The contract runs three years and is delivered through HIVE's subsidiary, BUZZ High Performance Computing. It involves 2,304 NVIDIA Grace Blackwell GPUs—chips designed for frontier AI model training and inference—installed at Bell's purpose-built data center in Merritt, British Columbia.
Cohere, a large language model company that builds AI systems for enterprises and governments, will use that compute layer to run its platform for Canadian clients.
Cohere is a fitting anchor for the arrangement. The Toronto company is one of the few anywhere building foundation models—the base-layer AI that powers enterprise chatbots, government document processing, and everything in between—and recently announced a merger with Germany's Aleph Alpha that values the combined company at roughly $20 billion. Bell and Cohere had an existing partnership dating to July 2025; this contract is the compute infrastructure underneath it.
“Canada helped pioneer modern artificial intelligence. What we have lacked is not talent, it is industrial infrastructure to commercialize that talent at scale before others do it for us,” Frank Holmes, executive chairman at HIVE Digital Technologies, said in a statement. “This partnership with Bell and Cohere is a defining moment. BUZZ HPC is the GPU factory layer that transforms Canada's AI ambitions from political promises into productive national assets”
Crypto mining returns are volatile and get harder as more miners compete for block rewards. Things get even harder during crypto winters, downturns in the market, as rewards become increasingly less attractive as the price of crypto assets go down but costs stay the same or go up.
On the opposite side, AI compute demand is growing fast and clients—especially government agencies—sign multi-year contracts at locked-in rates. Trading one bubble for another, maybe, but at least this one has a government mandate behind it.
Once the deployment goes live—expected between late 2026 and early 2027—HIVE expects roughly $70 million in new annual recurring revenue on top of the $35 million it already books from existing GPU operations. Its contracted HPC revenue target now exceeds $100 million.
The company also has a larger project in the works: a 320-megawatt AI data center in the Greater Toronto Area designed to house more than 100,000 Nvidia GPUs at full build-out.
