2023, with approximately three-quarters of research and development headcount consisting of software engineers.

117

Growth Strategies

Increase sales to our existing customers. We have established a strong land-and-expand track record with our existing customer base, which comprises leading enterprises and research institutions looking to harness the potential of AI. We intend to deepen these relationships by expanding our product and service offerings tailored to their evolving needs. Our strategy focuses on demonstrating the value proposition of our solutions through initial engagements, cross-selling complementary products, and identifying new use cases within existing customers. Our deep technical expertise and customer-focused professional services consistently deliver value beyond initial expectations. This success helps drive increased investment from existing clients.

Expand our customer base. We believe a core component of our growth strategy centers on expanding our diverse customer base across industries and applications. We plan to aggressively pursue opportunities in relevant sectors such as healthcare, pharmaceutical, biotechnology, government, financial services, sovereign, and energy, to name a few, where our AI acceleration capabilities can address critical computational bottlenecks. We will seek to drive this expansion by focused sales and marketing initiatives, highlighting the transformative potential of our technology with targeted use cases. We intend to leverage our existing success stories and strategic partnerships to both bolster our credibility within new markets and establish key channels for customer acquisition. For example, we entered into a partnership with a world-leading European AI company to create sovereign and secure AI solutions for governments and enterprises resulting in a first-time buy of over $4 million. Additionally, we will continue investing in the development of tools and resources that streamline the onboarding experience for new customers to enable seamless adoption of our solution.

Further penetrate into the rapidly growing inference market. We recently launched our inference solution. A large opportunity for our growth is to make our inference solution widely available. The immense amount of memory bandwidth and capacity on our chip allows us to deliver significantly lower generation latency and higher generation throughput over GPUs. By making API-based inferencing available through our cloud offering, we could significantly accelerate our adoption into inferencing use cases. Based on the inference TAM of $43 billion in 2024, growing at an estimated 63% CAGR to $186 billion in 2027, we believe our expansion into inference will be a significant growth driver.

Benefit from opportunities in large adjacent AI and compute-intensive markets. We intend to leverage our differentiated solution to address ever-evolving computing challenges in emerging end markets and applications. We are actively enabling applications in fields like life sciences, materials science, and financial modeling, where our cutting-edge AI computing solutions can unlock new discoveries and solve complex problems. Geographically, our strategy includes deepening our investment in partnerships with large Sovereign AI initiatives and new markets to accelerate AI adoption. For instance, AI’s economic value in the Middle East’s Gulf Cooperation Council countries is projected to be as much as $150 billion, or approximately 9% of such countries’ combined gross domestic product, according to a McKinsey report issued in 2023. We believe our strategic partnership with G42 positions our solutions at the forefront of innovation and expansion in these emerging AI markets.

Accelerate our existing product roadmap as well as develop new products for emerging use cases. We believe continued technological innovation is a cornerstone of our future growth. Our substantial investment in research and development fuels advancements to further differentiate our wafer-scale technology and develop our software to optimize performance, accessibility, and ease of use of our platform. Our close partnerships with leading AI researchers, industry pioneers, and the open-source community grant us early insights into cutting-edge market developments. Moreover, as AI infrastructure requirements scale, we expect emerging use cases to require new products with added functionalities to solve data, networking, and memory bottlenecks. With our continued focus on innovation, we intend to develop and introduce new products and form factors that will enable us to service a larger portion of our market opportunity.

118

Advance product adoption by proliferating cloud deployment of our AI solutions. We intend to accelerate our growth by expanding access to our revolutionary AI systems through cloud deployment. We expect this strategy to make our unique, high-performance computing capabilities available to a significantly broader customer base. Cloud solutions reduce the upfront capital investment required for customers, enabling more rapid experimentation and wider adoption of our technology. By offering our technology as a cloud service, we can streamline workflows for data scientists and machine learning engineers, allowing them to focus on innovation instead of infrastructure management.

Our Technology

We believe we have built the fastest commercially available AI training and inference solution in the world. Our solution combines our industry-leading wafer-scale engine with a supercomputer system architecture and a co-designed software suite.

Every component of our platform is specifically designed to meet the computational needs of AI workloads. We combine massive computational resources, fast on-chip memory, and high-bandwidth interconnect to simplify and accelerate AI training and inference. Our product offerings are available through a unified platform that is both easy to use and simple to deploy.

Our platform’s exceptional performance is driven by a portfolio of fundamental technological innovations, that resolve challenges in the computing industry that had been open for decades.

WSE-3 – The World’s Largest, Most Powerful Commercially Available Chip

Throughout the history of computing, the drive towards larger chip sizes with more integrated components has consistently produced significant performance and efficiency improvements. This has driven the entire semiconductor industry towards Moore’s Law, which predicts that the number of transistors in an integrated circuit will double every two years. However, the unprecedented demand for AI computing has underscored the need for improvements that go beyond the traditional cadence of compute advancements.

To meet this demand, we have designed the Cerebras Wafer-Scale Engine – the largest chip ever sold, and which has broken Moore’s Law. With four trillion transistors, the WSE-3 achieves a milestone that Moore’s Law did not predict would occur until 2034. This leap marks a significant advancement in chip design and manufacturing, setting new standards for computational power and efficiency.

119

“Moore’s Law” Over the Past 40 Years

The white line shows the Moore’s Law trend of the largest processor chips in the industry. The orange line shows the new trajectory created by the Cerebras wafer-scale technology.

Wafer-scale integration is a unique capability. In the entire history of the computer industry, only Cerebras has delivered a wafer-scale product to market. To achieve wafer-scale integration, we invented and productized key chip design technologies:

•Multi-die interconnect. Traditionally, small die—regions of silicon containing an integrated circuit—are replicated independently on a silicon wafer and then cut up (“diced”) into small, separate chips. We have developed the technology to connect these otherwise independent die together at the wafer level, at the semiconductor fabrication plant. The inter-die connectivity uses a special cross-reticle connection that is integrated into the overall fabrication process.

•Fault-tolerant architecture. A primary factor in the commercial viability of a semiconductor is the yield. The fabrication process introduces tiny defects – and chips with these defects are discarded and considered yield loss. Recognizing that large chips are more prone to defects than small chips, we invented new techniques designed to withstand defects, rather than seek to avoid them. Combining architectural innovations with insights from adjacent industries, we used a combination of redundant compute cores and redundant routing to address the yield problems. The wafer behaves like a hyper-scale data center, using a “fail in place” mechanism to handle the inevitable failures at scale. Flaws are designed to be recognized, shut down, and routed around. Redundant cores are used to re-form a logically functional whole.

The enormous scale of our wafer-scale engine provides four fundamental technical advantages for AI:

1.Massive, tightly-coupled compute. All modern, high-quality GenAI models require too much compute to run on a single GPU. Our wafer-scale processor is purpose-built for these AI workloads and tightly couples a massive amount of compute resources (900,000 cores, 52 times more than the leading GPU currently in the market) onto the same piece of silicon. By integrating the equivalent compute of a cluster of GPUs onto a single chip, the WSE can train or run inference on large AI workloads without breaking up the problem across a large number of small devices. Keeping so much compute in one place also allows the WSE to

120

achieve significantly higher performance and power efficiency, compared to GPU clusters that require heavy communication across hundreds or thousands of small chips. Put simply, the WSE avoids the bandwidth, latency, and power tax of distributing communication-heavy AI computation across chips.

Comparison of Cerebras Wafer-Scale Integration Versus Traditional GPU Cluster Integration

2.Fast, efficient interconnect. For both inference and training, AI requires extraordinary amounts of data movement. On GPU clusters, most of the data movement occurs between individual GPU chips using traditional I/O interconnects such as NVLink or Infiniband. On the Cerebras wafer-scale chip, most of the data movement is performed locally using the on-chip fabric that achieves nearly 30,000 times higher interconnect bandwidth and is over 100 times more power-efficient per bit of data moved, compared to leading GPU interconnects. This architecture achieves industry-leading interconnect performance and efficiency compared to current GPU servers. This directly translates into faster acceleration for both training and inference, as it minimizes communication across slow, physical networking interfaces.

3.Outsized on-chip memory and memory bandwidth. AI workloads use memory to store the model state (parameters) and active data samples running through the model (activations). Fast access to this data is critical to the performance of both training and inference. The WSE uses a unique memory-in-compute architecture, which integrates memory and processing cores on the same chip, minimizing the distance that data must travel between memory and compute. Memory-in-compute architectures provide a step-change improvement in bandwidth compared to traditional external-memory architectures. This directly results in dramatically reduced latency and power consumption. Traditionally, memory-in-compute architectures were restricted to only small-scale problems because the tight integration of memory and compute was constrained by the small size of single chips. With wafer-scale integration, we are able to scale to enormous on-chip memory (SRAM) capacity (44 GB) on a single chip, while achieving 7,000 times more memory bandwidth (21 petabytes per second) compared to the leading GPU.

121

WSE-3 Achieves 7,000 Times More Memory Bandwidth Compared to NVIDIA H100

4.Native hardware acceleration for sparsity (zeros in data). AI models are designed to be over-parameterized and created with more parameters than required to produce the result. The goal of training a neural network is to find the important subset of the parameters to produce the best result. Therefore, AI models can contain a large number of zero-valued parameters, which are less important, and do not impact the result. These zero-valued parameters are considered “sparse” and can mathematically be skipped when computing the model, since multiply-by-zero is always zero. The WSE-3 architecture has the memory bandwidth and built-in fine-grained dataflow hardware mechanisms to skip multiplies by zero, only performing work for the non-zero important parameters, and thereby accelerating performance. We believe the WSE-3 is the only commercially available hardware that can accelerate every sparsity pattern because our hardware supports fully unstructured sparsity. This sparsity acceleration uses less power by skipping unnecessary compute and can provide a significant additional performance advantage both for training and inference.

Sparsity Within Neural Networks. Dense networks can be made sparse by removing the less important parameters (connections between nodes) and retaining only the important information. We believe Cerebras has the only hardware architecture on the market capable of accelerating all forms of sparsity, including unstructured sparsity.

122

CS-3 System Powered by the WSE-3 – Innovative Power and Cooling for Our Wafer-Scale Chip

The CS-3 system houses the WSE-3 and is engineered to tackle the unique thermal and power challenges of wafer-scale integration. To support the wafer-scale chip and make it easy to deploy in modern data centers, we have developed and productized key system and packaging technologies:

•Thermal expansion-tolerant packaging. The physical connection between the chip and the surrounding infrastructure must be tolerant to the physical stresses caused by thermal expansion over a wide temperature range. At wafer-scale, the thermal expansion stress is significantly higher than traditional chip packages because of the size of the chip. We have developed a unique packaging technique using flexible connections to compensate for the high degree of thermal stresses, designed to enable reliable and high performance in all workload scenarios.

•Wafer-scale power and cooling. Since the wafer-scale chip is significantly larger in size than traditional chips, it requires innovative methods of power and cooling that cannot be satisfied by typical methods designed for smaller chips. To address this challenge, we have developed a novel, perpendicular power and water-cooling delivery system designed to provide uniform power and cooling to the entire surface of the wafer, maximizing power efficiency and thermal management.

Additionally, the CS-3 provides all the physical infrastructure, interfaces, and management to easily integrate into a standard data center. The CS-3 is a 16RU chassis with high-availability, redundant, and hot-swappable power supplies and fans. The system uses high-speed, standards-based 100G Ethernet connections to communicate to the cluster and rest of the data center. We have also developed an advanced system manager that is constantly monitoring and adjusting the unique wafer-scale operating conditions and providing system administrator interfaces.

Cerebras CS-3 System. The left diagram shows the wafer packaging, called the “engine block” providing direct power and cooling to the wafer. The right diagram shows the entire CS-3 system, which houses a single WSE-3 and engine block, and provides the physical infrastructure to integrate into a standard data center.

The Cerebras AI Supercomputer – Purpose Built for Scaling Training and Inference Performance

While a single CS-3 already has the equivalent performance of a GPU cluster, there is often a need to accelerate even further by using multiple CS-3 systems. Our specialized wafer-scale AI Supercomputer architecture is designed to scale near-linearly on up to 2,048 CS-3 systems, and to bring exceptional acceleration for training and inference

123

on even the largest AI models, without distribution complexity. Through co-design with the WSE-3 chip and CS-3 system, Cerebras has developed and productized key cluster-scaling technologies:

•Training parameter storage. The cluster uses a device called MemoryX to centrally store model parameters during training, enabling seamless model scaling. Parameters are streamed to the CS-3 system for training computations, allowing even a single CS-3 to train the largest models, limited only by MemoryX storage capacity. This separation of storage from compute allows for scaling the model size independent of compute capacity. In the CS-3 wafer-scale cluster, SKUs support models ranging from 30 billion to 24 trillion parameters.

•Data-parallel training interconnect. The cluster features a specialized interconnect, SwarmX, enabling data-parallel scaling for training large models. SwarmX includes active devices for weight broadcast and gradient reduction, simplifying operations compared to traditional GPU clusters. In the CS-3 wafer-scale cluster, SwarmX uses 400G and 800G links, designed to support up to 2,048 CS-3 systems in a single cluster.

•Inference runtime processing. The AI Supercomputer has proprietary software that handles inference runtime operations. Model parameters are stored in on-chip memory for ultra-low latency during inference. The inference runtime subsystem provides low latency data input to each CS-3, coordinates their operations, and interfaces with external inference systems. It also manages numerous simultaneous inference streams while maintaining ultra-low latency in the CS-3 wafer-scale cluster.

The Cerebras Wafer-Scale Cluster. The single cluster architecture has resources to support training (shown on the left) and inference (shown on the right). These resources are assigned to CS-3 systems (shown in the middle) such that any CS-3 system can be used either for training or inference.

CSoft Compiler Co-designed With the Hardware to Deliver Easy-To-Use Performance

The Cerebras Software platform (CSoft) is foundational to the usability and user productivity advantages of the Cerebras solution. CSoft allows ML users to program for Cerebras Supercomputer clusters as simply as they would for a single computer, without any distributed computing complexity. This enables rapid iteration and allows users

124

to focus purely on ML experimentation, model quality, and business value creation. To achieve these goals, we developed and productized key software and AI technologies:

•PyTorch tracing and integration. PyTorch is the industry-standard programming language of ML practitioners because it is easy to use and abstracts away the hardware. PyTorch allows users to focus on the ML algorithm, programming in the high-level Python language while avoiding low-level languages such as CUDA. CSoft is tightly integrated with PyTorch so that programs written in PyTorch can be extracted and then mapped to the Cerebras architecture efficiently. The PyTorch integration uses an advanced tracing technique that can analyze the AI model prior to executing it, to identify all the model components to be globally optimized. This tracing technique is designed to be leverageable for other ML frameworks that gain future popularity as well, allowing us to adapt as the needs of the ML community change.

•Performance optimizing graph compiler. Once the AI model has been extracted from PyTorch, the CSoft Graph Compiler compiles the model for the Cerebras hardware and automatically generates the machine code that runs on each core of each WSE-3. The compile process uses a series of proprietary steps to transform the high-level program to efficient, low-level machine code. To find an optimal mapping, we developed a series of advanced optimization algorithms using constraint solving and simulated annealing to produce high-performance out-of-the-box on the CS-3 hardware.

•AI model and workflow libraries. On top of the PyTorch and compiler foundation, the remaining critical enabler to state-of-the-art AI is the ML algorithmic techniques needed to create and run the model for training or inference. We have developed and validated state-of-the-art ML techniques that have led to many research publications and releases of best-in-class models, including GPT model scaling laws, long context architectures, multi-modal image and text models, and hyper-parameter tuning recipes. The list of techniques continues to grow as we further our applied ML research and learnings from customer engagements. We have distilled these techniques into a library of AI models, layer abstractions, and workflows that enable users to directly achieve state-of-the-art ML results, without having to build the model or learn the techniques from scratch. This library is called the Cerebras Model Zoo, an open-source repository of reference model implementations using the latest ML techniques, data preprocessing tools, and training recipes. Model Zoo includes a wide range of GenAI model examples supporting a diverse set of downstream applications and provides users with an easy foundation to build upon that already leverages state-of-the-art advancements and best practices.

The CSoft Software Stack. User programs are extracted from industry standard PyTorch, optimized and lowered to execute on the CS-3 systems with high performance without further user intervention.

125

Extraordinary Performance, Efficiency, and Ease of Use – Fundamental Technology Innovation Produces Industry-Leading User Value

The above Cerebras technology advancements combine to provide direct customer benefit across key areas, solving fundamental challenges faced by GPU clusters:

1.High performance training: over 10 times faster training time-to-solution than the leading 8-way GPU systems of the same generation (as reported by many of our customers), designed to scale to 256 exaFLOPS.

2.High performance inference: over 10 times faster output generation speeds than GPU-based solutions from top CSPs.

3.High power efficiency: three times higher performance per watt than the leading 8-way GPU system.

4.Easy to program, with low developer switching cost.

Sales and Marketing

Our sales and marketing strategy centers on deep market understanding and customer-centric product development. We leverage our extensive market knowledge, proven track record in delivering large-scale compute solutions, and close customer collaborations to optimize our product roadmap. This is designed to ensure our solutions consistently deliver significant value to our customers.

Alongside a focused sales force, we maintain a dedicated Field ML team. This team provides customers with access to leading AI expertise, ensuring they are positioned to leverage our technology effectively. Field ML teams are supported by Applied ML teams, product applications engineers, marketing, and business development/strategy teams, fostering a comprehensive customer support structure.

We focus our sales and marketing efforts on industry leaders, specifically large enterprises domestically and abroad with rich data assets. Our customers are seeking to leverage their rich proprietary data and combine it with Cerebras’ industry leading compute and AI expertise to build a durable competitive advantage. Among our customers, word of success travels quickly, and as a result, it is very important to our future that we maintain strong and collaborative relationships and that we invest behind the largest and most successful of our customers. We utilize master purchase agreements, purchase orders, and statements of work, to define work scope, price, quantities, delivery terms, warranties, and software subscriptions. We predominantly sell our solutions directly to customers via on-premise hardware or via the cloud, based on a consumption-based model.

Research and Development

We are committed to relentless innovation in both hardware and software to address the rapidly-evolving computational needs of GenAI.

In the hardware domain, our research and development efforts focus on chip and system design. We spearheaded the development of wafer-scale integration, resulting in three generations of industry-leading processors at the 16nm, 7nm, and 5nm process nodes. While the idea that bigger silicon would lead to computational efficiencies was novel in 2016, the rest of the industry has now moved in this direction. Cerebras’ innovations in wafer-scale technology helped solve a technical problem that had been open in the processor industry for 75 years. Core members of our founding and early engineering team are world-leaders in chip design and took big risks to help drive the computational industry forward. Our rigorous design methodology utilizes cutting-edge simulation tools and partnerships with Electronic Design Automation providers to create high-quality processors and accelerated development cycles.

126

In the software domain, we develop specialized compilers that translate AI models for optimal performance on our hardware, maximizing performance and efficiency. Our researchers are leaders in sparsity techniques, a critical approach for efficient AI workloads, as evidenced by our publications in this field. These techniques are represented in new and advanced models that we make available to our customers.

We dedicate significant resources to ongoing research and development. We invest heavily in attracting and retaining a global team of highly skilled engineers across dedicated facilities in the United States, Canada, and India. This unwavering commitment to innovation fuels our growth and positions us as a leader in the GenAI landscape.

Manufacturing and Suppliers

We operate a fabless manufacturing model, strategically partnering with industry leaders for the production of our AI compute systems which include ICs, boards, and systems. Our core manufacturing partners include:

TSMC, a leading semiconductor foundry, fabricates our cutting-edge WSEs. Advanced Semiconductor Engineering (“ASE”) handles specialized processes, including the deposition of redistribution layers, and we manage final wafer packaging, assembly, and testing in our Sunnyvale, California facility.

We also use a small number of third parties to manufacture subassemblies and critical components such as printed circuit boards, I/O subsystems, cooling assemblies and power delivery modules. The manufacturing process is subject to extensive testing and verification.

Our supply chain is designed for flexibility and for quality, as we plan to ramp up production to meet the growing global demand for our AI compute systems. Simultaneously, we are committed to rigorous quality control throughout the manufacturing process to confirm reliability in even the most demanding environments at our customer facilities. Our contract manufacturing partners perform system assembly, extensive testing, and verification protocols are in place at every stage including post assembly. Final system-level burn-in and test is conducted by Cerebras. Our quality processes include high production test coverage, full product traceability, and extensive post assembly burn-in. We employ a dedicated quality team that continuously monitors feedback during manufacturing and after deployment. This data-driven approach allows us to improve our product quality and reliability, and help us meet the stringent demands of our customers worldwide.

Intellectual Property

Protecting our intellectual property and proprietary technology, including our AI products and solutions, is an important aspect of our business. We rely on a combination of intellectual property rights, including patent, trademark, trade secret, and other related laws in the United States and internationally as well as confidentiality procedures and contractual provisions to protect, maintain, and enforce our proprietary technology, intellectual property rights, and brand. Our intellectual property portfolio includes patents, trademarks, proprietary software, and trade secrets.

As of June 30, 2024, we owned 85 issued patents and 15 pending patent applications globally. Of these, 44 are issued U.S. patents and nine are pending U.S. patent applications. Our issued patents and pending patent applications generally relate to the design and fabrication of wafer-scale processors, the assembly, packaging, and cooling of wafer-scale processors and hardware, and software architectures for accelerated deep learning. The expiration dates of the U.S.-issued patents are between 2038 and 2042, not taking into account any applicable patent term extensions. We routinely review our development efforts to assess the existence and patentability of new inventions.

We have a policy of requiring employees and consultants to execute confidentiality agreements upon the commencement of an employment or consulting relationship with us. Our employee and independent contractor agreements also require relevant employees and independent contractors to assign to us all rights to any inventions made or conceived during their employment or engagement with us. In addition, we typically require individuals and

127

entities with which we discuss potential business relationships to sign non-disclosure agreements that contain customary confidentiality provisions.

Competition

We offer a purpose-built AI compute platform. Our CS product family primarily competes against solutions from NVIDIA Corporation, Advanced Micro Devices, Inc., Intel Corporation, Microsoft Corp., and Alphabet Inc., among others, as well as internally developed custom application-specific integrated circuits and a variety of private companies, some of which are focused on inference-only offerings.

We believe that our ability to remain competitive will depend on how well we are able to anticipate the features and functions that customers will require and whether we are able to deliver consistent volumes of our products at acceptable levels of quality and at competitive prices. We expect competition to increase from both existing competitors and new market entrants with products that may be lower priced than ours or may provide better performance or additional features not provided by our products. In addition, it is possible that new competitors or alliances among competitors could emerge and acquire significant market share. Some of our competitors have greater marketing, financial, distribution, and manufacturing resources than we do and may be more able to adapt to customers or technological changes. We expect an increasingly competitive environment in the future.

Human Capital

As of September 26, 2024, we had 401 employees, including 252 in the United States, and we have employees located internationally, including in Canada, and in India. We maintain a full-time workforce and supplement our workforce with contractors and consultants.

To our knowledge, none of our employees are represented by a labor union or party to a collective bargaining agreement. We consider our relationships with our employees to be good. Our human capital resources objectives include, as applicable, identifying, recruiting, retaining, incentivizing, and integrating our existing and new employees. The principal purposes of our equity incentive plans are to attract, retain, and reward personnel through the granting of stock-based compensation awards in order to increase stockholder value and the success of our company by motivating such individuals to perform to the best of their abilities and achieve our objectives.

Facilities

Our corporate headquarters is located in Sunnyvale, California, where we lease approximately 68,000 square feet for office space, research and development, and testing, pursuant to a lease agreement that expires in November 2027, subject to the terms thereof. We lease additional facilities in San Diego, Canada, and India for research and development.

We also enter into agreements for offsite colocation facilities to house and operate our AI supercomputers. We enter into these agreements for our own corporate purposes as well as on behalf of our customers. Currently, our data center facilities are in California, and we manage another data center facility in Texas on behalf of a customer.

We believe that our facilities are suitable to meet our current needs. We intend to expand our facilities or add new facilities as we grow, and we believe that suitable additional or alternative spaces will be available on commercially reasonable terms, if required.

Government Regulations

We are subject to many U.S. federal and state laws, rules, and regulations, as well as laws, rules, and regulations imposed by various non-U.S. governmental authorities, including those related to intellectual property, tax, import and export requirements, anti-corruption, economic and trade sanctions, national security and foreign investment, foreign exchange controls and cash repatriation restrictions, data privacy and security requirements, competition, advertising, employment, product regulations, environment, health, and safety requirements, and consumer laws.

128

These laws and regulations are complex, are constantly evolving, and may be interpreted, applied, created, or amended, in a manner that could harm our business.

The import and export of our products and technology are subject to laws and regulations, including international treaties, U.S. and various non-U.S. export controls and sanctions laws, customs regulations, and other trade rules. The scope, nature, and severity of such controls varies widely across different countries and may change frequently over time. Such laws, rules, and regulations may delay the introduction of some of our products or impact our competitiveness through restricting our ability to do business in certain countries or territories or with certain parties (including certain governments) or certain jurisdictions. U.S. export restrictions also require us to obtain licenses from the U.S. Department of Commerce to allow the export or transfer of our products (including our software and technology), and there can be no assurance that export permissions will be granted.

See the section titled “Risk Factors” for additional information regarding risks we face related to government regulation.

Legal Proceedings

From time to time, we may be subject to legal proceedings, claims, and investigations in the ordinary course of business. We are not presently a party to any litigation to which the outcome, we believe, if determined adversely to us, would individually or taken together have a material adverse effect on us. We cannot predict the results of any such proceedings, claims, or investigations, and despite the potential outcomes, the existence thereof may have a material adverse impact on us due to diversion of management time and attention as well as the financial costs related to resolving such matters.

129

MANAGEMENT

Executive Officers and Directors

The following table sets forth information regarding our executive officers and directors as of September 30, 2024:

Name

Age

Position(s)

Executive Officers and Employee Director:

Andrew D. Feldman

Chief Executive Officer, President, and Director

Robert Komin

Chief Financial Officer

Dhiraj Mallick

Chief Operating Officer

Non-Employee Directors:

Paul Auvil(1)(3)

Director

Glenda Dorchak(1)(2)

Director

Thomas Lantzsch(2)

Director

Lior Susan(3)

Director

Steve Vassallo(2)(3)

Director

Eric Vishria(1)

Director

_______________

(1)Member of the audit committee.

(2)Member of the compensation committee.

(3)Member of the nominating and corporate governance committee.

Executive Officers and Employee Director

Andrew D. Feldman is one of our co-founders and has served as our Chief Executive Officer and President and as a member of our board of directors since April 2016. From February 2012 to June 2014, Mr. Feldman served as Corporate Vice President and General Manager at Advanced Micro Devices, Inc. (“AMD”), a semiconductor company. From November 2007 to February 2012, Mr. Feldman served as Chief Executive Officer at SeaMicro, a dense microserver company acquired by AMD. From August 2003 to December 2006, Mr. Feldman served as Vice President, Marketing and Product Management at Force10 Networks, Inc., a computer networking company acquired by Dell, Inc. From March 2000 to August 2003, Mr. Feldman served as Vice President, Corporate Marketing and Corporate Development at Riverstone Networks Inc., a networking switching hardware company. Mr. Feldman holds an M.B.A. from Stanford University and a B.A. in Economics and Political Science from Stanford University.

We believe Mr. Feldman is qualified to serve as a member of our board of directors because of the perspective and experience he brings as our co-founder and Chief Executive Officer. See “—Involvement in Certain Legal Proceedings” for certain details regarding historical legal proceedings involving Mr. Feldman.

Robert Komin has served as our Chief Financial Officer and Treasurer since March 2024. Mr. Komin previously served as Chief Financial Officer of Sunrun Inc. (“Sunrun”), a residential solar and storage company, from March 2015 to May 2020, and then continued as a consultant until January 2021. From September 2013 to January 2015, Mr. Komin served as Chief Financial Officer at Flurry, Inc., a mobile analytics and advertising company. From August 2012 to August 2013, Mr. Komin served as Chief Financial Officer at Ticketfly, Inc., a music ticketing and marketing services provider. From January 2010 to July 2012, Mr. Komin served as Chief Operating Officer and Chief Financial Officer at Linden Research, Inc., a creator of virtual digital entertainment and cybercurrency. Mr. Komin previously served as a member of the board of directors and audit committee of Bird

130

Global Inc., a micromobility company, from June 2021 to April 2024. Mr. Komin holds an M.B.A. from Harvard Business School and a B.S. in Accounting and General Science from the University of Oregon.

Dhiraj Mallick has served as our Chief Operating Officer since September 2023. From June 2018 to September 2023, Mr. Mallick served as our Senior Vice President of Engineering and Operations. From November 2015 to May 2018, Mr. Mallick served as Vice President of Innovation, Pathfinding and Architecture, Data Center Group at Intel Corporation, a multinational technology company. From April 2012 to August 2015, Mr. Mallick served as General Manager and Corporate Vice President at AMD. Since January 2020, Mr. Mallick has served on the Global Advisory Group of the Global Semiconductor Alliance, a semiconductor and technology industry organization. Mr. Mallick holds an M.S. in Electrical Engineering from Stanford University and a B.S. in Electrical Engineering from the University of Rochester.

Non-Employee Directors

Paul Auvil has served as a member of our board of directors since July 2024. Mr. Auvil previously served as Chief Financial Officer of Proofpoint, Inc., an enterprise security company, from March 2007 to February 2023. From September 2006 to March 2007, Mr. Auvil was an entrepreneur-in-residence with Benchmark, a venture capital firm. From August 2002 to July 2006, Mr. Auvil served as Chief Financial Officer at VMware, Inc., a cloud-computing and virtualization company. From April 1998 to January 2002, Mr. Auvil served as Chief Financial Officer at Vitria Technology, Inc., an eBusiness platform company. Mr. Auvil held various executive positions at VLSI Technology, Inc., a semiconductor and circuit manufacturing company, from August 1988 to March 1998, including serving as the Vice President of the Internet and Secure Products Division. Mr. Auvil has served as a member of the board of directors of Elastic N.V., a platform for search-powered solutions, since October 2023. Mr. Auvil previously served as a member of the boards of directors of 1Life Healthcare, Inc. (doing business as One Medical), a primary care organization acquired by Amazon, Inc., from September 2019 to February 2023, Quantum Corporation, a data storage company, from August 2007 to November 2017, Marin Software Incorporated, a cloud-based advertisement management platform company, from October 2009 to April 2017, and OpenTV Corp., a provider of interactive television software and services, from January 2010 to April 2010. Mr. Auvil holds an M.M. from the Kellogg Graduate School of Management at Northwestern University and a B.E. in Electrical Engineering from Dartmouth College.

We believe Mr. Auvil is qualified to serve as a member of our board of directors because of his extensive experience in the technology industry and as an executive and a member of the boards of directors of technology companies.

Glenda Dorchak has served as a member of our board of directors since July 2024. Ms. Dorchak previously served as Executive Vice President and General Manager of Global Business for Spansion Inc., a flash memory manufacturer, from April 2012 to June 2013. Ms. Dorchak served as Chief Executive Officer of VirtualLogix, Inc., a virtualization software solutions company, from January 2009 to October 2010, and as Chief Executive Officer of Intrinsyc Software, a software company, from July 2006 to November 2008. From March 2001 to July 2006, Ms. Dorchak served in various roles at Intel Corporation, including as Vice President and Chief Operating Officer Intel Communications Group, Vice President and General Manager Intel Broadband Products Group, and Vice President and General Manager Intel Consumer Electronics Group. Ms. Dorchak served as Chairman and Chief Executive Officer at Value America, an e-retailer, from September 1998 to November 2000. From July 1974 to September 1998, Ms. Dorchak held various management and executive positions at IBM Corporation, a semiconductor and circuit manufacturing company, including General Manager and Director of IBM Direct, and Director, General Business, of IBM Personal Systems Group North America. Ms. Dorchak has served as a member of the boards of directors of Wolfspeed, Inc., a provider of silicon carbide materials and semiconductor products, since January 2020, Globalfoundries Inc., a semiconductor contract manufacturing and design company, since June 2019, and Ansys Inc., an engineering simulation software company, since July 2018. She previously served as a member of the boards of directors of Viavi Solutions Inc., a provider of network test, monitoring, and assurance solutions, from November 2019 to October 2021, Mellanox Technologies, Ltd., a multinational supplier of computer networking products, from June 2009 to April 2020, Quantenna Communications, a communication device

131

company, from June 2018 to June 2019, and Energy Focus Inc., a developer of energy-efficient LED lighting systems and controls, from July 2015 to February 2019.

We believe Ms. Dorchak is qualified to serve as a member of our board of directors because of her experience in our industry and as a member of the boards of directors of public companies.

Thomas Lantzsch has served as a member of our board of directors since September 2024. Mr. Lantzsch previously served as Senior Vice President and General Manager, Internet of Things, of Intel Corporation from January 2017 to January 2023. From December 2006 to November 2016, Mr. Lantzsch served as Executive Vice President of Strategy and Corporate Development of Arm Inc., a semiconductor company. Mr. Lantzsch has served as a member of the board of directors of Canatu Oyj, a carbon nanomaterial developer company, since November 2023. Mr. Lantzsch holds an M.S. in Finance from the Naveen Jindal School of Management at the University of Texas at Dallas and a B.S. in Electrical Engineering from Michigan State University.

We believe Mr. Lantzsch is qualified to serve as a member of our board of directors because of his expertise and experience working with and for technology and semiconductor companies.

Lior Susan has served as a member of our board of directors since April 2023. Since January 2015, Mr. Susan has served as Founder and Managing Partner of Eclipse Ventures, a venture capital firm. Mr. Susan is a co-founder of Bright Machines, Inc., a software company, and has served as its Executive Chairman since January 2018 and as its Chief Executive Officer from January 2018 to May 2018, from December 2021 to December 2022, and from August 2023 to the present. From June 2012 to January 2015, Mr. Susan served as Founder and General Partner at LabIX, the hardware investment platform of Flextronics International Ltd., an end-to-end supply chain solutions company. Mr. Susan served as an Advisor at Intucell Ltd., a self-optimizing network software company, from April 2008 until it was sold to Cisco in 2012. Mr. Susan has served as a member of the board of directors of Owlet, Inc., a health technology company, since March 2015, and has served as the chairman of its board of directors since July 2021. He also serves as a member of the boards of directors of several private companies, including Augury, Inc., Bright Machines, Inc., Chord Commerce, Inc., Cybertoka Ltd., Datapelago, Inc, Datorios, Inc., Dutch Pet, Inc., Flex Logix Technologies, Inc., InsidePacket, Ltd., Senser, Ltd., and Skyryse, Inc. Mr. Susan is former member of an elite Special Forces unit in the Israel Defense Force.

We believe Mr. Susan is qualified to serve as a member of our board of directors because of his expertise and experience working with and investing in technology companies.

Steve Vassallo has served as a member of our board of directors since May 2016. Since October 2007, Mr. Vassallo has served as a general partner and in various other roles at Foundation Capital, a venture capital firm. From September 2004 to September 2006, Mr. Vassallo served as Vice President of Product and Engineering at Ning Interactive Inc., a social platform. From May 1999 to September 2002, Mr. Vassallo served as director of engineering at Immersion Corporation, a haptic technology company. Mr. Vassallo previously served as a member of the board of directors of Sunrun from May 2008 to June 2019. Mr. Vassallo also serves as a member of the boards of directors of several private companies. Mr. Vassallo holds an M.B.A. from Stanford University, an M.S. in Electromechanical Engineering from Stanford University, and a B.S. in Mechanical Engineering from Worcester Polytechnic Institute.

We believe Mr. Vassallo is qualified to serve on our board of directors because of his extensive experience in the technology industry.

Eric Vishria has served as a member of our board of directors since May 2016. Since July 2014, Mr. Vishria has served as a General Partner of Benchmark, a venture capital firm. From August 2013 to August 2014, Mr. Vishria served as Vice President, Digital Magazines and Verticals at Yahoo! Inc., a web services provider. From November 2008 to August 2013, Mr. Vishria served as co-founder and Chief Executive Officer of RockMelt, Inc., a social media web browser. He has served as a member of the boards of directors of Amplitude, Inc., a digital optimization company, since December 2014, and Confluent, Inc., a data solutions company, since September 2014.

132

He also serves as a member of the boards of directors of several private companies. Mr. Vishria holds a B.S. in Mathematical and Computational Science from Stanford University.

We believe Mr. Vishria is qualified to serve on our board of directors because of his extensive experience as a venture capital investor and a member of the boards of directors of other technology companies.

Involvement in Certain Legal Proceedings

Andrew D. Feldman was previously one of six named defendants in the action SEC v. Pereira, No. 3:06-cv-06384-CRB (N.D. Cal.), in which the SEC alleged, among other things, that Mr. Feldman, as Vice President, Corporate Marketing and Corporate Development of Riverstone Networks, Inc., negotiated, reviewed, approved, or was otherwise aware of sales transactions in 2001 and 2002 that were improperly accounted for by Riverstone and aided and abetted Riverstone in violating U. S. securities laws. Without admitting or denying the allegations of the complaint, Mr. Feldman settled the claims against him in 2008 by entering into an agreement with the SEC permanently restraining and enjoining Mr. Feldman from violating federal securities laws and requiring Mr. Feldman to pay $289,507 plus interest. In connection with the same alleged facts, Mr. Feldman also pled guilty in December 2007 to one count of circumventing accounting controls of an issuer in violation of 15 U.S.C. sections 78m(b)(5) and 78ff, and was sentenced to three years of probation and fined $5,000 in connection with an action brought by the U.S. Department of Justice captioned USA v. Feldman, No. 3:07-cr-07-00731-0001-CRB (N.D. Cal.).

Family Relationships

There are no family relationships among any of our executive officers or directors.

Board Structure and Composition

Director Independence

Our board of directors currently consists of seven members. Our board of directors has determined that all of our directors, other than Mr. Feldman, qualify as independent directors in accordance with the Nasdaq Listing Rules. Mr. Feldman is not considered independent by virtue of his position as an executive officer of the company. Under the Nasdaq Listing Rules, the definition of independence includes a series of objective tests, such as that the director is not, and has not been for at least three years, one of our employees and that neither the director nor any of his or her family members has engaged in various types of business dealings with us. In addition, as required by the Nasdaq Listing Rules, our board of directors has made a subjective determination as to each independent director that no relationships exists that, in the opinion of our board of directors, would interfere with the exercise of independent judgment in carrying out the responsibilities of a director. In making these determinations, our board of directors reviewed and discussed information provided by the directors and us with regard to each director’s relationships as they may relate to us and our management.

Classified Board of Directors

In accordance with our amended and restated certificate of incorporation, which will be effective immediately prior to the completion of this offering, our board of directors will be divided into three classes with staggered three-year terms. At each annual general meeting of stockholders, the successors to directors whose terms then expire will be elected to serve from the time of election and qualification until the third annual meeting following their election. Our directors will be divided among the three classes as follows:

•The Class I directors will be Andrew D. Feldman, Paul Auvil, and Eric Vishria, and their terms will expire at the annual meeting of stockholders to be held in 2025;

•The Class II directors will be Glenda Dorchak and Steve Vassallo, and their terms will expire at the annual meeting of stockholders to be held in 2026; and

133

•The Class III directors will be Thomas Lantzsch and Lior Susan, and their terms will expire at the annual meeting of stockholders to be held in 2027.

We expect that any additional directorships resulting from an increase in the number of directors will be distributed among the three classes so that, as nearly as possible, each class will consist of one-third of the directors. The division of our board of directors into three classes with staggered three-year terms may delay or prevent a change of our management or a change in control.

Leadership Structure of the Board of Directors

Our amended and restated bylaws and corporate governance guidelines to be adopted immediately following the effectiveness of the registration statement of which this prospectus forms a part will provide our board of directors with flexibility to combine or separate the positions of chairperson of the board of directors and Chief Executive Officer and to implement a lead director in accordance with its determination regarding which structure would be in the best interests of our company.

Our board of directors currently believes that our existing leadership structure, under which our chief executive officer, Mr. Feldman, serves as chairman of our board of directors, is effective.

Our board of directors will continue to periodically review our leadership structure and may make such changes in the future as it deems appropriate.

Our board of directors has elected Eric Vishria to serve as lead independent director. As lead independent director, Mr. Vishria will preside at all meetings of the board of directors at which the chairman of the board of directors is not present, including executive sessions, and perform such additional responsibilities as set forth in our corporate governance guidelines.

Voting Arrangements

The election of the members of our board of directors is currently governed by our amended and restated voting agreement that we entered into with certain holders of our capital stock and the related provisions of our current amended and restated certificate of incorporation. Pursuant to our amended and restated voting agreement and current amended and restated certificate of incorporation, Mr. Feldman was elected by certain holders of our common stock, voting together as a single class, and Messrs. Susan, Vassallo, and Vishria were elected by the holders of our Series A redeemable convertible preferred stock.

Our amended and restated voting agreement will terminate and the provisions of our current amended and restated certificate of incorporation by which our directors were elected will be amended and restated in connection with this offering. After this offering, the number of directors will be fixed by our board of directors, subject to the terms of our amended and restated certificate of incorporation and amended and restated bylaws that will become effective immediately prior to the completion of this offering. Each of our current directors will continue to serve as a director until the election and qualification of his or her successor, or until his or her earlier death, resignation, or removal.

Role of Board in Risk Oversight Process

Risk assessment and oversight are an integral part of our governance and management processes. Our board of directors encourages management to promote a culture that incorporates risk management into our corporate strategy and day-to-day business operations. Management discusses strategic and operational risks at regular management meetings, and conducts specific strategic planning and review sessions during the year that include a focused discussion and analysis of the risks facing us. Throughout the year, senior management reviews these risks with the board of directors at regular board meetings as part of management presentations that focus on particular business functions, operations or strategies, and presents the steps taken by management to mitigate or eliminate such risks.

134

Our board of directors does not have a standing risk management committee, but rather administers this oversight function directly through our board of directors as a whole, as well as through various standing committees of our board of directors that address risks inherent in their respective areas of oversight. While our board of directors is responsible for monitoring and assessing strategic risk exposure, our audit committee is responsible for overseeing our major financial and cybersecurity risk exposures and the steps our management has taken to monitor and control these exposures. The audit committee also approves or disapproves any related person transactions. Our nominating and corporate governance committee monitors the effectiveness of our corporate governance guidelines. Our compensation committee assesses and monitors whether any of our compensation policies and programs has the potential to encourage excessive risk-taking. The risk oversight process also includes receiving regular reports from our committees and members of senior management to enable our board of directors to understand our risk identification, risk management, and risk mitigation strategies with respect to areas of potential material risk, including operations, finance, legal, regulatory, cybersecurity, strategic and reputational risk.

Board Committees

Effective as of the date the registration statement of which this prospectus forms a part is declared effective by the SEC, our board of directors will have three standing committees: an audit committee; a compensation committee; and a nominating and corporate governance committee. Each committee is governed by a charter that will be available on our website following completion of this offering. Members serve on these committees until their resignation or until otherwise determined by our board of directors.

Audit Committee

Effective as of the date the registration statement of which this prospectus forms a part is declared effective by the SEC, the members of our audit committee will consist of Paul Auvil, Glenda Dorchak, and Eric Vishria. Mr. Auvil will be the chairperson of our audit committee. The composition of our audit committee meets the requirements for independence under the current Nasdaq Listing Rules and Rule 10A-3 of the Exchange Act. Each member of our audit committee is financially literate. In addition, our board of directors has determined that Mr. Auvil is an “audit committee financial expert” within the meaning of the SEC rules. This designation does not impose on such directors any duties, obligations, or liabilities that are greater than are generally imposed on members of our audit committee and our board of directors. Our audit committee is directly responsible for, among other things:

•appointing, retaining, compensating, and overseeing the work of our independent registered public accounting firm;

•assessing the independence and performance of the independent registered public accounting firm;

•reviewing with our independent registered public accounting firm the scope and results of the firm’s annual audit of our financial statements;

•overseeing the financial reporting process and discussing with management and our independent registered public accounting firm the financial statements that we will file with the SEC;

•pre-approving all audit and permissible non-audit services to be performed by our independent registered public accounting firm;

•reviewing policies and practices related to risk assessment and management;

•reviewing our accounting and financial reporting policies and practices and accounting controls, as well as compliance with legal and regulatory requirements;

•reviewing cybersecurity matters;

135

•reviewing, overseeing, approving, or disapproving any related-person transactions;

•reviewing with our management the scope and results of management’s evaluation of our disclosure controls and procedures and management’s assessment of our internal control over financial reporting, including the related certifications to be included in the periodic reports we will file with the SEC; and

•establishing procedures for the confidential anonymous submission of concerns regarding questionable accounting, internal controls, or auditing matters, or other ethics or compliance issues.

Compensation Committee

Effective as of the date the registration statement of which this prospectus forms a part is declared effective by the SEC, the members of our compensation committee will consist of Glenda Dorchak, Thomas Lantzsch, and Steve Vassallo. Ms. Dorchak will be the chairperson of our compensation committee. Each of Ms. Dorchak and Messrs. Lantzsch and Vassallo is a non-employee director, as defined by Rule 16b-3 promulgated under the Exchange Act and meets the requirements for independence under the current Nasdaq Listing Rules. Our compensation committee is responsible for, among other things:

•reviewing and approving the compensation of our executive officers, including reviewing and approving corporate goals and objectives with respect to compensation;

•authority to act as an administrator of our equity incentive plans;

•reviewing and approving, or making recommendations to our board of directors with respect to, incentive compensation and equity plans;

•reviewing and recommending that our board of directors approve the compensation for our non-employee directors; and

•establishing and reviewing general policies relating to compensation and benefits of our employees.

Nominating and Corporate Governance Committee

Effective as of the date the registration statement of which this prospectus forms a part is declared effective by the SEC, the members of our nominating and corporate governance committee will consist of Steve Vassallo, Paul Auvil, and Lior Susan. Mr. Vassallo will be the chairperson of our nominating and corporate governance committee. Messrs. Vassallo, Auvil, and Susan meet the requirements for independence under the current Nasdaq Listing Rules. Our nominating and corporate governance committee is responsible for, among other things:

•identifying and recommending candidates for membership on our board of directors, including the consideration of nominees submitted by stockholders, and on each of the board’s committees;

•reviewing and recommending our corporate governance guidelines and policies;

•reviewing proposed waivers of the code of business conduct and ethics for directors and executive officers;

•overseeing the process of evaluating the performance of our board of directors; and

•assisting our board of directors on corporate governance matters.

Code of Business Conduct and Ethics

In connection with this offering, our board of directors will adopt a code of business conduct and ethics that applies to all of our employees, officers, and directors, including our Chief Executive Officer, Chief Financial

136

Officer, and other executive and senior financial officers. Upon completion of this offering, the full text of our code of business conduct and ethics will be posted on the investor relations section of our website. We intend to disclose future amendments to our code of business conduct and ethics, or any waivers of such code, on our website or in public filings.

Indemnification and Insurance

We maintain directors’ and officers’ liability insurance. Our amended and restated certificate of incorporation and amended and restated bylaws will include provisions limiting the liability of directors and officers and indemnifying them under certain circumstances. We have entered or will enter into indemnification agreements with each of our directors and officers to provide our directors and officers with additional indemnification and related rights. See the section titled “Description of Capital Stock—Limitations on Liability and Indemnification Matters” for additional information.

Compensation Committee Interlocks and Insider Participation

None of the members of our board of directors who will serve on our compensation committee upon the effectiveness of the registration statement of which this prospectus forms a part is or has been an officer or employee of our company. None of our executive officers currently serves, or in the past fiscal year has served, as a member of a compensation committee (or if no committee performs that function, the board of directors) of any other entity that has an executive officer serving as a member of our board of directors.

137

EXECUTIVE AND DIRECTOR COMPENSATION

Executive Compensation

The following is a discussion and analysis of compensation arrangements of our named executive officers (“NEOs”). This discussion contains forward looking statements that are based on our current plans, considerations, expectations and determinations regarding future compensation programs. Actual compensation programs that we adopt may differ materially from currently planned programs as summarized in this discussion. As an “emerging growth company” as defined in the JOBS Act, we are not required to include a Compensation Discussion and Analysis section and have elected to comply with the scaled disclosure requirements applicable to emerging growth companies.

We seek to ensure that the total compensation paid to our executive officers is reasonable and competitive. Compensation of our executive officers is structured around the achievement of individual performance and near-term corporate targets as well as long-term business objectives.

Our NEOs for the year ended December 31, 2023 were:

•Andrew D. Feldman, our Chief Executive Officer and President;

•Dhiraj Mallick, our Chief Operating Officer; and

•Anthony E. Maslowski, our former Chief Financial Officer.

Mr. Maslowski ceased serving as our Chief Financial Officer on March 15, 2024.

Summary Compensation Table

The following table sets forth total compensation paid to our NEOs for 2023.


Name and Principal Position	Year	Salary ($)	Bonus ($)	Stock Awards ($)(1)	Option Awards ($)(1)	Non-Equity Incentive Plan Compensation ($)(2)	All Other Compensation ($)	Total ($)
Andrew D. Feldman Chief Executive Officer	2023	400,000	—	117,659	444,000	342,500	—	1,304,159
Dhiraj Mallick Chief Operating Officer	2023	385,417	—	3,514,000	900,350	187,500	—	4,987,267
Anthony E. Maslowski Former Chief Financial Officer	2023	350,000	—	—	592,000	100,000	—	1,042,000

_______________

(1)Amounts shown represent the grant date fair value of stock options and RSUs granted during 2023 as calculated in accordance with ASC Topic 718. See “Stock-Based Compensation” in Note 12 to our audited consolidated financial statements included elsewhere in this prospectus for the assumptions used in calculating this amount.

(2)Amounts shown represent performance-based cash bonuses earned based on the achievement of a certain corporate financial objective during 2023.

138

Narrative to Summary Compensation Table

Annual Base Salaries

We pay each of our NEOs a base salary to compensate them for services rendered to our company. The base salary payable to each NEO is intended to provide a fixed component of compensation reflecting the executive’s skill set, experience, role, and responsibilities.

For 2023, the compensation committee of our board of directors established the annual base salary of each NEO for 2023 as follows: Mr. Feldman, $400,000; Mr. Mallick, $375,000; and Mr. Maslowski, $350,000. In August 2023, our compensation committee increased Mr. Mallick’s annual base salary to $400,000, and in February 2024, our compensation committee increased the annual base salary of Mr. Feldman to $425,000.

Our board of directors and compensation committee may adjust base salaries from time to time in their discretion.

Annual Bonuses

We maintain an annual performance-based cash bonus program in which each of our NEOs participated in 2023. Under our annual performance-based cash bonus program, our compensation committee has established a target bonus opportunity for each NEO, with the NEO earnings based on the achievement of a certain corporate financial objective established by our compensation committee. The target bonus opportunity established for each NEO for 2023 was as follows: Mr. Feldman, $250,000; Mr. Mallick, $175,000; and Mr. Maslowski, $100,000. In August 2023, Mr. Mallick’s annual target bonus opportunity was increased to $200,000.

In January 2024, our compensation committee determined the corporate financial objective was achieved under our annual performance-based cash bonus program and approved the following bonuses for our NEOs based on 2023 performance: Mr. Feldman, $342,500; Mr. Mallick, $187,500; and Mr. Maslowski, $100,000. The annual bonus approved for Mr. Feldman included an adjustment above achievement of the corporate financial objective of $92,500 based on our compensation committee’s overall assessment of Mr. Feldman’s performance during 2023.

For 2024, our compensation committee adopted a similar annual performance-based cash bonus program.

Our board of directors and compensation committee may adjust annual target bonus opportunities or award discretionary bonuses from time to time.

Equity-Based Compensation

We have granted stock options and RSUs to our NEOs to attract and retain them, as well as to align their interests with the interests of our stockholders. Our stock options are generally exercisable prior to vesting (with any unvested shares subject to repurchase at the original exercise price upon any termination of service) and generally vest over one, three or four years, subject to continued service to the company. Our RSUs generally require satisfaction of both a service-based vesting condition, which is generally satisfied over a four-year period, and a liquidity-based vesting condition, which will be satisfied upon completion of this offering.

In February 2023, we granted Mr. Feldman an option to purchase 150,000 shares of our Class A common stock and 23,438 RSUs. The RSUs were granted in satisfaction of Mr. Feldman’s 2022 performance-based bonus and will fully vest upon completion of this offering or, if earlier, the date of a change in control. The option was immediately exercisable and vests in substantially equal monthly installments over four years from January 1, 2023, subject to Mr. Feldman’s continued service. Vesting of the option is accelerated in full in the event Mr. Feldman’s employment is terminated by us without cause or Mr. Feldman for good reason during the period commencing three months prior to a change in control and ending 12 months following the closing of the change in control. The option has a term of 10 years, subject to earlier termination upon a termination of employment, and an exercise price per share of $5.02.

139

In February 2023, we granted each of Mr. Mallick and Mr. Maslowski an option to purchase 200,000 shares of our Class A common stock. Each option was immediately exercisable and vests in substantially equal monthly installments over four years from January 1, 2023, subject to the executive’s continued service. Vesting of each option is accelerated in full in the event the executive’s employment is terminated by us without cause or the executive for good reason during the period commencing three months prior to a change in control and ending 12 months following the closing of the change in control. Each option has a term of 10 years, subject to earlier termination upon a termination of employment, and an exercise price per share of $5.02.

In August 2023, we granted Mr. Mallick three additional options, each providing Mr. Mallick the right to purchase 35,000 shares of our Class A common stock for an exercise price of $5.02 per share. Each option was immediately exercisable and vests in 12 substantially equal monthly installments from August 1, 2023, 2024, and 2025, respectively, such that each option will be fully vested on the first anniversary of the applicable vesting start date. For each option, Mr. Mallick may choose between keeping the option or giving up the option in return for a cash bonus of $300,000. If the bonus is chosen, the applicable option will terminate. Otherwise, each option has a term of 10 years, subject to earlier termination upon a termination of employment.

In connection with this offering, we intend to adopt a 2024 Incentive Award Plan, referred to below as the 2024 Plan, in order to facilitate the grant of cash and equity incentives to employees (including our NEOs), directors, and consultants of our company and certain of our affiliates and to enable us to obtain and retain services of these individuals, which is essential to our long-term success. We expect that the 2024 Plan will become effective on the business day immediately prior to the date of effectiveness of the registration statement of which this prospectus forms a part, subject to approval of such plan by our stockholders. For additional information about the 2024 Plan, see the section titled “Equity Compensation Plans” below.

Other Elements of Compensation

Retirement Savings and Health and Welfare Benefits

We currently maintain a 401(k) retirement savings plan for our employees, including our NEOs, who satisfy certain eligibility requirements. Our NEOs are eligible to participate in the 401(k) plan on the same terms as other full-time employees. The U.S. Internal Revenue Code of 1986, as amended (the “Code”) allows eligible employees to defer a portion of their compensation, within prescribed limits, on a pre-tax basis through contributions to the 401(k) plan. We believe that providing a vehicle for tax-deferred retirement savings though our 401(k) plan adds to the overall desirability of our executive compensation package and further incentivizes our employees, including our NEOs, in accordance with our compensation policies.

All of our full-time employees, including our NEOs, are eligible to participate in our health and welfare plans, including medical, dental, and vision benefits; medical and dependent care flexible spending accounts; short-term and long-term disability insurance; and life and accidental death and dismemberment insurance.

Perquisites and Other Personal Benefits

We provide perquisites and other personal benefits to our NEOs when we believe it is necessary to attract or retain the NEO. None of our NEOs received any perquisites during 2023.

140

Outstanding Equity Awards at Year End

The following table lists all outstanding equity awards held by our NEOs as of December 31, 2023.


			Option Awards(1)(2)					Stock Awards(3)
Name	Vesting Commencement Date		Number of Securities Underlying Unexercised Options (#) Exercisable	Number of Securities Underlying Unexercised Options (#) Unexercisable	Option Exercise Price ($)		Option Expiration Date	Number of Shares or Units of Stock That Have Not Vested (#)	Market Value of Shares or Units of Stock That Have Not Vested ($)(3)
Andrew D. Feldman	2/15/2019		1,150,000		2.40		5/13/2029
	3/1/2022	(4)	262,500	337,500	2.72		12/7/2030
	1/1/2023	(4)	45,833	104,167	7.89		1/11/2032
	1/1/2023	(4)	34,375	115,625	5.02		2/13/2033
	2/14/2023	(5)						23,438	128,440
Dhiraj Mallick	6/28/2018		469,410		$	0.98	7/16/2028
	6/28/2021	(6)	166,666	33,334	$	2.72	7/6/2030
	6/15/2021	(4)	62,500	37,500	$	2.89	3/14/2031
	10/1/2021	(7)						325,000	1,781,000
	8/23/2022	(4)	100,000	200,000	$	6.47	8/22/2032
	1/1/2023	(4)	45,833	154,167	$	5.02	2/13/2033
	8/1/2023	(8)	11,666	23,334	$	5.02	7/31/2033
	8/1/2024	(8)		35,000	$	5.02	7/31/2033
	8/1/2025	(8)		35,000	$	5.02	7/31/2033
	8/1/2023	(9)						700,000	3,836,000
Anthony E. Maslowski	1/28/2020	(10)						14,867	81,471
	1/1/2023	(6)	30,555	69,445	$	7.89	1/11/2032
	1/1/2023	(4)	45,833	154,167	$	5.02	2/13/2033

_______________

(1)Each option is exercisable as to all shares underlying the option with any shares purchased upon exercise subject to the same vesting conditions applicable to the option. In the event of any termination of employment, unvested shares may be repurchased by us for the exercise price of the related option. The portion of the option included under the “Number of Securities Underlying Unexercised Options Unexercisable” represents the unvested portion of the option notwithstanding that it is fully exercisable.

(2)Except as otherwise provided, each option vests at a rate of 25% of the shares underlying the option on the first anniversary of the vesting commencement date and as to 1/48th of the shares underlying the option on each monthly anniversary of the vesting commencement date thereafter, in each case, subject to the executive continuing to provide services to us through the applicable vesting date.

(3)Amount reported calculated by multiplying $5.48, which our board of directors determined equaled fair market value of our Class A common stock as of December 31, 2023, by the number of unvested shares comprising or underlying the stock award.

(4)The option vests as to 1/48th of the shares underlying the option on each monthly anniversary of the vesting commencement date, subject to the executive continuing to provide services to us through the applicable vesting date.

(5)The RSUs vest upon the completion of this offering, subject to the executive continuing to provide services to us through such completion.

(6)The option vests as to 1/36th of the shares underlying the option on each monthly anniversary of the vesting commencement date, subject to the executive continuing to provide services to us through the applicable vesting date.

(7)Each RSU vests on the date both service-based and liquidity-based vesting conditions are satisfied. The service-based vesting condition is satisfied as to 1/48th of the total number of RSUs on each monthly anniversary of the

141

vesting commencement date, subject to the executive continuing to provide services through the applicable date. The liquidity-based vesting condition will be satisfied upon the completion of this offering.

(8)The option vests as to 1/12th of the shares underlying the option on each monthly anniversary of the vesting commencement date, subject to the executive continuing to provide services to us through the applicable vesting date.

(9)Each RSU vests on the date both service-based and liquidity-based vesting conditions are satisfied. The service-based vesting condition is satisfied as to 1/36th of the total number of RSUs on each monthly anniversary of the vesting commencement date, subject to the executive continuing to provide services through the applicable date. The liquidity-based vesting condition will be satisfied upon the completion of this offering.

(10)Constitutes restricted stock acquired upon exercise of an option prior to vesting, which are subject to repurchase at the original exercise price upon any termination of service. The restricted stock vests in full on January 28, 2024, subject to the executive continuing to provide services through such date.

Executive Compensation Arrangements

We are party to offer letters with each of our NEOs, other than Mr. Feldman. Each offer letter provides for an initial base salary, target bonus opportunity, two initial stock option grants with accelerated vesting of the two option grants in the event of certain terminations of employment in connection with a change in control, and participation in our benefits plans.

Mr. Maslowski began serving as our Chief Financial Officer in January 2020 and ceased serving as our Chief Financial Officer on March 15, 2024. In February 2024, we entered into a separation agreement with Mr. Maslowski after a commitment to an extended tenure was not reached. Under the separation agreement, in exchange for a general release of claims against us and our affiliates, we paid Mr. Maslowski $225,000, which constituted six months of his base salary at the rate in effect as of his date of termination and 50% of his target bonus opportunity. The separation agreement also provided Mr. Maslowski with access to COBRA coverage for 36 months from his separation date, and (i) for the first 24 months of such period, the premiums will be paid for by us and (ii) for the remaining 12 months of such period, Mr. Maslowski will be responsible for payment of the full premiums. The separation agreement also provided Mr. Maslowski with the full vesting acceleration of each of his outstanding stock options and shares of restricted stock, up to 12 months to exercise the stock option granted to him in February 2023, and up to 18 months to exercise the stock option granted to him in January 2022.

Equity Compensation Plans

The following summarizes the material terms of the long-term incentive compensation plan and employee stock purchase plan in which our NEOs will be eligible to participate following this offering and our existing equity plans, under which we have previously made periodic grants of equity and equity-based awards to our NEOs and other employees.

Cerebras Systems Inc. S-1