What happened
According to CryptoBriefing and The Energy Mag, HIVE Digital Technologies announced plans for a CAD $3.5 billion, 320-megawatt AI "gigafactory" in the Greater Toronto Area, and the company's stock jumped roughly 35% on the news. The Energy Mag reports the project will be developed by HIVE's subsidiary BUZZ High Performance Computing Inc., is targeted to host more than 100,000 GPUs at full build-out, and is expected to be operational in the second half of 2027. The Energy Mag reports BUZZ acquired roughly 25 acres, paying about CAD $46 million for a 21-acre main parcel and CAD $12 million for an adjacent four-acre parcel, and that the site carries a 320 MW utility power allocation. CryptoBriefing and The Energy Mag report HIVE currently operates around 5,500 GPUs, controls more than 850 MW of power capacity globally, and has secured about CAD $30 million in AI cloud services contracts.
Technical details
According to The Energy Mag, BUZZ said the Ontario campus will be designed for vertically integrated AI supercomputing with closed-loop cooling and a no-water-use approach. Editorial analysis - technical context: At this scale, operators typically prioritize high-efficiency cooling, substations with direct utility feeds, and mechanical redundancy; these choices materially affect capital intensity, site lead times, and PUE considerations for practitioners planning similar deployments.
Context and significance
The Energy Mag frames the announcement within the policy theme of "sovereign AI infrastructure," noting proximity to the Toronto-Waterloo corridor and local research institutions. Editorial analysis: For the AI infrastructure market, a single 320 MW allocation and a >100k-GPU build-out represents a sizeable incremental demand signal for GPUs, power delivery, and local grid upgrades, and it highlights how companies with existing power contracts or land-banked substations can enter the data center market faster than greenfield entrants.
What to watch
For practitioners: financing and capital structure for multi-billion-dollar builds, including whether HIVE materially expands external project financing or joint-venture partners. For practitioners: power-delivery timelines and local utility approvals, since utility interconnection and substation upgrades commonly drive schedule risk for 100+ MW deployments. For practitioners: GPU procurement and supply-chain timing, because scaling toward 100,000 GPUs depends on multi-quarter vendor allocations and logistics.
Bottom line
According to the reporting, HIVE's announcement is a large-scale infrastructure bid that combines land acquisition, substantial power allocation, and an explicit GPU target; industry observers and practitioners should treat it as a concrete example of crypto-mining assets and expertise being repurposed into AI compute capacity.
Key Points
- 1HIVE's CAD $3.5B, 320 MW plan signals large-scale repurposing of miner power assets toward AI compute, increasing utility and GPU demand.
- 2Site-level technical choices, like closed-loop no-water cooling and direct substation access, materially affect capital and schedule risk for gigawatt-scale builds.
- 3Observers should track financing, interconnection timelines, and GPU procurement as the critical constraints for turning announced capacity into usable AI compute.
Scoring Rationale
A large, well-specified AI data center announcement with **320 MW** and **100,000+ GPUs** is a notable infrastructure development for practitioners, affecting power planning and GPU demand. It is company-level rather than industry-defining, so it rates as a major but not paradigm-shifting story.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
