Artificial-intelligence (AI) data centres have become increasingly vital across various sectors, including science, business, national defence, and health care, highlighting the importance of reliable computational infrastructure in Canada. These facilities support advanced AI applications that improve health-risk detection, accelerate software development, and lower barriers for complex research, underscoring the need for dependable data storage, processing capabilities, and domestic expertise.

Unlike traditional factories that must be physically located where production occurs, data centres’ geographical placement is less critical since computational services can be supplied remotely. However, the broader economic benefits depend heavily on developing domestic skills in building, managing, and improving these facilities—a phenomenon known as innovation spillover. Furthermore, data centres have taken on strategic importance, exemplified by the recent U.S. decision to restrict access to Anthropic’s AI models, illustrating how geopolitical considerations can directly impact the availability of AI tools.

Given the increasing integration of AI into societal functions, Canada faces risks in depending heavily on foreign-controlled AI infrastructure and technologies. While the country cannot dictate how other nations manage access to AI resources, reducing reliance on external actors remains a priority for national innovation and sovereignty.

Despite their importance, data centre projects often encounter resistance at the local level due to concerns over increased energy costs, noise pollution, and limited permanent job creation compared to other capital-intensive developments. Additionally, the economic benefits are often diffuse, with profits tied to intellectual property held outside Canada, leading to skepticism about tangible advantages for local communities. Recent polling indicates widespread public support for domestic AI infrastructure in principle, but strong opposition exists when proposed sites are near residents’ own communities. This reflects a fundamental issue where costs are localized while benefits are broadly dispersed.

To address these challenges, experts suggest a policy framework that ensures Canadians capture the strategic and innovative gains justifying data centre construction. This includes prioritizing projects with Canadian ownership stakes or partnerships and enforceable commitments to allocate computing resources to domestic researchers, startups, and public institutions. While Canada remains open to global investment, greater emphasis is placed on what domestic stakeholders learn from these ventures and on maintaining control over data and processing during times of geopolitical tension.

Another recommendation is that data centres should not impose higher electricity costs on local residents. Operators should either provide their own power or finance equivalent new supply and grid capacity, matched closely to their real-time usage to prevent peak-hour price spikes. Although initiatives like “bring-your-own-power” can mitigate some concerns, opposition based on environmental and community impact persists, as demonstrated by Manitoba Premier Wab Kinew’s recent rejection of a data centre proposal involving natural gas turbines.

Fiscal arrangements also require reform. Municipalities currently face bidding wars that often erode their tax base through local incentives. A coordinated federal approach to subsidies and tax credits could provide a fairer, consistent system that allows local governments to retain full property-tax revenues and receive additional federal support tied to hosting data centres. Such arrangements could improve community acceptance by directly linking data centre presence with public benefits like infrastructure and services.

Finally, addressing multinational profit shifting is critical to ensuring communities gain a more visible share of the economic benefits. While Alberta’s 2 percent levy on computer hardware is straightforward to administer, it has drawbacks, such as taxing intermediate inputs and targeting a single industry. Alternative methods, including formula-based profit apportionment requiring international cooperation, have been proposed but present complex challenges.

The opposition to data centres is seen as understandable given current conditions, but experts argue that blocking construction outright is not the solution. Instead, a more balanced approach that aligns incentives, protects local interests, and secures national strategic goals is needed to realize the full potential of AI infrastructure development in Canada.