Canada’s data center boom is accelerating fast, but the biggest wave of new AI-linked infrastructure is heading toward a province with a far more carbon-intensive grid, raising major questions about energy, ownership, and the future of digital infrastructure.
Key Takeaways
Canada’s data center sector is undergoing a structural shift, with planned capacity nearly 10 times the current operating base and new projects significantly larger than existing facilities.
Future growth is heavily concentrated in Alberta, which accounts for about 93% of planned capacity, even though its power grid is far more carbon-intensive than the national average.
Provinces with cleaner power, including Quebec, Ontario, and British Columbia, have introduced grid-access constraints for large new loads, while Alberta is pursuing a more active investment-attraction strategy.
Ownership of Canadian data center infrastructure is predominantly foreign, with provider data available, and Canadian pension funds are gaining AI infrastructure exposure mainly through cloud and semiconductor equities, not through direct data center ownership.
Paper: Data Centred: The Shifting Landscape of Canada's Digital Infrastructure. Image Credit: Visuals6x / Shutterstock
In a recent article posted as a working paper on the Social Science Research Network (SSRN) preprint server, researchers Alexander Carlo and Lyndsey Rolheiser at York University, Toronto, Canada, investigated the rapidly evolving landscape of data centers in Canada, driven primarily by increased demand for computational power associated with artificial intelligence (AI) advances since 2022.
Rising Demand for Data Centers
Advances in large-scale machine learning systems have sharply increased demand for data center capacity, particularly in hyperscale facilities designed for AI workloads that require sustained high power and reliable infrastructure.
Canada currently maintains a modest operational base of data centers, but upcoming projects suggest a near tenfold expansion in capacity, underscoring a major shift in scale and spatial distribution. Notably, Alberta emerges as the dominant province for planned capacity, despite having a grid emissions intensity nearly five times the national average.
Other provinces, like Quebec, Ontario, and British Columbia, which offer cleaner and generally less costly electricity, have introduced legislation and allocation frameworks that constrain access to grid power for large new loads, influencing where data centers are sited. This dynamic generates a tension between sustainability goals and economic development ambitions.
Ownership patterns further complicate this landscape, with foreign firms, predominantly U.S.-headquartered, accounting for most identified providers of operational facilities, while Canadian pension funds have increased exposure to AI infrastructure primarily via public equities rather than direct ownership of physical data centers.
Tracking Canada’s Data Center Growth
The study uses detailed data from Aterio, a proprietary source that tracks all stages of data center development in Canada and the United States. This database provides asset-level information, including geographic coordinates, power capacity, development stage, and ownership.
Projects are monitored from public announcement through construction and activation, ensuring a comprehensive view of the sector’s growth. The authors also complement this core dataset with analyses of provincial energy systems, grid access policies, fiscal incentive frameworks, and water-related risks to connect physical infrastructure trends with institutional context.
Additionally, public filings from major Canadian pension funds are examined to gauge indirect financial exposure to AI infrastructure through equity investments. Various relevant government reports, media sources, and industry publications provide contextual background aligned with the empirical evidence.
Expansion, Location, and Ownership Trends
The findings reveal that Canada's current data center footprint comprises relatively small facilities, averaging around 11 MW in operational capacity. However, new projects are substantially larger, averaging 116 MW, reflecting a decisive movement toward hyperscale data centers optimized for AI workloads.
Geographically, operational centers cluster around urban hubs in Ontario, Quebec, and British Columbia, whereas upcoming developments concentrate overwhelmingly in Alberta, with many increasingly located in rural areas far from metropolitan centers. These Alberta projects can consume electricity equivalent to tens of thousands of households, posing challenges for local energy infrastructure and raising questions about emissions trajectories given the province’s fossil fuel-reliant grid.
Provincial policies significantly shape these patterns. Quebec, Ontario, and British Columbia have constrained grid access to manage new load growth and support sustainability objectives, complicating further data center expansion.
Alberta, conversely, leverages a deregulated energy market and has launched a targeted Artificial Intelligence Data Centers Strategy, combining investment-attraction measures, an evolving power-allocation framework, and a dedicated tax levy on computing equipment to attract large-scale investment. However, this strategy entails trade-offs: Alberta’s electricity grid emissions intensity may create reputational risks for operators committed to carbon disclosure and sustainability, and the reliance on fossil fuels introduces the possibility of stranded assets under evolving climate policies.
Ownership analysis confirms that U.S.-based providers dominate Canada’s data center landscape, where provider information is available, with limited direct Canadian ownership of physical assets. Construction activity is predominantly carried out by domestic firms, but strategic control remains in foreign hands.
Canadian institutional investors, including major pension plans, have increased their exposure to AI infrastructure primarily through publicly traded cloud computing and semiconductor companies rather than by acquiring stakes in data center operators. This disconnect between capital flows and ownership of physical infrastructure suggests potential vulnerabilities to Canada’s data sovereignty and raises questions about the future role of Canadian institutional capital in shaping the domestic AI infrastructure ecosystem.
The increasing public visibility of data centers since late 2024 has sparked growing attention to their environmental and social impacts. Media coverage reflects concerns about energy consumption, noise, land use, and associated externalities, positioning data centers as politically contentious infrastructure similar to pipelines or highways.
Implications on Growth and Sustainability
Canada’s data center sector is undergoing a fundamental transformation, driven by rapidly rising demand for computing capacity, much of it tied to AI expansion. A near tenfold increase in planned infrastructure, primarily concentrated in Alberta, signifies a major reallocation of electricity demand, land use, and capital investment. This shift poses critical policy challenges across provincial energy systems, regulatory frameworks, and environmental sustainability.
Alberta’s dominant position illustrates a trade-off between attracting large-scale AI infrastructure investment and managing the environmental implications of a carbon-intensive electricity grid. In conclusion, as AI continues to shape data center growth and political salience, further research is needed to understand the local socioeconomic impacts, energy system integration, and long-term sustainability of Canada’s digital infrastructure expansion.
Source:
- Carlo A. and Rolheiser L. (2026). Data Centred: The Shifting Landscape of Canada's Digital Infrastructure. Social Science Research Network. DOI: 10.2139/ssrn.6464099, https://ssrn.com/abstract=6464099