September 13, 2026
by Arnie Gess, president of Construction Links Network
Artificial intelligence may exist in the digital world, but the infrastructure required to power it is decidedly physical. Behind every AI model are data centres requiring enormous amounts of electricity, cooling, electrical equipment, concrete, steel, specialized engineering and skilled construction labour. As investment in AI infrastructure accelerates, a new question is emerging for Canada’s construction industry: Can we build the physical infrastructure fast enough to keep up?
Canada is actively trying to expand domestic AI computing capacity. The federal government now describes data centres as critical infrastructure and, on September 3, 2026, introduced Canada’s Responsible Data Centre Development Principles. The framework calls for projects to create lasting community benefits, cover infrastructure costs attributable to their development, minimize water and environmental impacts, and deliver strategic value to Canada.
The challenge is that constructing the next generation of AI infrastructure involves much more than building server facilities.
AI workloads require significantly more computing power, making electricity availability one of the defining issues facing new data centre development.
The Canada Energy Regulator (CER) now explicitly models data centre growth in its long-term electricity outlook. Its baseline scenarios assume an additional 1.5 GW of Canadian data centre load by 2030 and 3.5 GW by 2050, with much of the growth occurring in Ontario, Alberta and Quebec. Its higher-growth scenario reaches 12 GW by 2050.
Under that higher scenario, data centres account for roughly 100 TWh—or about one-third—of Canada’s projected growth in end-use electricity demand through 2050.
That means Canada’s AI infrastructure challenge quickly becomes an energy and construction challenge.
New facilities may require generation capacity, transmission infrastructure, substations, transformers, switchgear, backup power, cooling systems, water infrastructure, roads and other supporting construction before the computing equipment can even operate.
Few projects demonstrate this transformation better than Meta’s new AI-optimized data centre campus in Sturgeon County, Alberta.
The project represents an investment of more than $13 billion and is expected to provide 1 GW of computing capacity. Construction is underway and scheduled through 2029, with approximately 3,000 construction workers expected at peak activity.
The campus is also paired with a dedicated 932 MW natural gas power plant, while Meta has committed approximately $60 million toward local road and water infrastructure improvements.
And it isn’t an isolated proposal.
Alberta’s project pipeline includes other large proposed data centre developments.
Another proposed project, Malachite One in Yellowhead County, could include a data centre costing up to $15 billion, supported by a power-generation facility capable of producing one gigawatt of electricity.
The scale of these projects illustrates why AI infrastructure is becoming relevant far beyond the technology sector.
Electricity is increasingly influencing where data centres can be built and how quickly projects can proceed.
Canada’s new federal principles state that data centre developers should pay their proportionate costs for new generation, transmission, substations and grid upgrades attributable to their projects rather than shifting those costs to households and existing businesses. Projects should also avoid compromising grid reliability and, where necessary, contribute new supply, storage or other infrastructure.
For construction, that potentially creates an enormous secondary market.
The AI infrastructure boom isn’t simply creating demand for data centre contractors. It can generate work for utilities, civil contractors, electrical contractors, mechanical contractors, engineering firms, equipment suppliers, energy developers and companies involved throughout the construction supply chain.
In other words, the data centre itself may only be one component of a much larger construction opportunity.
Canada also has to find the people capable of delivering these projects.
Construction already employs more than 1.6 million Canadians, while federal labour-market information identifies the sector as having one of the country’s highest projected levels of occupational shortages.
Data centres intensify the issue because many systems require specialized expertise.
Electricians, power-system specialists, mechanical contractors, controls technicians, engineers, heavy-equipment operators and commissioning professionals may increasingly find themselves competing for labour with housing, transportation, energy and other major infrastructure projects.
The federal government created an Energy and Electricity Workforce Alliance in September specifically as Canada faces workforce pressures while expanding generation and modernizing its electricity systems.
AI may therefore create a somewhat unexpected constraint: computing capacity can only expand as quickly as physical infrastructure and the people capable of building it.
Those constraints also represent opportunity.
Contractors that develop expertise in mission-critical construction, electrical infrastructure, advanced cooling, power generation and complex project delivery could find themselves positioned in one of construction’s fastest-developing markets.
The opportunity extends to manufacturers and suppliers as well. Transformers, generators, switchgear, cooling equipment, electrical components, prefabricated systems, concrete products and specialized construction technology all form part of the physical ecosystem behind AI.
Canada also has a strategic reason to build more of this infrastructure domestically. The federal government says the country currently relies significantly on computing, cloud and data-storage infrastructure located outside Canada and is pursuing additional sovereign AI capacity.
That potentially makes AI infrastructure development both an economic opportunity and a national infrastructure priority.
The AI revolution is often described in terms of algorithms, chips and software.
Construction sees another side of it.
Every expansion of computing capacity ultimately requires physical assets: land, power, buildings, cooling, electrical systems, roads, equipment and skilled people.
Canada therefore faces two related challenges. It must attract the investment necessary to participate in the global AI economy while ensuring data centre development does not overwhelm electricity systems, water resources, communities or an already stretched construction workforce.
The companies that recognize this transition early may also discover one of the industry’s most significant emerging opportunities.
AI may be digital. The infrastructure behind it isn’t.
And whether Canada can capitalize on the AI boom may depend increasingly on whether its construction industry can build what AI requires.
