Nvidia and BUZZHPC deepen AI infrastructure ties as parent HIVE targets $100M revenue

BUZZHPC, the high-performance computing arm of HIVE Digital Technologies, has completed an integration of Nvidia’s silicon and software on its platform, adding another milestone to what has become one of Canada’s more quietly ambitious AI infrastructure plays.

The move cements BUZZHPC’s standing as an Nvidia Cloud Partner and positions the company closer to its parent’s stated goal of surpassing $100 million in annual recurring revenue.

What BUZZHPC is actually building

BUZZHPC has been running GPU clusters across Canada and the Nordics since 2017, long before “sovereign AI” became a buzzword in policy circles.

The platform runs on Nvidia’s current top-tier hardware lineup: Hopper H100 and H200 GPUs alongside the newer Blackwell B200 and GB200 NVL72 systems. These are connected through Quantum InfiniBand networking, which is the high-speed fabric that keeps large GPU clusters from bottlenecking each other during intensive AI training runs.

The most consequential deal in BUZZHPC’s recent history came in June 2026: a three-year GPU cloud contract with Bell Canada worth approximately $220 million. The agreement covers the deployment of 2,304 Nvidia Grace Blackwell GPUs.

A separate partnership with Bell Canada, announced in 2025, covers a 5 to 6.5 megawatt project in Manitoba, with planned expansions into Quebec.

The sovereign AI pitch

Sovereign AI refers to AI infrastructure that keeps data physically within a country’s borders, subject to local law rather than foreign jurisdiction. For Canadian enterprises and government agencies that handle sensitive data, the question of where compute actually lives is increasingly a procurement requirement, not a preference.

HIVE Digital Technologies raised $115 million through a convertible note in 2026, with proceeds directed toward expanding GPU cloud capacity. The ARR target of over $100 million gives investors a concrete benchmark to hold management against.

HIVE itself started as a Bitcoin miner, which gave it early expertise in running large clusters of power-hungry processors at scale. That operational muscle transferred reasonably well when the company pivoted toward GPU cloud services as AI demand overtook crypto mining economics.

Reading the risk alongside the momentum

Announced contracts and deployed infrastructure are not the same thing. The Bell Canada agreement covers a deployment timeline that stretches across three years. Revenue recognition for cloud infrastructure deals of this type tends to be spread over the contract term rather than booked upfront, which means the $220 million figure represents potential revenue, not cash already in hand.

The research notes there is no official record or announcement confirming a completed silicon and software integration on the BUZZHPC platform as of the latest information available, raising questions about the specifics of their operational advancements.

What BUZZHPC does have is a physical footprint in Canada with renewable energy access and a major telco as an anchor customer. A multi-year, nine-figure commitment from Bell Canada provides a revenue floor that most GPU cloud startups would trade a lot to have.