Alpha Compute has closed a $32.2 million revenue contract with a leading frontier artificial intelligence laboratory for high-performance GPU acceleration. The two-year agreement delivers $16.1 million in Annual Recurring Revenue (ARR) to Alpha Compute, with an expected upfront payment of $7.5 million.
Overview
The deal centers on Alpha Compute's custom-designed, PCIe-based GPU accelerator cards, which are optimized for dense matrix multiplication and other key AI computations. The agreement covers a dedicated cluster of 504 NVIDIA B200 GPUs, providing the unnamed frontier AI laboratory with exclusive, high-performance compute access to accelerate its next-generation model development.
What it does
The NVIDIA B200 GPU delivers unprecedented performance for AI and HPC workloads, making it the preferred architecture for frontier model training at scale. Alpha Compute's GPU assets deliver privacy-preserving computation to partners and applications, including Telegram, Animoca Brands, GAMEE, and Midnight Network.
Tradeoffs
The contract highlights the escalating importance of hardware acceleration in AI research and development. Alpha Compute is actively expanding its sales pipeline of enterprise compute agreements with AI laboratories, sovereign entities, and enterprise customers seeking dedicated, high-performance GPU infrastructure outside of traditional hyperscaler environments.
The key details of the transaction are:
- Total Contract Value: $32,200,000
- Annual Recurring Revenue: $16,100,000
- Upfront Payment: $7,500,000
- GPU Deployment: 504 × NVIDIA B200 Tensor Core GPUs
- Data Center Location: Canada
- Energy: 100% Hydroelectric Power
- Agreement Term: Two (2) Years
- Tenant: Unnamed Leading Frontier AI Laboratory
In practical terms, this deal underscores the growing demand for specialized hardware in large-scale AI workloads. As AI research and development continue to advance, the need for high-performance GPU acceleration will likely increase, making companies like Alpha Compute key players in the field.