AI Data Centre Pipeline Concentrates in Alberta, Challenging Grid Capacity and Policy

Canada’s planned data-centre capacity for AI workloads has ballooned to nearly ten times the existing infrastructure, according to a map and research from The Globe and Mail and York University. This rapid expansion is primarily driven by large-scale AI training and inference needs, marking a significant shift away from the web hosting that defined previous Canadian server farms.

This geographic shift creates immediate policy pressure. While provinces that built Canada’s initial data-centre sector, including Quebec and Ontario, are beginning to manage grid access more cautiously for new oversized loads, Alberta is now attracting the bulk of future industrial growth. The infrastructure demands mirror those of its energy sector, meaning existing fights over water use, land allocation, and Indigenous consultation are reappearing around server farms instead of pipelines.

The scale of these projects is immense. For example, a Bell-backed AI data centre near Regina, Saskatchewan, has been expanded fourfold from 300 megawatts to a planned 1.2 gigawatts. This expansion also increases permanent employment projections from 80 positions to 500 as the project scope grows.

For investors and policy teams, this means that Alberta is becoming a critical pillar for Canada’s AI diversification strategy, though one heavily reliant on abundant, cheap energy rather than entirely new economic bases. The policy challenges surrounding grid capacity in these remote industrial zones are now central to Canadian tech infrastructure.