Canada Should Prioritize Growth and Position Itself as an AI Infrastructure Leader
Diverse energy mix, renewable options, vast untapped land, cold climates, slowing economy
Good morning to the Great White North.
Full disclosure: I pitched this to the Globe and Mail, but it wasn’t the right fit for the publication this time. However, I do hope to have a chance to write for the national paper one day.
A few thoughts/ points to share: 1) I collaborated with a Vancouver-based policy analyst who cannot put his name on this 2) I also included some insights I got from a uni classmate who works in the energy sector in Alberta 3) I do feel strongly about this topic and writing this piece made me miss home dearly.
P.S. I used Canadian spelling for this piece ;p
As the AI boom increases demand for computing power, hyperscalers in the U.S. and China compete for land and energy to expand their data centres. Many hyperscalers are looking for alternative options and are considering Canada for its energy, land, and resources.
Amid an economic recession, Canada can consider positioning itself as a global leader within the infrastructure layer of the growing AI industry, which can then potentially attract and nurture innovation in the model and application layers as well.
Canada has considerable advantages in positioning itself as a critical player in the AI infrastructure conversation, including vast unurbanized land, diverse energy sources, a favorable cold climate, sophisticated pipeline infrastructure, a strong talent pool, and advanced data security. As we all navigate this pivotal moment, Canada should choose to bring jobs to the nation, optimize the use of its abundant renewable resources, and establish itself as a major AI player.
Alberta is Already Open for Business
Hyperscalers like Google, Amazon, and Microsoft are forming agreements with the Alberta government to establish data centres in this resource-rich province. As of October, 239 data centres were operating in Canada, 22 of which were located in Alberta (12 in Calgary and nine around Edmonton).
According to the Alberta Electric System Operator’s (AESO) forecast for its reserve margin, Alberta’s planning reserve margin is quite healthy. The higher the reserve margin, the less risk of blackouts, so utility companies are constantly finding ways to balance a healthy reserve.
In the U.S., the targeted healthy reserve margin typically is within 17-20%, so looking at this chart from AESO, it is likely that either the government will nudge companies to a) scale back on available resources or b) add demand (potentially AI-driven).
Capital Power’s management said in a recent earnings call that “... the focus for hyperscalers has been on the U.S. market, but as reliability and speed to market have increasingly become more important in the calculus for where to locate these hyper data centres, Alberta has become more prominent in that conversation.”
Taking a quick glance at the ongoing projects related to the AI boom, we are seeing new projects, capital, and job prospects across the whole energy supply chain:
Power generators are exploring opportunities for data centres in Alberta. TransAlta Corp. announced discussions with major proponents seeking locations for these facilities.
Oil and gas companies such as Edmonton-based Capital Power have submitted applications for two projects totaling 1,500 MW of load for new data centres at the Genesee Generating Station.
Infrastructure providers such as Enbridge are focusing on optimizing their operations in anticipation of a surge in demand for renewable energy from data centres and generative AI.
Data centres such as eStruxture, Canada’s leading AI and hyper scale-ready data centre platform, just announced that it is building Alberta’s largest data centre with an investment of CAD 750 million due to the recent surge in demand for computing power and Beacon Data Centres, a private development firm has proposed three large AI data centre hubs in Alberta.
Optimizing The Diverse Energy Mix
The Alberta Electric System Operator (AESO), which operates the province’s power grid, is working with at least six proposed data centre proposals, according to the latest public data. The regulator, Canadian Energy Centre, is A-OK for more business.
Lawyers with Calgary-based McMillan LLP wrote in July that between 2018 and 2023, data centres in Alberta generated approximately $1.3 billion in revenue, growing on average by about eight percent per year. Opportunities like this will continue to grow as hyperscalers seek locations to build data centres outside the U.S. due to the power shortage and land restrictions.
Skyline of Calgary, home to many of Canada’s largest power companies.
U.S. big tech has been leading the search for suitable places to ramp up access to land, power, and, eventually, data centres that can be powered by renewables. Notably, Amazon has been laying out the groundwork and already secured renewable power from Alberta, building out a near 500-megawatt capacity wind far and backing one of the largest solar photovoltaic projects in the world in the province, over the last few years.
This is echoed by Alberta Technology Minister Nate Glubish, saying that proposals for new data centre projects may well come with new clean energy to fuel them - a "bring your own power" approach. Looking at Alberta alone, natural gas supplies about 60 percent of the province’s energy needs but is followed by wind and solar at 27 percent, thus these projects could be powered by a mix of energy sources, including sustainable options.
Despite Environmental Concerns, Can Canada Prioritize Growth?
In neighbouring British Columbia, the situation is a bit more complex. Although the province offers some of Canada’s lowest energy rates and generates over 98% of its power from renewable sources, and it already serves as home to 35 data centres (third highest in the country), lawmakers in BC appear much more wary to position the province as a welcome home for the incoming AI data centre boom.
In 2022, the BC provincial government paused new connections to its electricity grid for cryptocurrency mining operations, which former Energy Minister Josie Osborne described as consuming “massive amounts of electricity” while adding “very few jobs.” While crypto data centres currently require more concentrated energy than their AI counterparts, the logic behind this has been used to justify turning away data center projects for AI.
In a context where the province’s Clean Energy Strategy has pledged to achieve a 100% clean integrated grid by 2030, and yet recent large energy projects are not forecasted to meet the clean energy needs of a growing population. The concern for bringing in data centres is that it will just add more pressure to the energy mix. But as much as achieving 100% clean energy is very noble, so is finding a way to balance growth and sustainability amid an economic slowdown.
AI offers a whole ecosystem of new employment opportunities compared to the crypto mining industry if the infrastructure is accessible. Singapore has opened up work visas to talent investing in and building AI, signed deals with Malaysia to offer data centres, and brought in talent from across the world – notably China and the U.S.- to build up the entire full-stack ecosystem.
We’re seeing in Canada now that lawmakers are finding themselves at a crossroads. The global demand for artificial intelligence is undeniable, with an insatiable appetite for computing power and data centres that show no signs of abating. U.S. hyperscalers are grappling with a power shortage, where new facilities take too long to come online to meet immediate growth needs. Policymakers must see the bigger picture here. Yes, sustainability is vital, but if we miss out on this opportunity, then Canada could be left behind in this AI arms race.
Also, the CBC just published this report on Canada’s nuclear power potential a day after I published my piece here. It focuses specifically on SMRs (which I wrote about here) and why we cannot yet use fully renewable energy to power data centres.