Wednesday, July 22, 2026
Every time you prompt an AI chatbot or stream a video, a physical server somewhere generates a blast of heat. To keep the internet from melting, the global tech sector is rushing to build hyperscale data centres and they are eyeing Canadian communities for their cheap land, power, and abundance of freshwater.
For the people running Canada’s water systems, this is no longer a hypothetical tech trend. A single large data centre can demand millions of litres of water a day. That is a massive operational strain that our 50-to-100-year infrastructure lifecycles and decades-old light industrial zoning bylaws were never designed to handle.
The friction hit a boiling point when Hamilton City Council narrowly defeated a first-in-Canada, one-year moratorium on data centres in a tense 10-6 vote. While the pause button wasn’t pushed, the message to water managers across the country was clear: federal and provincial guardrails do not exist, applications can land on your desk tomorrow, and local utilities are completely on the front lines of protecting community resources.
To get past the corporate sustainability hype and look at the hard fluid dynamics, infrastructure costs, and wastewater realities, we sat down with Dr. Anne Pasek. As an Associate Professor at Trent University and Canada Research Chair, she is one of the few independent experts pulling back the curtain on the actual material footprint of the AI gold rush.
Water Canada: The recent policy debate in Hamilton exposed a major regulatory gap. What does it say about Canadian environmental governance that individual municipalities are having to step in where federal and provincial regulations are absent?
Dr. Pasek: It’s an indictment of the current system. The federal AI strategy prioritizes industry growth without considering the infrastructure or who bears the burdens. It assumes existing light-industrial regulations are sufficient, but hyperscale data centres—which have grown exponentially in the last decade—are substantively different from the facilities those bylaws were designed to handle. It is not surprising that cities, as the front line of planning, are the ones forced to seek remedies. This could have been mitigated had federal policymakers mandated environmental reviews and community outreach as conditions for accessing public AI funds.
Water Canada: When speaking with major metropolitan utilities, data centre draws are often a fraction of capacity. But in smaller municipalities, a single facility could demand 20 per cent of the daily budget. How does this localized reality distort the conversation?
Dr. Pasek: This debate isn’t just about engineering constraints; it’s an interrogation of values. It asks whether the industrial use case is worth the water consumed. This is particularly sharp in water-stressed regions like Alberta, where developers demand constant, non-curtailable access—clashing with public conservation needs during peak stress.
Water Canada: When a data centre uses millions of litres of water a day, who absorbs the capital cost of municipal upgrades? Are developers paying their fair share?
Dr. Pasek: In almost all jurisdictions, there is no accounting structure for the systemic wear and tear caused by these massive loads. Unlike energy, where “ring tariffs” can force companies to pay for upgrades, water management often relies on “water positive” pledges—fixing leaky pipes or evaporation issues. These are soft promises, rarely commensurate with the scale of the development. Because data centres provide low per-square-foot job creation, this is provoking a necessary rethinking of how municipalities evaluate the balance of benefits and harms.
Water Canada: What is the capital risk to a utility that expands its water treatment infrastructure for this boom, only to end up with stranded assets if the tech landscape shifts?
Dr. Pasek: We are in a bubble. If a utility takes on debt to service these loads and the bubble pops, the municipality faces a significant revenue and physical infrastructure problem—an empty, windowless concrete box is difficult to repurpose. City planners should treat these deep-pocketed tech companies as stepping stones to desired municipal system goals, rather than letting the developers drive the infrastructure agenda.
Water Canada: In your community toolkit, what are the primary engineering and operational blind spots to look out for when a developer pitches a project?
Dr. Pasek: First, push back on Non-Disclosure Agreements (NDAs); they poison public trust and prevent scrutiny. Microsoft’s move to pivot away from NDAs suggests they aren’t mission-critical. Second, demand clear timelines where developers are forced to lock into a specific plan; proposals are often kinetic and redrafted constantly. Finally, be wary of the “disappearing waste” model. Smaller municipalities often lack the expertise to evaluate the toxicological and thermal qualities of industrial discharge, and “trucking waste off-site” is not a long-term solution.
Water Canada: Tech companies often promise best practices or efficiency to ease concerns. How should municipal engineers and water advocates critically evaluate these metrics?
Dr. Pasek: “Best practices” has no coherent definition in this industry and is often a distraction. We should move away from arguments about efficiency percentages—which can still result in a net overdraw of local capacity—and look exclusively at total, system-wide water volume numbers to ensure they align with the community’s long-term sustainability limits.
This Q&A is the second instalment in our summer series on the intersection of water and data centres. Stay tuned as we continue to explore the critical challenges facing Canada’s utilities—and be sure to catch the rest of the series as it rolls out on Water Canada.
Read the first article in the series: Hamilton Council rejects landmark data centre pause, but “the work will still go ahead
