TL;DR:
- HIVE Digital Technologies closed a $220 million, three-year sovereign AI cloud contract through its subsidiary BUZZ HPC.
- The deal involves 2,304 NVIDIA Grace Blackwell GPUs and integrates Bell Canada, Cohere and Hypertec into a Canadian AI infrastructure.
- HIVE projects the contract will contribute approximately $70 million in annual recurring revenue once fully deployed.
HIVE Digital Technologies breaks into the artificial intelligence infrastructure market following the close of a sovereign GPU cloud contract valued at approximately $220 million. The agreement, signed for three years by its subsidiary BUZZ HPC alongside Bell Canada, Cohere and Hypertec, allows the company — until recently focused exclusively on Bitcoin mining — to diversify its revenue streams, as mining as a business continues to adjust to post-halving margins.
According to the company, the project will deploy 2,304 NVIDIA Grace Blackwell GPUs under the Bell AI Fabric framework, and is estimated to generate around $70 million in annual revenue once fully implemented.
The Miners’ Path Toward AI
The significance of HIVE’s agreement goes beyond its face value. Over the past two years, public miners have sought to demonstrate that their energy contracts, data centers and technical equipment can serve markets beyond Bitcoin block production. High-performance computing and artificial intelligence emerged as the clearest alternative, given that the underlying infrastructure — energy, cooling, facilities and operations — can be redirected toward enterprise clients with high compute demand.
This does not imply that mining and GPU infrastructure are interchangeable. Hardware requirements, service standards and client profiles differ substantially. Nevertheless, the direction is clear: operators with access to cheap energy and scalable sites are looking to convert those assets into more predictable revenue streams.
HIVE: Sovereign AI
The scale of the contract carries its own weight. Bell AI Fabric, Cohere, Hypertec and BUZZ HPC present the agreement as part of a domestic sovereign AI initiative, where computing capacity, connectivity and model development remain within a national infrastructure. That approach is gaining relevance as governments and companies prioritize AI systems that meet local data, privacy and security requirements.
For HIVE, the immediate challenge is execution. GPU procurement, availability commitments, customer concentration and capital intensity represent real risks.
