Canada produced some of the most influential minds in modern AI. For years, however, the compute infrastructure to match that intellectual legacy simply did not exist domestically. Researchers and companies built on foreign hardware, in foreign facilities, under foreign jurisdiction. Buzz HPC, a subsidiary of Hive Digital Technologies, is making a serious attempt to change that picture.
The company announced plans for a 320MW data center in the Greater Toronto Area, targeting a go-live date in the second half of 2027. At full build-out, the facility will house more than 100,000 GPUs across two parcels of land totaling 25 acres. Hive purchased the land for approximately 58 million Canadian dollars, and the site already carries a 320MW power allocation, which removes one of the more common bottlenecks in large data center development.
The location is notable beyond geography. Positioned between the University of Toronto and the Vector Institute to the east and Waterloo to the west, the site sits at the center of Canada’s most established AI research corridor. That proximity to academic talent and research institutions tends to matter for organizations building long-term AI infrastructure rather than simply chasing short-term compute demand.
Total investment in the project is expected to reach 3.5 billion Canadian dollars. During construction, the facility will generate at least 800 jobs. The cooling system takes a closed-loop, no-water-use approach, which addresses a growing concern for large-scale data center operators facing water availability constraints in certain regions.
Aydin Kilic, president and CEO of Hive Digital Technologies, noted that this facility forms part of a broader Canadian expansion. Hive currently operates a 70MW site in Grand Falls, New Brunswick, and runs 5,500 GPUs for AI compute today. Together with the new GTA site, the company expects to support roughly 130,000 GPUs domestically. Furthermore, Hive is working with Bell Canada across several Canadian deployments, including a recent 6.5MW deployment in Merritt, British Columbia.
For Canada’s AI ecosystem, the significance of purpose-built sovereign compute infrastructure extends beyond any single company. Consequently, when research talent and domestic infrastructure finally occupy the same geography, the competitive dynamics shift in ways that take years to fully appreciate.
