Ottawa has told the companies building Canada’s next generation of data centres what responsible development should look like. Its new Responsible Data Centre Development Principles aim to protect Canadians from the costs and environmental pressures that can accompany these massive projects. Data-centre operators are expected to pay the electricity-system costs they create, conserve water, be transparent with host communities and ensure that their investments deliver lasting value to Canada.
Major technology companies have signed on. But the principles remain voluntary. If these are the minimum conditions Canada believes should govern infrastructure that will consume large amounts of electricity, water and land, why are they optional? Canada needs this investment. We need more domestic computing capacity for competitiveness and resilience, and AI capabilities that can help tackle difficult problems, including climate change, energy-system management, and the protection and restoration of nature. We need AI. But we need sustainable AI.
Every major economy is competing for this capital. Canada should compete by offering something increasingly valuable: clear and durable rules. Business is about risk-adjusted returns. Investors can price a known cost. What is much harder to price is uncertainty. A company considering a multibillion-dollar data-centre investment needs to know whether the electricity and water it will require will actually be available. It also needs to understand the environmental limits within which it must operate and how Indigenous rights and interests will shape the project. If restoration or other environmental investments will be required, those costs need to be understood before capital is committed.
Above all, investors need confidence that the basic rules will not fundamentally change after billions of dollars have been spent. Uncertainty is risk. Risk has a price.
Clear rules support investment
Making Ottawa’s principles binding could therefore be pro-investment, not anti-investment. Rules established before capital is committed allow proponents to design projects around them, price the costs and compare locations. They also reduce the risk of late-stage conflict and changing requirements.
We are already seeing the cost of not having that clarity. This week, Mississauga temporarily paused major new data-centre development for up to a year while it studies the sector and determines the right long-term planning approach. The city is not rejecting data centres; it is working out the rules before more projects proceed. That is understandable, but it also illustrates the value of establishing those rules earlier.
The answer is not another layer of overlapping project-by-project approvals. Canada should establish more of the environmental rulebook before individual projects arrive. The legal tool already exists. The federal Impact Assessment Act provides for regional assessments. Unlike a project review, which examines one proposal, a regional assessment can examine the combined effects of past, existing and future development across an area, including cumulative impacts.
In regions likely to attract clusters of major data centres, governments should use that mechanism to understand ecological capacity and establish credible regional limits. Federal, provincial and Indigenous governments could then translate those limits into clear, durable agreements and requirements. Developers would know the rules before they invest.
That is not red tape. That is investment certainty.
Go beyond minimizing harm
Ottawa’s principles call on data centres to minimize their environmental impacts. Canada should go further.
For generations, we have drawn economic value from forests, wetlands, watersheds, soils and biodiversity while deferring maintenance of those natural assets. Any executive understands deferred maintenance: postpone it long enough and the asset deteriorates, risks rise and the eventual repair bill grows. With nature, some losses can become irreversible.
Major data centres should therefore do more than compensate for the damage they cause. They should help improve the natural asset base on which our economy and well-being depend. A measurable biodiversity net gain should become part of the standard for major projects. This is not unprecedented. From November 2, nationally significant data-centre projects in England applying for development consent will be required to deliver at least a 10% biodiversity net gain. Canada need not copy its model, but the principle is worth adopting: government establishes the outcome, makes it binding and lets business determine the most efficient way to deliver it. That requirement would also direct investment into restoration and natural-asset improvement, creating local projects, businesses and jobs alongside the data-centre investment itself.
Technology companies may be more receptive than governments assume. Several already make commitments on water, climate and nature. Rather than relying on voluntary corporate ambition, Canada should give all proponents the same clear baseline.
We should not compete for AI investment by implying that our electricity, water, land and ecosystems are cheap or limitless. We should compete by offering resources managed responsibly, predictable decisions, Indigenous partnership and rules durable enough that an investment approved today will still make sense decades from now. Ottawa’s voluntary principles are a good start. Now make them mandatory.
In a world of growing ecological and policy uncertainty, Canada can offer investors something increasingly valuable: certainty.
Peter van Dijk is a former senior financial-services executive, an adjunct professor at Brock University and a PhD researcher in environmental sustainability at the University of Ottawa.
The Weekly Roundup
Get all our stories in one place, every Wednesday at noon EST.
