KIDZ AI Inc. Enters 60-Month GPU Services Agreement with Canopy Wave

What Happened

• KIDZ AI Inc. (through wholly owned subsidiary Catalyst Compute LLC) announced on July 17, 2026 that it entered into a service order and terms of service with Canopy Wave, Inc. Catalyst will deploy a dedicated cluster of 32 GPU nodes (256 NVIDIA HGX B300 GPUs) and provide GPU processing, CPU server and storage capacity. The services will run for an initial 60‑month term beginning when hardware setup is complete; total service fees over that term are $44,626,944. A press release was furnished on July 21, 2026.

Key Details

• Effective date: July 17, 2026 (agreement becomes effective after both parties sign and Catalyst places a non‑cancellable purchase order).
• Hardware: 32 nodes with a total of 256 NVIDIA HGX B300 GPUs plus specified CPUs, memory, storage and networking.
• Fees & term: $44,626,944 in aggregate service fees over an initial 60‑month term; services commence upon hardware setup.
• Uptime & remedies: Catalyst must provide ≥99.5% monthly uptime (subject to exclusions); service credits apply for certain GPU hardware failures.
• Termination: Canopy Wave may not terminate in the first 24 months. After that, a convenience termination requires payment equal to 75% of remaining service fees (subject to mitigation/avoided costs); there are customary termination rights for breach, insolvency, regulatory change and prolonged force majeure. Canopy Wave can terminate without the early fee if Catalyst’s GPU hardware faults cause failure to meet the 99.5% uptime in any 10 months of a rolling 12‑month period (with some cure periods).

Why It Matters

• The contract provides multi‑year, predictable revenue (roughly $44.6M) for Catalyst Compute, supporting KIDZ AI’s infrastructure services business.
• It involves significant upfront capital/hardware commitments (non‑cancellable purchase order and dedicated equipment) and operational SLAs (99.5% uptime) that create both revenue opportunity and performance/operational risk.
• Key investor considerations: timing of service commencement (depends on hardware setup), capital expenditures to procure/deploy the cluster, and exposure to potential service credits or early termination fees if performance or demand changes. A press release was filed as Exhibit 99.1 to the 8‑K.