お知らせ • Apr 29
Alpha Compute Confirms That Its First Large-Scale GPU Cluster Is Now in Final Testing and Is Targeted to Hand over to AI Compute Customers on May 8, 2026
Alpha Compute Corp. provided an update on the deployment status of its NVIDIA Blackwell GPU infrastructure. The Company confirms that its first large-scale GPU cluster, a 504-chip NVIDIA B200 server deployment located in Canada, ALPHA-01, is now in final testing and is targeted to hand over to AI compute customers on May 8, 2026. Upon the launch of the ALPHA-01 cluster, Alpha Compute's balance sheet is projected to reflect total assets of $48.1 million, total liabilities of $36.8 million, and total equity of $11.3 million. These enterprise-grade deployments follow Alpha Compute’s test bed deployments of NVIDIA B300s, B200s and H200s, which have been purchased, installed and upgraded to AI Confidential Computing from November 2025 through the present. Alpha Compute has been able to prove out its confidential computing thesis and capabilities, and is now in growth mode. ALPHA-01 B200 Cluster Go-Live Target of May 8, 2026; Alpha Compute's inaugural large-scale deployment consists of 504 NVIDIA B200 GPUs hosted in a Canadian data center. The cluster was originally targeted for a March 31, 2026 launch; however, delivery of critical hardware components from the manufacturer was subject to supply -chain delays that pushed the timeline by four weeks. The machines are now fully installed and in their final testing phase. The data center management team is on track to deploy the cluster into production for customer use on May 8, 2026. ALPHA-01 deployment is designed for scalability. Alpha Compute holds a contractual Right-of-First-Refusal (ROFR) on an expansion to more than 1,000 NVIDIA B200 chips at the same facility, with a targeted expansion go-live in August 2026 (ALPHA-03), providing a clear path to doubling Canadian compute capacity within the same operating year. ALPHA-01 Cluster: Deployment Summary; Location: Canada, Hardware: 504 NVIDIA B200 GPUs, Energy: 100% Hydro, Status: Final testing phase; data center management targeting May 8, 2026 handover, Original Timeline: March 2026 (delayed due to hardware delivery delays), Expansion ROFR: 1,000+ NVIDIA B200 GPUs; targeted August 2026 go-live. ALPHA-02 Sweden B300 Cluster: June 2026 Target; Alpha Compute has secured a binding term sheet for bridge financing to fund its second major deployment, a 576-chip NVIDIA B300 cluster at Equinix's AtNorth datacenter in Sweden, one of Northern Europe's premier high-performance computing facilities. The cluster is targeted to go live in June 2026. ALPHA-02 Cluster: Deployment Summary; Location: Equinix AtNorth Datacenter, Sweden, Hardware: 576 NVIDIA B300 GPUs, Energy: 100% Hydro, Status: Bridge financing binding term sheet signed, Target Go-Live: June 2026, Expansion ROFR: 1,000+ NVIDIA B300 GPUs; targeted September 2026 go-live, ALPHA-03 and ALPHA-04 Deployments. These deployments follow Alpha Compute's earlier testbed installations of NVIDIA B300, B200, and H200 GPUs that were purchased, installed, and upgraded to AI Confidential Computing between November 2025 and the present. Having validated its confidential computing thesis, the Company is now in growth mode. All four clusters operate NVIDIA Blackwell-generation GPUs within Trusted Execution Environments (TEEs) using Intel TDX — the foundation of Alpha Compute's confidential compute mandate. Once fully operational, the combined footprint of currently planned deployments is projected to generate $72 million in annual revenue ($6 million per month). Alpha Compute Deployment Roadmap; ALPHA-01 Cluster - May 8, 2026: Canada B200s (504 chips), ALPHA-02 Cluster - June 2026: Sweden B300s at AtNorth/Equinix (576 chips), ALPHA-03 Cluster - August 2026: Canada B200s (Expansion to 1,000+ chips, ROFR), ALPHA-04 Cluster - September 2026: Sweden B300s (Expansion to 1,000+ chips, ROFR).