Alberta’s energy industry has always been an economic engine. Now, it’s taking on a new challenge: powering the AI revolution.
As artificial intelligence advances, the demand for high-performance computing is surging. Massive data centres are being built to handle AI workloads, but they require large amounts of electricity — creating pressure on grids, governments, and energy producers to keep up.
Alberta, with its abundant natural gas reserves and growing expertise in carbon capture and storage (CCS), has an opportunity to position itself as a hub for AI-ready infrastructure.
But the province isn’t yet a leader in the data centre race. According to Cushman Wakefield’s 2024 Global Data Centre Market Comparison, Alberta doesn’t rank among the top ten established or emerging data centre markets in the Americas. Energy supply, regulatory hurdles, and international competition all stand in the way.
Can Alberta overcome these obstacles and carve out a role in the future of AI infrastructure?
These questions will be at the forefront of a panel discussion at YYC Data Con, taking place in Calgary this week. The session will feature Wish Bakshi, an AI systems engineer specializing in energy applications, Carson Kearl, a senior energy transition analyst at Enverus; and Mark Taylor, executive VP at Captus Generation.
Ahead of their discussion, Digital Journal spoke with them to get their insights on Alberta’s potential as a data centre hub.
An energy-rich province with room to grow
Alberta has the resources to support large-scale data centres, but its infrastructure is still developing.
“The current data center environment in Alberta is in its infancy and ripe for greenfield technological innovation,” says Bakshi.
The province has natural advantages that make it attractive for AI-driven operations.
Its cold climate reduces the need for energy-intensive cooling, and it has some of the cheapest natural gas in North America. Decades of investment from the oil and gas industry have also created a workforce with expertise in industrial-scale operations.
Yet, Alberta remains far behind major data centre markets.
“Alberta is not competitive with tier one markets — PJM Dom, Dallas, SF, Columbus,” Kearl says.
PJM Dom refers to the PJM Interconnection Dominion Zone, a key part of the PJM Interconnection, which is one of the largest electricity markets in North America. It covers multiple U.S. states and ensures reliable electricity for high-demand industries, including data centres.
The Dominion Zone specifically includes Virginia and parts of North Carolina, home to Northern Virginia’s “Data Center Alley ” — the world’s largest data centre hub. This region benefits from strong power infrastructure, low energy costs, and proximity to key internet exchange points, making it a prime location for AI and cloud computing infrastructure.
However, Kearl sees an opportunity for the province to compete at the next level.
“Where Alberta can be truly competitive, and perhaps even an outlier, is in tier two markets,” he says. “If we can minimize time to power for developers, that will go a long way in drawing their eyes north from Cleveland, Denver, Chicago, and other major U.S. cities.”
Competing with these top-tier markets will require Alberta to reduce regulatory barriers and accelerate infrastructure development.
The ability to quickly bring data centre projects online is crucial. Alberta’s regulatory and permitting processes will play a major role in determining whether the province can attract investment in the years ahead.
Balancing AI’s energy demands with sustainability
AI models require immense computing power, and power requires energy. That’s where Alberta’s resources — especially natural gas and CCS — come into play.
“A challenge across North America (and around the world) is to develop very reliable, sustainable, new power generation to match the demands of the growth in data centre capacity,” says Taylor.
Captus is developing a 4GW natural gas-fired power generation facility with carbon capture in Pincher Creek. The plan is to co-locate data centres directly on the site, eliminating the need for new power lines and reducing land-use conflicts.
“The average temperature over a full year is less than 4 C which greatly reduces cooling demand,” says Taylor. “Access to fuel from two or three discrete sources and the ability to generate power on-site combined with an existing connection to the grid provide very high reliability of power for the data centres.”
While Alberta has the energy potential to support AI infrastructure, integrating large-scale data centres with the power grid is a challenge.
“If the cost-effective location for generating power is remote from the location for sequestering carbon and remote from the location of the data centres, the cost burden has the potential to stop development,” Taylor warns.
Kearl sees policy and investment as key to unlocking Alberta’s potential.
“Unleash capital toward resource production, power generation (gas-fired or nuclear) and power infrastructure,” he says. “American data centre companies who can afford to spend tens of billions in capital will provide their own infrastructure on the data side if we provide fertile ground for development.”
Yet global politics could complicate things.
“U.S. regulatory risk poses the biggest threat to Alberta AI infrastructure,” Kearl adds. “If AI is treated as a national security mandate, which it is right now, American companies will not be training foundational models in Alberta.”
AI’s role in optimizing energy use
AI may be driving energy demand, but it’s also being used to make energy systems more efficient. AI-driven forecasting models help data centres optimize power usage.
“Load forecasting models that leverage AI models allow data centres to optimize energy usage,” says Bakshi. “There are also GPU-specific AI solutions that ensure maximum usage of the hardware while reducing idle time for training models.”
In energy markets, AI is already helping utilities and power producers manage supply and demand.
AI gives operators a clearer picture of what’s happening behind the scenes, says Bakshi. It helps them spot things that are usually tough to track, like differences between expected and actual GPU performance. It also helps industries that use a lot of power manage demand more efficiently, reducing strain on the grid and lowering operating costs.
However, AI’s ability to predict power demand for large-scale computing is still evolving.
“This is a space of research,” Bakshi says. “The challenge with large data centres is that they require a lot of power instantly. And predicting incoming load with AI is often challenging because anticipating consumer behaviour is difficult. However, AI models are already leveraged to capture voltage or grid frequency disturbances which are often leading indicators of incoming volatility.”
Kearl outlines three key factors that will determine Alberta’s success in the AI infrastructure race.
“Close relationships between Canada and the U.S., the province even more-so,” he says. “A path to decarbonization to attract the magnificent seven [Apple, Microsoft, Amazon, Alphabet, Meta, Nvidia, and Tesla] scale of developer; a clear policy road for building data centers, power infrastructure, generation assets, and resource production.”
Alberta has the energy resources, technical expertise, and policy momentum to become a key player in AI infrastructure. But execution will determine whether it can compete on a global stage.
This conversation will continue at YYC Data Con this week, where Bakshi, Kearl, and Taylor will be participating in a panel discussion on Alberta’s digital backbone and the future of AI-driven infrastructure.
For those looking to dive deeper into the future of AI and energy in Alberta, the panel will offer a firsthand look at how industry leaders are thinking about the province’s role in the global AI revolution.
YYC Data Con runs from February 27 to March 1 in Calgary.
This article was created with the assistance of AI. Learn more about our AI ethics policy here.
