Canada wants to be an AI leader. Whether it can achieve that ambition may depend as much on substations, cooling systems and grid connections as it does on algorithms, computing power and talent.
The rapid growth of AI is forcing data centre operators and enterprises to rethink how they plan for power, cooling, resiliency and future capacity. Infrastructure that was once treated as a back-end consideration is increasingly becoming a deciding factor in where projects are built—and whether they can proceed at all.
For James See, National Sales Director – Systems at Schneider Electric Canada, the challenge is not simply generating enough electricity. Canada must also ensure power can be delivered where and when it is needed while encouraging infrastructure that operates efficiently, supports the grid and delivers value to surrounding communities.
DataCentre.ca spoke with See about whether Canada’s physical infrastructure is keeping pace with its AI ambitions, the growing importance of energy in data centre site selection, and how efficiency, technology and public-private collaboration can help the country scale responsibly.
Canada is making AI a national priority, but AI ultimately needs physical infrastructure—power, cooling, connectivity, and resilient facilities. From your perspective, is Canada thinking seriously enough about the infrastructure layer behind its AI ambitions?
JS: Canada has positioned itself as a major player in AI research and innovation, but in many of the conversations we’re having with customers, there’s still a gap between AI ambition and infrastructure readiness.
AI may be a digital transformation, but it is fundamentally powered by physical infrastructure. Every model, every workload, depends on reliable power, efficient cooling, and resilient facilities. What AI is doing is accelerating demand across all of those dimensions at a pace the industry hasn’t experienced before, and that’s where the next phase of Canada’s AI strategy will be won or lost.
The reality is that AI is creating a step change in demand for electricity, cooling, and digital infrastructure. Training and deploying advanced AI models requires significantly more power density than traditional IT workloads, and that has implications for everything from grid planning to data centre design. What is encouraging is that governments, utilities, and industry are increasingly recognizing this challenge. Ontario’s proposed Data Centre Playbook is one example of infrastructure planning moving closer to the centre of AI and economic policy.
The next step is ensuring that these frameworks create clarity and predictability while encouraging projects that use energy efficiently, strengthen the electricity system and deliver meaningful economic and community benefits. Schneider Electric is contributing to that effort by helping organizations modernize electrical infrastructure, improve energy efficiency, and build more resilient, AI-ready facilities.
AI may be digital, but its success will depend on very physical foundations. The organizations moving fastest are the ones aligning infrastructure planning early with their digital strategy.
AI changes the scale and speed of demand—but it also changes how early infrastructure decisions need to be made.
As AI workloads grow, data centres are becoming some of the most energy-intensive pieces of the digital economy. What are you seeing in Canada in terms of demand from data centre operators and enterprises preparing for AI?
JS: We are seeing a significant shift in how organizations think about capacity planning. Historically, data centre operators focused on incremental growth. AI is changing that dynamic.
Operators are now evaluating how to accommodate much higher-density computing environments, while enterprises are assessing whether their existing facilities can support AI-enabled applications and workloads. Questions around power availability, cooling capacity, backup power, and energy efficiency are moving to the forefront much earlier in project planning. What’s particularly notable is the speed at which demand is evolving.
Organizations don’t just want more capacity, they want infrastructure that can scale quickly, operate efficiently, and remain resilient as workloads continue to grow. As a result, we are seeing increased interest in modular infrastructure, advanced cooling technologies, and digital tools that provide greater visibility into energy and operational performance. For many organizations, infrastructure is no longer a back-end consideration—it’s becoming a gating factor for AI deployment.
Do you believe Canada’s electricity grid is ready for the next wave of AI-driven growth, or are there specific bottlenecks that need to be addressed first?
JS: Canada’s electricity system benefits from a relatively low‑carbon generation mix, with about 84% of electricity generation coming from non‑emitting sources such as hydro, nuclear, wind, and solar, according to the Government of Canada.
From what we’re seeing across projects, the biggest constraint isn’t necessarily generating electricity; it’s ensuring that power can be delivered where and when it’s needed. In many areas, utilities are facing growing demand from electrification, population growth, industrial development, and now AI-related infrastructure.
Grid interconnection timelines, transmission constraints, and local capacity limitations are emerging as critical factors for data centre developers.
These challenges are not unique to Canada, but they are becoming more visible as AI accelerates demand. Addressing them will require long-term planning and closer coordination between utilities, governments, and the private sector.
The reality is that AI is not just a compute challenge – it’s an energy and infrastructure challenge at scale.
Technology can also help data centres become more flexible and grid-aware. Greater visibility, distributed energy resources, energy storage, microgrids and intelligent energy management can help facilities manage demand, strengthen resiliency and, where market rules and local infrastructure allow, support the broader electricity system. The goal should not simply be to connect more load, but to develop infrastructure that understands and actively manages its impact on the grid.
Canada has an advantage in clean electricity, but access to power is not evenly distributed across the country. How should data centre operators think about site selection in a market where energy availability may become as important as connectivity?
JS: Site selection is evolving quickly. While connectivity has historically been the primary driver, we’re now seeing power availability become a determining factor in many projects. In some cases, projects are being delayed or even re-scoped based on access to power alone.
Data centre developers need to take a more holistic view, looking at grid capacity, utility expansion plans, permitting timelines, long-term energy reliability and sustainability objectives. They also need to consider how a project will affect local infrastructure and how it will demonstrate tangible value to its host community. It is no longer only about where capacity exists today, but where a facility can operate, expand and maintain community confidence over time.
This is also leading to increased interest in secondary markets or emerging regions where infrastructure can support long-term growth. The most successful projects will be those that balance connectivity, energy access, sustainability, and resiliency from the outset rather than treating them as separate considerations.
The question has shifted from “Where can we build?” to “Where can we build and scale reliably over the next decade?”. Also, A technically viable site is not automatically a socially or economically viable site.
What role can energy efficiency play in helping Canada scale AI infrastructure without overwhelming local grids or slowing down new data centre development?
JS: From a practical standpoint, energy efficiency is one of the fastest ways to unlock additional capacity. Every kilowatt that can be saved through more efficient infrastructure reduces pressure on the grid and allows existing capacity to support more digital growth. Even incremental improvements can translate into meaningful capacity at scale, especially in high-density AI environments.
We are seeing innovation across the entire data centre ecosystem, from more efficient power distribution and UPS systems to direct-to-chip liquid cooling for high-density AI environments. Closed-loop and water-conscious cooling approaches, noise-reduction engineering and opportunities for waste-heat recovery can also help reduce impacts beyond electricity use.
Digital monitoring and analytics are equally important because operators need to measure performance, identify inefficiencies and show regulators and communities how facilities are performing over time.
The combination of electrification, digitalization, and real-time visibility is what allows efficiency improvements to scale—not just remain incremental. In many cases, the most sustainable megawatt is the one you never have to generate.
For businesses that are not data centre operators but are adopting AI, automation, and more electrified operations, what should they be doing now to prepare their facilities and energy systems?
JS: Many organizations are focused on the software side of AI adoption, but it’s equally important to assess whether their facilities are ready to support increased digital activity.
The priority should be threefold: understanding existing capacity, improving visibility into energy use, and planning for resiliency and future load growth. As digital workloads increase, the demand on electrical infrastructure grows—often beyond what facilities were originally designed to support.
Organizations are also recognizing that as operations become more digital, even short disruptions can have significant business impacts. That’s driving greater investment in monitoring, backup power, and energy management systems.
Tools that provide real-time energy visibility are becoming a starting point for many customers, helping them make informed decisions and avoid constraints later.
The organizations that prepare early will be better positioned to scale AI adoption without unexpected infrastructure limitations.
Resiliency is becoming a bigger issue as businesses become more dependent on digital systems. How should companies balance sustainability, uptime, and cost as they modernize their infrastructure?
JS: Historically, organizations often viewed these priorities as competing objectives. Increasingly, we are seeing that they can reinforce one another when approached strategically.
Modern infrastructure solutions allow organizations to improve energy efficiency, strengthen resiliency, and manage costs simultaneously. Better visibility into operations enables more informed decisions, while predictive maintenance and digital monitoring can reduce downtime and improve asset performance.
The key is taking a lifecycle approach rather than focusing solely on upfront costs. Investments that improve reliability and efficiency often deliver meaningful operational savings over time while helping organizations meet their sustainability goals.
The organizations that succeed will be those that view resiliency, sustainability, operational performance and grid impact as part of the same conversation. Increasingly, infrastructure needs to deliver value not only within the facility, but also operate responsibly within the electricity system and community around it.
What kind of public-private collaboration will be needed to ensure Canada can support long-term AI growth—not just in major hubs, but across regions and industries?
JS: From what we’re seeing, no single stakeholder can address this challenge independently.
Supporting AI growth at scale will require coordination across governments, utilities, technology providers, infrastructure developers, educational institutions and host communities. Areas like grid expansion, permitting, workforce development, and infrastructure investment all need to be aligned. Just as importantly, project assessment needs to be clear, evidence-based and predictable so that proponents understand what is expected and communities can evaluate benefits and impacts with confidence.
There is also an opportunity to think more broadly about regional growth. AI adoption will not be limited to major technology hubs. Manufacturers, healthcare organizations, educational institutions, and businesses across the country will increasingly rely on AI-powered technologies.
Supporting that growth means building infrastructure that is scalable, resilient, and accessible across regions. Canada has many of the ingredients needed to become a global AI infrastructure leader, including clean energy resources, technical expertise, and a strong innovation ecosystem. The opportunity now is to connect those strengths together through sustained collaboration and long-term planning.
The potential is clear—but how quickly we scale will depend on how effectively we align infrastructure, policy, investment, and public confidence.