🔍 Read the full analysis: Canada’s Power Grid: The Unsung Hero Of AI Progress on ThorstenMeyerAI.com

Canada’s widely touted hydroelectric power resources are facing significant constraints, with recent provincial restrictions and regulatory disputes limiting the availability of cheap electricity for large data centers. This situation is closely related to the gigawatt gap in global AI infrastructure. This challenges the common assumption that Canada has an energy surplus ideal for supporting AI infrastructure, impacting both domestic and international AI development plans.

Over the past year, provinces like Quebec and British Columbia have implemented measures that restrict new power procurement for large data centers. Quebec, which has the largest hydroelectric capacity in Canada, has proposed a higher tariff of roughly 13 cents per kilowatt-hour for data centers exceeding 5 MW, nearly double the existing large-industrial rate of 6.82 cents. This proposal remains under regulatory review, contested by a coalition of data center operators, and has yet to be approved.

Similarly, British Columbia has allocated only 400 MW over two years for new data-center power, capped at 145 MW per project, a fraction of the capacity needed for major AI hubs. This highlights the importance of understanding the power bottleneck in AI data centers. Ontario and Alberta have different approaches: Ontario requires proponents to cover marginal connection costs, while Alberta has suspended certain regulations to encourage data-center growth but faces a queue of proposals exceeding 10 GW, with only about 1,200 MW expected to be connected by 2028. These restrictions illustrate that Canada’s hydro power is not as readily available as previously assumed, especially for large-scale AI infrastructure. The constraints are part of broader challenges in AI energy consumption and infrastructure resilience.

Canada’s total hydroelectric capacity is over 78 GW, with Quebec, BC, Ontario, Manitoba, and Newfoundland & Labrador supplying approximately 60% of national generation. However, the actual capacity to support new large data centers is limited by provincial policies, existing infrastructure, and regulatory hurdles. The country’s energy fundamentals—such as proximity to US markets, cold climate benefits, and ongoing nuclear expansion—remain strong, but the current constraints significantly temper the expected advantage for AI development.

Implications for AI Infrastructure Planning

The constraints on Canadian hydro power mean that the assumption of abundant, cheap, and accessible energy for AI data centers is no longer valid. This impacts not only Canada’s domestic ambitions but also international strategies that rely on Canadian energy resources. As global demand for AI infrastructure grows, regions with unrestrained power access will become more attractive, potentially shifting investment away from Canada. The regulatory and provincial restrictions highlight that energy availability is a key bottleneck, complicating efforts to scale AI computing capacity and influencing negotiations with European partners.

Furthermore, the current limits could lead to increased competition for power, higher costs, and delays in project timelines. For countries and companies planning to leverage Canadian hydro for AI, these developments suggest a need to reassess infrastructure and supply chain strategies, and to consider alternative energy sources or locations that can support large-scale data centers without regulatory hurdles.

Canadian Hydro Power and Growing Data Demands

Canada’s hydroelectric resources have long been viewed as a strategic advantage for supporting high-demand industries, including AI. With over 78 GW of installed capacity, the country produces roughly 60% of its electricity from hydro, making it one of the cleanest and most reliable sources globally. Quebec alone benefits from extensive hydro infrastructure, with costs as low as C$76/MWh in 2023, making it attractive for energy-intensive industries.

However, recent policy shifts reveal a different picture. Quebec has restricted new power procurement for large data centers since 2024, citing concerns over grid stability and the need to prioritize existing consumers. Hydro-Québec’s proposal for a higher tariff aims to limit growth but faces opposition from industry groups. Meanwhile, British Columbia’s limited allocation and provincial rationing further constrain the supply. These measures reflect a broader trend of managing existing capacity carefully, rather than expanding it freely, which contradicts the narrative of Canada’s hydro as an unlimited resource for AI growth.

Globally, data center power demand has surged from 59 GW in 2020 to 96 GW in 2024, with US markets like Virginia experiencing seven-year delays for grid connections. Europe’s major hubs—Frankfurt, Dublin, Amsterdam—are congested, and analysts warn that AI investments are likely to shift to regions with fewer restrictions. Canada’s current situation underscores that the scale of power needed for AI is a significant challenge, not just a matter of resource availability but also of regulatory and infrastructural capacity.

Unclear Impact of Regulatory Delays on Future Capacity

It remains uncertain how quickly and to what extent the current restrictions and regulatory disputes will be resolved. The approval process for Quebec’s higher tariff proposal is ongoing, and British Columbia’s limited allocations may be adjusted in response to industry needs or political pressures. Additionally, the potential for new infrastructure investments or policy shifts that could relax constraints is still unknown, making it difficult to predict Canada’s future capacity to support large AI data centers definitively.

Monitoring Policy Changes and Infrastructure Developments

The immediate next steps involve observing regulatory decisions in Quebec and British Columbia, as well as provincial government responses to industry demands. Industry stakeholders and international investors will be watching for any policy adjustments that could expand or further restrict power availability. Additionally, Canada’s federal government may consider new interconnection projects or incentives to alleviate constraints. Long-term, the focus will be on how these regulatory and infrastructural developments influence the global distribution of AI data-center investment and the strategic positioning of Canada’s energy resources.

Key Questions

How much hydro power does Canada currently have available for new data centers?

Canada has over 78 GW of installed hydro capacity, but current restrictions and regulatory delays limit the amount of power available for new large data centers, with Quebec proposing a higher tariff and other provinces capping allocations.

Why are Canadian provinces restricting power for data centers?

Provinces like Quebec and BC cite grid stability, existing commitments, and the need to prioritize current consumers as reasons for restricting new power procurement, which also helps control costs and manage infrastructure constraints.

What does this mean for global AI infrastructure plans?

The constraints in Canada could shift investment toward regions with fewer restrictions, such as the US or parts of Europe, potentially altering the global distribution of AI data centers and impacting the scalability of AI models.

Could policy changes relax these restrictions in the future?

It is uncertain; regulatory processes are ongoing, and political or economic pressures could lead to changes. However, current trends suggest a cautious approach to expanding capacity.

What are the broader economic implications of these constraints?

Limits on power availability could increase costs for data-center operators, slow AI development, and potentially raise electricity prices for consumers, influencing national competitiveness and technological progress.

Source: ThorstenMeyerAI.com