Canada’s AI ambitions face an energy reality: Can the grid keep up?

Artificial intelligence is rapidly becoming a strategic economic priority for Canada. Data centres represent billions of dollars in capital investment. They can stimulate local economic development, support domestic AI capabilities, strengthen digital sovereignty, and drive investment in electricity infrastructure. Some analysts argue that large data-centre projects could act as “anchor tenants” that help justify new transmission lines, renewable generation, storage facilities and grid modernisation programmes.

As governments and technology companies race to develop domestic AI capacity, a less visible challenge is emerging behind the scenes: electricity. While concerns about data centres often focus on water consumption or land use, many analysts argue that Canada’s biggest obstacle to becoming an AI powerhouse is ensuring sufficient electricity supply, grid capacity, and access to low-carbon energy. The issue is straightforward. AI requires data centres, and modern AI data centres require vast amounts of power.

AI’s growing appetite for energy

Traditional data centres have long been part of Canada’s digital infrastructure. However, AI-focused facilities operate on a different scale. Training and operating large language models, running cloud-based AI services, and supporting advanced analytics all require dense clusters of high-performance computer processors that consume considerable electricity. [climateinstitute.ca], [cer-rec.gc.ca]

Globally, electricity demand from data centres is rising rapidly. The International Energy Agency has projected a significant increase in demand over the coming years, driven largely by AI workloads. Canada is expected to participate in this growth because of its relatively cool climate, strong telecommunications infrastructure, political stability, and access to clean electricity.

According to the Canada Energy Regulator, there are hundreds of data centres operating across the country, with concentration in major hubs such as Toronto and Montreal. Utilities and system operators are increasingly incorporating data-centre demand into long-term planning forecasts. More than 80% of Canada’s electricity comes from non-emitting sources, including hydroelectric, nuclear, wind and solar generation. This gives Canada a significant advantage over jurisdictions that depend heavily on fossil fuels. Furthermore, colder climates can reduce cooling requirements, improving energy efficiency and lowering operating costs.

Provincial governments and utilities are actively competing to attract data-centre investments. Hyperscale operators, including major technology firms, are evaluating Canadian locations for large-scale AI infrastructure. Alberta, Quebec, Ontario and British Columbia have each developed distinct strategies to capture a share of this growing market. Recent announcements illustrate the scale of these ambitions. Meta’s planned Alberta campus could eventually require up to 1.8 gigawatts of electricity, an amount comparable to the demand of a major city.

The grid capacity challenge

The problem is not that Canada lacks energy resources. The challenge is whether electricity can be delivered where and when it is needed. AI data centres require continuous, reliable power twenty-four hours a day. Unlike many industrial operations, interruptions are unacceptable. As a result, grid operators must not only generate enough electricity but also maintain sufficient transmission and distribution infrastructure to support new loads.

Many provinces are already anticipating substantial growth in electricity demand due to electrification of transport, industrial decarbonisation, heat pumps, and population growth. AI data centres add another major source of demand to an already crowded agenda. Canada’s federal government has therefore linked future AI development with broader electricity planning through its proposed National Electricity Strategy, recognising that transmission upgrades, generation projects, energy storage and inter-provincial grid connections will all be needed to support future growth.

Perhaps the most significant policy question is whether AI infrastructure could end up competing with other sectors for scarce low-carbon electricity. The electrification of transportation, building heating and manufacturing represents a cornerstone of Canada’s climate objectives. These sectors also require clean electricity to reduce greenhouse gas emissions. If AI data centres absorb large amounts of available generating capacity, concerns arise over whether homes, businesses and industrial decarbonisation projects may face delays or higher electricity prices.

The Canadian Climate Institute has warned that without careful policy design, data-centre growth could crowd out other forms of electrification or create incentives for additional fossil-fuel generation. Conversely, if managed correctly, data centres could help finance grid infrastructure that benefits the wider economy. This balancing act is becoming a defining policy challenge.

Alberta versus Quebec: two different approaches

Provincial responses highlight the complexity of the issue. Alberta has shown openness to models that allow data-centre operators to provide or finance their own electricity generation. While this may accelerate development, critics argue that reliance on natural gas generation could increase emissions and create challenges for climate objectives. Quebec, by contrast, benefits from abundant hydroelectric resources but has become increasingly cautious about allocating large amounts of low-cost electricity to data-centre operators. Policymakers are seeking to ensure that valuable clean electricity delivers maximum economic and societal benefit.

Ontario’s strategy centres around grid expansion, refurbishing existing nuclear assets, and developing new nuclear technologies such as small modular reactors. British Columbia is also assessing how best to allocate available power to high-demand projects.

In terms of innovation, future AI facilities may become more energy efficient. Developers are increasingly exploring advanced cooling technologies, demand-response programmes, and flexible operating models that could reduce pressure on electricity systems. Yet the more current, and difficult, question is whether electricity systems can expand quickly enough to support that growth while maintaining affordability, reliability and climate commitments.