The Real Economy

Canada’s AI power play: Why data centre investment could be a global game-changer

Businesses and governments must plan now to capitalize on the moment

September 29, 2026
#
Economics The Real Economy

As the global race to own artificial intelligence intensifies, Canada can differentiate itself thanks to two critical components: data centres and abundant domestic energy resources to power them.

AI runs on data centres, and global data centre capacity is growing by double digits annually amid the seemingly insatiable demand for AI.

That growth curve is projected to steepen in the coming years—and so is the projected growth in demand for electricity and water to power these data centres.

Canada has a clear competitive advantage: abundant and affordable electricity and water, available land, a cool climate, and a relatively speedy permitting process that doesn't make investors wait years to break ground. It also helps that Canada was named the most energy-efficient country in the world to build data centres.

Compared to the U.S. and its lengthy and complex approval processes or Europe’s pricey electricity, it is easy to see why Canada is such an attractive host for data centres.

Canada appears ideally positioned to capitalize on this opportunity, but it must approach the situation with intentionality—especially when it comes to cost and clean energy considerations.

Businesses should explore building and scaling their own AI capabilities domestically instead of just being hosts for foreign companies, in order to make the most of the emerging infrastructure. Meanwhile, government policies must be carefully designed for businesses and consumers to capture the upside of the benefits instead of just footing the water bill and grid strain while the profits flow elsewhere.

The lay of the land

While Canada’s current data centre footprint is modest, the announced and under-construction pipeline of data centres is roughly 14 times larger than existing capacity.

Ninety-two per cent of that pipeline is in Alberta, largely due to its deregulated electricity market and a provincial government actively courting the tech industry.

Not every proposed data centre will be built, and some of that 14-times-bigger pipeline is speculative. Due to phantom projects and long-shot pitches, only about 28 per cent of proposed data centres are projected to ever materialize.

That said, data centres are not a passing trend. They are poised to become permanent infrastructure that will be critical for Canada’s AI development and its role as a technological innovator.

The infrastructure catch

Existing capacity does not guarantee future capacity.

AI's appetite for electricity is growing faster than most contemporary grids were built to handle. Having cheap power today does not automatically translate to the grid’s ability to absorb the insatiable demand from future data centres.

That kind of capacity requires sizable and thoughtful investments in infrastructure.

More data centres mean more strain on a grid that's already being asked to decarbonize and expand at the same time—two often-contradictory objectives.

And if history is any indicator, the grid’s capacity is growing at a rate much slower than what would be required to accommodate the projected surge in demand from future data centres.

Energy abundance is also not synonymous with cleanliness. Canada’s electricity generation is mostly renewable, with hydro comprising 58 per cent of total generation in May 2026, followed by natural gas at 19 per cent and nuclear at 13 per cent.

While the federal government recently announced an agreement on Churchill Falls and other electricity projects in Newfoundland and Labrador valued at about $70 billion, which it touted as “the largest clean energy investment in North American history,” Canada still has a long way to go in terms of fully decarbonizing the grid.

That consideration is especially relevant for these projects because, although hydroelectricity is technically renewable, it has more negative environmental effects than solar and wind power. 

Severe and frequent natural disasters, amplified by climate change, have become increasingly costly—with insured damages exceeding $2.4 billion in Canada in 2025, according to the Insurance Bureau of Canada.

This presents an interesting paradox: If this new and expanded electricity infrastructure and future data centres are to last, they must be built to be climate-resilient.

Data centres’ environmental impact is twofold: They use up enormous amounts of electricity, and they enable more emissions from increased oil and gas extraction.

Recent research suggests that AI’s biggest climate impact comes from productivity gains in the oilfield rather than data centres. While this could be promising for the energy industry, it comes with a hefty bill in terms of climate considerations.

The cost-benefit dilemma

The benefits of data centres are dispersed and global, whereas costs are local and borne by communities where they are located.

Consumers, businesses and global governments stand to benefit from AI development—particularly with the productivity gains realized across fields ranging from logistics and data analysis to health care.

Canadian businesses have a chance to capitalize on the fact that their country is a compelling place to build data centres.

Currently, ownership of data centres is largely foreign—meaning a sizable share of the profits is leaving Canada altogether. While small and medium-sized data centres are primarily Canadian-owned, large data centres are predominantly owned by U.S.-headquartered firms.

An increase in Canadian-owned data centres would allow domestic businesses to propel themselves forward in their tech and AI journeys instead of acting as hosts for foreign tech firms.

As for community considerations, benefits to the local economy arise primarily in the construction phase. Once built, data centres are capital-intensive projects; only hyperscale facilities are meaningful job creators because they require a supporting telecommunications ecosystem.

Other potential benefits include know-how, the spillover effect of knowledge and the possibility of developing tech centres. Data governance is another benefit, especially in this era of geopolitical volatility.

The costs, on the other hand, are hyperlocal: higher electricity prices, grid strain, water use  and noise. It is unsurprising that opposition to data centres has emerged in some communities—and municipal and provincial governments are beginning to respond.

Policies to meet the moment

Federally, the Canadian Sovereign AI Compute Strategy has put a $15 billion incentive package on the table to build out domestic computing capacity and infrastructure.

The federal government also launched a new national AI strategy to introduce legislation, investments and programs to help drive AI adoption in a way that benefits Canadian households and businesses.

Ontario is proposing that large data centres (above 1 megawatt) pay a premium rate, with closed-loop cooling systems getting fast-tracked for approval. The goal is to keep costs of water and electricity off the backs of Ontario households.

Most recently, Oakville, Ont., became the first municipality to enact a one-year moratorium on data centres, with Mississauga following suit.

Alberta is letting data centres bring their own power through on-site natural gas generation, solving the grid-strain problem by having operators sidestep the grid entirely.

But policy can only go so far. Looking ahead, it’s incumbent on businesses and governments to make the most of this emerging opportunity by building this infrastructure in a way that allows companies to fully utilize it to grow their technological capabilities.

RSM contributors

Subscribe to The Real Economy Canada

An economic outlook for Canada and insights into key issues Canadian businesses are facing today

 

More from The Real Economy

The Real Economy

The Real Economy Canada

A quarterly economic report for middle market business leaders.

Industry insights

Industry-specific insights for the middle market.