Canada’s AI Data Centre Strategy Leaves Key Questions Unanswered
January 9, 2026
By Lucas Bettle
The use of generative AI is growing rapidly, and countries around the world are rushing to build the application-specific infrastructure needed. Canada is no exception, with its “Sovereign AI Compute Strategy” providing $2 billion in federal funding for a wide range of projects. While seemingly moving full steam ahead, Canada has yet to put specific regulations in place for AI and data centres, and many economic, environmental, and social problems remain unaddressed.
AI Data Centres and Canada
AI data centres differ from conventional data centres due to the high concentration of GPUs and petascale storage required to train and operate large models (Canadian Energy Regulator, 2024). They operate at much higher power and cooling requirements than data centres handling other IT workloads (Merwat, 2024).
In Canada, infrastructure is being driven largely by the Canadian Sovereign AI Compute Strategy, which focuses on implementing domestic, high-performance capacity under Canadian jurisdiction to ensure economic competitiveness, national security, and data protection (Innovation, Science and Economic Development Canada, 2025). The government positions this approach as beneficial for the public, supporting academic and public-sector research through a shared national platform (Digital Research Alliance of Canada, 2025).
Electricity use is among the most contentious issues surrounding data centres. Globally, data centres already consume an estimated 1.4 to 1.7% of electricity, about 415 TWh in 2024. As AI-intensive workloads grow, this is expected to more than double or even quadruple by 2030 (International Energy Agency, 2025). For perspective, Canada’s total electricity generation in 2023 was 620 TWh. Canadian regulators now single out AI data centres as a major driver of future load growth (Canadian Energy Regulator, 2024).
Canada has several notable advantages in AI infrastructure. Predominantly low-emission electrical generation and a cool climate make the country an ideal site for major AI data centre projects (Canadian Energy Regulator, 2024). Hydropower alone accounts for roughly 60% of Canada’s electricity generation, providing AI data centres with lower emissions profiles than the U.S. and other regions (Eddy, 2025).
Planned Large-Scale AI Projects in Canada
There has been a significant commitment to large-scale projects across Canada, with planned data centres under the Canadian Sovereign AI Compute Strategy and commercial developments (Digital Research Alliance of Canada, 2025). Collectively, Canadian deployments that are operational, under construction, committed, and in early-stage development add up to 10 GW of capacity, with over three-quarters of that still in the pipeline (Eddy, 2025).
The TELUS Sovereign AI Factory in Rimouski, Quebec, opened in September 2025 and is branded as Canada’s first fully sovereign AI compute facility, with TELUS claiming that the facility runs on 99% renewable energy from regional hydropower. TELUS also claims that the new design reduces water consumption by 75% (TELUS, 2025). The energy consumption of the facility has not been publicly disclosed.
Another major AI data centre came online in Cambridge, Ontario, in late 2025. The project by Canadian AI company Cohere had a total cost of $725 million, with $240 million of that coming from federal financing under the Canadian Sovereign AI Compute Strategy (Hemmadi, 2025).
In Alberta, eStruxture’s CAL-3 project near Calgary is a planned 90 MW facility scheduled to go live in the second half of 2026, claiming to incorporate advanced liquid and air cooling and waste-heat recovery (eStruxture, 2024).
The proposed Wonder Valley AI Data Centre Industrial Park in Greenview, Alberta, would be the largest such development in the world. The $70 billion project would create a 7.5 GW campus to host multiple hyperscale operators, powered by natural gas from the Greenview Industrial Gateway (Municipal District of Greenview, 2024). However, the project remains largely speculative as of 2025, lacking confirmed tenants and facing unresolved concerns around emissions, water withdrawals, and Indigenous rights (Ridley, 2025).
A partnership between HIVE Digital Technologies and Bell Canada is focusing on deploying AI capacity at existing Bell facilities throughout the country, with initial rollout planned for late 2025, with a 5 MW deployment in Manitoba (Peng, 2025).
The University of Toronto has received a $42.5 million federal investment to build out state-of-the-art AI infrastructure to support Canadian researchers across a wide range of fields (Kalvapalle, 2025).
The Canadian Regulatory Landscape Surrounding AI Data Centres
The landscape surrounding AI data centres in Canada has yet to be firmly established. The Artificial Intelligence and Data Act was introduced in 2022 in federal Bill C-27 with the aim of regulating high-impact AI and imposing duties on developers and deployers (Innovation, Science and Economic Development Canada, 2025). The bill advanced through committee but died on the order paper when Parliament prorogued in early 2025, leaving Canada without a dedicated AI statute and creating a governance gap (Fraser, 2025).
AI data centres currently fall under a patchwork of general laws. The Personal Information Protection and Electronic Documents Act governs how private-sector organizations collect, use, and share personal information (Innovation, Science and Economic Development Canada, 2025). Concerns around energy use, site approval, grid connections, and emissions currently fall under existing rules for environmental assessments, the Clean Electricity Regulations, utility regulators, and municipal zoning (Canadian Energy Regulator, 2024).
Pressing Concerns Surrounding AI Data Centre Projects
AI data centres are among the fastest-growing electricity loads in Canada and could soon become one of the largest consumers overall. Data centres could represent low double-digit percentages of incremental load by the 2030s (Morton, 2025). Globally, power generation for data centres already emits 180 million tonnes of CO2e annually, with projections of 300 to 500 million tonnes of CO2e by 2035 (Rutgers, 2025). For perspective, Canada’s annual GHG emissions for 2023 were 694 million tonnes of CO2e.
Electricity use from AI data centres is likely to multiply several-fold by 2030. This additional demand could compete with the electrification of transport, buildings, and heavy industry, potentially impacting progress toward climate targets (Merwat, 2024).
Freshwater impacts are also a contentious topic. AI data centres require substantial water withdrawals for cooling. Canadian projects could increase stress on already-strained basins and freshwater systems (Smith, 2025). In 2023, U.S. data centres used about 64 billion litres of water for cooling, with water used in power generation to run the centres reaching 800 billion litres (Shehabi, 2024).
As both the Canadian government and private industry invest billions in AI data centres, uncertainties surrounding long-term AI demand pose serious concerns (Merwat, 2024). If demand for AI does not meet expectations, expensive facilities could be left without any practical application. Furthermore, many projects see value flows to global AI firms, sending profits out of the country while risks ultimately fall on taxpayers (Harland, 2025).
Balancing Canadian AI Growth with Real-World Constraints
Canada is investing big in AI with its Sovereign AI Compute Strategy. However, AI data centre projects pose a variety of risks with regard to climate targets, electricity use, freshwater consumption, and long-term returns. Despite these pressing concerns, Canada remains without a regulatory framework for AI. With many projects already planned and underway, the country can’t afford to go without a clear policy that addresses both environmental and economic factors.
References
Canadian Energy Regulator. (2024). Market Snapshot: Energy demand from data centers is steadily increasing, and AI development is a significant factor.
Digital Research Alliance of Canada. (2025). The Digital Research Alliance of Canada Launches Cross-Country Consultation Series on Canada’s Sovereign AI Compute Infrastructure Strategy.
Eddy, N. (2025). Canada Emerges as Global Data Center Powerhouse. Retrieved from Data Center Knowledge.
eStruxture. (2024). eStruxture Announces Alberta’s Largest Data Center: Introducing the Groundbreaking CAL-3 Facility.
Fraser, D. (2025). The Demise of the Artificial Intelligence and Data Act (AIDA): 5 Key Lessons. Retrieved from McInnes Cooper.
Harland, K. (2025). Is there a smart way to integrate artificial intelligence data centres into Canada’s electricity grids? Retrieved from Canadian Climate Institute.
Hemmadi, M. (2025). Ottawa is putting $240M into Cohere’s $725M Canadian AI compute plan. Retrieved from The Logic.
Innovation, Science and Economic Development Canada. (2025). Artificial Intelligence and Data Act.
Innovation, Science and Economic Development Canada. (2025). Canadian Sovereign AI Compute Strategy.
International Energy Agency. (2025). AI is set to drive surging electricity demand from data centres while offering the potential to transform how the energy sector works.
Kalvapalle, R. (2025). AI compute infrastructure at U of T receives $42.5 million federal investment. Retrieved from U of T News.
Merwat, S. (2024). Power Struggle: How AI is challenging Canada’s electricity grid. RBC Climate Action Institute.
Morton, D. (2025). Top reliability challenges to Canada’s energy system. Retrieved from Energy Regulation Quarterly.
Municipal District of Greenview. (2024). World’s Largest AI Data Centre Industrial Park ‘Wonder Valley’ coming to the Greenview Industrial Gateway.
Peng, A. (2025). HIVE Digital’s BUZZ HPC partners with Bell Canada to deploy 5 MW for AI. Retrieved from Yahoo Finance.
Ridley, S. (2025). The AI data centre boom is here. What will it mean for land, water and power in Canada? Retrieved from The Narwhal.
Rutgers, J.-S. (2025). Pretty much everything you need to know about Manitoba’s new obsession with AI data centres. Retrieved from The Narwhal.
Shehabi, A. (2024). 2024 United States Data Center Energy Usage Report. Lawrence Berkeley National Laboratory.
Smith, A. (2025). Data’s dark, thirsty side. Retrieved from The Canadian Bar Association.
TELUS. (2025). TELUS opens Canada’s first fully Sovereign AI Factory.
