Canada has spent generations as hewers of wood and drawers of water, exporting raw materials such as oil, logs and electricity. The AI boom gives us an opportunity to move up the value chain to export compute.
AI requires enormous amounts of electricity, cooling, land and infrastructure. Canada happens to have an unusual abundance of all four. We have hydroelectric power, enormous natural gas reserves, freshwater resources, a cooler climate, vast amounts of land and immediate access to the US market. Canada may not produce the world’s dominant AI model, but we can become one of the best places in the world to run them.
The Data Centre Backlash
AI data centres are facing growing public opposition across North America. Communities are concerned about electricity consumption, water use and noise. They also question why residents should bear higher infrastructure or energy costs so some of the world’s largest technology companies can operate enormous computing facilities.
Canada shouldn’t dismiss those concerns. It should design its data-centre strategy around solving them.
Protect the Water Supply
Large data centres should be required to use closed-loop or similarly low-consumption cooling systems wherever technically feasible.
Canada’s freshwater resources are an advantage, but that doesn’t mean they should be consumed without limits. The goal should be to use Canada’s climate and water resources efficiently while minimizing ongoing withdrawals from local supplies.
Data Centres Should Pay for Their Power
Canadians should not pay higher electricity prices because hyperscalers are consuming enormous amounts of power.
Data centres should be encouraged where genuinely excess electricity exists without materially increasing costs for existing customers. Where additional capacity is required, developers should help finance it.
For the largest projects, governments could establish public-private partnerships linking new data centres directly to new power generation and grid infrastructure. A company that wants gigawatts of electricity should help create gigawatts of new supply.
That could turn AI demand into a catalyst for expanding Canada’s electricity system rather than a competitor for its existing capacity.
Put Data Centres Where They Belong
Data centres are industrial infrastructure. They shouldn’t be built beside residential neighbourhoods.
Canada has no shortage of land. Large facilities can be located in appropriate industrial areas with significant buffer zones, while requiring noise suppression for cooling equipment, backup generators and other infrastructure.
We don’t need to choose between AI infrastructure and quality of life. We need to build the infrastructure in the right places.
From Megawatts to Compute
Canada already has many of the resources the AI economy desperately needs, energy, water, land, and a cold climate.
The opportunity is to use those advantages to capture more of the AI value chain while ensuring communities don’t bear the costs.
Canada has the resources to become a global leader in responsible, high-value computing. By turning more of our energy, land and infrastructure into compute, we can create jobs, strengthen our economy and export a valuable finished product to the world.