Canada’s ambitions to become a global leader in artificial intelligence infrastructure are running into a hard physical limit: electricity.
A new report from TD Economics, authored by economist Likeleli Seitlheko, warns that while demand for data centres is surging across the country, the ability to power them is not keeping pace. The result is a growing disconnect between Canada’s AI aspirations and the underlying infrastructure required to support them.
Nowhere is that tension more visible than in Alberta. The province has received requests for roughly 21 gigawatts of data centre capacity—well above its current peak electricity demand of 12.8 gigawatts. Yet only a fraction of that demand can realistically be connected in the near term. According to TD, just 1.2 gigawatts of new capacity is expected to come online by 2028.
The issue is not a lack of interest or capital. It is access to the grid.
Across Canada, utilities and system operators are facing mounting pressure from a wave of proposed data centre developments, many tied to artificial intelligence workloads that require significantly more power than traditional facilities. While legacy data centres typically consumed less than 50 megawatts, modern AI clusters can demand 100 megawatts or more—sometimes approaching the scale of small cities.
This shift is fundamentally changing how infrastructure is planned and built. Where proximity to users and network latency once dictated site selection, power availability has become the primary constraint. Developers are now competing for access to limited grid capacity, creating bottlenecks that are delaying or outright stalling projects.
Alberta is not alone. Jurisdictions across Canada—including Ontario, Quebec, and British Columbia—are experiencing similar pressures, as electrification trends collide with the rapid rise of AI-driven compute demand. Globally, the challenge is even more pronounced, with some regions in Europe facing multi-year delays and outright moratoriums on new data centre connections due to grid limitations.
For Canada, the implications are significant.
The federal government has made sovereign AI infrastructure a strategic priority, positioning domestic compute capacity as essential to economic competitiveness, innovation, and data sovereignty. But without sufficient power, those ambitions risk remaining largely theoretical.
Seitlheko notes that addressing the gap will require closer coordination between governments, utilities, and developers, along with significant investment in generation and transmission infrastructure. At the same time, policymakers will need to balance industrial growth with the need to maintain affordable electricity for households and businesses.
The stakes are high. Global investment in AI and data centre infrastructure is expected to reach into the trillions over the coming years, and countries that can deliver both compute capacity and reliable power will be best positioned to capture that growth.
Canada has the demand. It has the capital. And it has the policy momentum.
What it does not yet have—at least at the scale required—is the power.nd 100 megawatts or more—sometimes approaching the scale of small cities.
This shift is fundamentally changing how infrastructure is planned and built. Where proximity to users and network latency once dictated site selection, power availability has become the primary constraint. Developers are now competing for access to limited grid capacity, creating bottlenecks that are delaying or outright stalling projects.
Alberta is not alone. Jurisdictions across Canada—including Ontario, Quebec, and British Columbia—are experiencing similar pressures, as electrification trends collide with the rapid rise of AI-driven compute demand. Globally, the challenge is even more pronounced, with some regions in Europe facing multi-year delays and outright moratoriums on new data centre connections due to grid limitations.
For Canada, the implications are significant.
The federal government has made sovereign AI infrastructure a strategic priority, positioning domestic compute capacity as essential to economic competitiveness, innovation, and data sovereignty. But without sufficient power, those ambitions risk remaining largely theoretical.
TD Economics suggests that addressing the gap will require closer coordination between governments, utilities, and developers, along with significant investment in generation and transmission infrastructure. At the same time, policymakers will need to balance industrial growth with the need to maintain affordable electricity for households and businesses.
The stakes are high. Global investment in AI and data centre infrastructure is expected to reach into the trillions over the coming years, and countries that can deliver both compute capacity and reliable power will be best positioned to capture that growth.
Canada has the demand. It has the capital. And it has the policy momentum.
What it does not yet have—at least at the scale required—is the power.