For most investors, the AI boom still brings names like Nvidia, Microsoft or Celestica to mind. But lately I’ve been spending more time thinking about something much less exciting on the surface: electricity. AI data centres consume an enormous amount of power, and that power has to come from somewhere. Alberta alone has roughly 19,565 megawatts of proposed new data-centre load seeking grid connections.
For perspective, the most electricity Alberta has ever consumed in a single hour was around 12,800 MW. Not all of those data-centre requests will ever get built, of course, but the gap tells you how enormous the opportunity—and the problem—could become. We’re already seeing it happen. Meta is building its first Canadian data centre in Sturgeon County, Alberta, with an investment of more than $13 billion and planned capacity of roughly 1 gigawatt. That got me thinking: instead of trying to guess which AI model wins, why not look at the Canadian companies supplying the power, transformers, natural gas infrastructure and construction needed to keep all those servers running?
Here are my top five.
Ready to start investing? Click the Qtrade banner below to learn more and open a Qtrade Direct Investing account.
#5 Pembina Pipeline (TSX: PPL)
Approximate share price: $65.82
Market capitalization: ~$38 billion
Industry: Energy infrastructure / midstream
Why It Made My List
Pembina isn’t suddenly becoming an AI company, and that’s actually part of what I like about it. The company owns 47.5% of the Greenlight Electricity Centre in Sturgeon County, a planned 932 MW natural-gas-fired power facility designed to provide dedicated electricity to Meta’s data centre. Pembina expects its net investment to be about $2.1 billion, with the project producing roughly $310 million of annual run-rate adjusted EBITDA attributable to Pembina once operating. The current target is the second half of 2030. That’s meaningful growth attached to a business that already existed long before anyone started talking about AI.
What I Like
Pembina generated $1.064 billion of adjusted EBITDA and $778 million of adjusted operating cash flow in the second quarter alone. What interests me most is that Greenlight doesn’t need to rescue the company. Pembina already owns pipelines, processing facilities and other infrastructure generating cash today. Data centres potentially create another source of long-term demand for both electricity and Western Canadian natural gas. I like growth more when the entire investment thesis doesn’t depend on it working perfectly.
One Thing I’d Watch
Greenlight isn’t expected online until 2030. That’s a long runway with billions of dollars still needing to be spent.
Would I Buy Today?
I’d happily own Pembina long term, particularly for income and infrastructure exposure. I just wouldn’t buy it expecting an explosive AI-stock-type return. The data-centre opportunity is an added growth engine, not the entire business.
#4 Aecon Group (TSX: ARE)
Approximate share price: $57.76
Market capitalization: ~$4.0 billion
Industry: Construction and infrastructure
Why It Made My List
Someone actually has to build all this stuff. An Aecon-led consortium was awarded the engineering and construction contract for Greenlight, with Aecon’s share worth approximately $1.7 billion. That amount is being added to the company’s third-quarter backlog. That’s not theoretical AI exposure. It’s contracted infrastructure work tied directly to the power required by a major data centre.
What I Like
Aecon was already showing improving results before this award. Second-quarter revenue increased 25% to $1.63 billion, while adjusted EBITDA roughly doubled to $82.4 million. Backlog stood at about $10.5 billion at the end of June. Then Greenlight came along. A large backlog gives Aecon something I really value in a construction business: visibility. It doesn’t guarantee profits, but management has a much clearer view of where future revenue should come from.
One Thing I’d Watch
Construction investors should never confuse backlog with guaranteed profit. Large projects can run into labour shortages, delays or unexpected costs, and a bad contract can destroy margins surprisingly quickly. Execution matters enormously here.
Would I Buy Today?
I like Aecon more than I did a year ago, but the stock has also run hard. Around $58, I’d probably wait for a better entry rather than chase it after such a strong move.
#3 Capital Power (TSX: CPX)
Approximate share price: $62.25
Market capitalization: ~$9.7 billion
Industry: Power generation
Why It Made My List
Capital Power might have one of the cleanest data-centre catalysts on this list. The company signed an agreement lasting more than 10 years to supply 250 MW of electricity to Meta’s Sturgeon County data centre, with the load expected to begin in the second half of 2028. That distinction matters to me. We’re not talking about management hoping AI demand eventually appears. There is an actual customer and an actual long-term agreement.
What I Like
Capital Power generated $351 million of adjusted EBITDA and $328 million of adjusted funds from operations in Q2. It also increased its dividend for the 13th consecutive year. That’s an interesting combination: an established power producer with a dividend-growth history now gaining exposure to one of North America’s fastest-growing sources of electricity demand.
One Thing I’d Watch
Power generation is extremely capital intensive, and Capital Power still needs to manage acquisitions, development spending and its balance sheet carefully. I’d also remember that the Meta agreement doesn’t begin tomorrow. The expected start is 2028.
Would I Buy Today?
Of the larger companies here, Capital Power is probably one I’d be most comfortable slowly accumulating rather than trying to perfectly time. I like the combination of existing cash flow, dividends and data-centre growth.
#2 Hammond Power Solutions (TSX: HPS.A)
Approximate share price: $307
Market capitalization: ~$3.5 billion
Industry: Electrical equipment / transformers
Why It Made My List
You can build all the power plants you want, but electricity still has to be transformed, distributed and controlled before it reaches rows of servers. That’s where Hammond gets interesting. The company makes dry-type transformers and power-quality equipment. In Q2, sales reached a record $325 million, up almost 45% from the previous year, while backlog was nearly double where it stood a year earlier. Management specifically pointed to U.S. data-centre investment, electrification and power infrastructure spending as major demand drivers.
What I Like
This is probably my favourite business model on the list. Hammond doesn’t need to know whether Meta, Amazon, Microsoft or another hyperscaler ultimately builds the most data centres. It sells equipment the entire industry needs. The recent $365 million acquisition of AEG Power Solutions also expands Hammond beyond transformers into a broader range of mission-critical power electronics.
One Thing I’d Watch
The price. Hammond has become a very popular stock, and expectations are now much higher than they were a couple of years ago. The AEG acquisition also needs to be integrated successfully. A great company can still be a disappointing investment if you pay too much for it.
Would I Buy Today?
I love the business more than I love the current valuation. At today’s price, I’d probably wait for a pullback or build a position very slowly.
#1 Enerflex (TSX: EFX)
Approximate share price: $35.71
Market capitalization: ~$4.4 billion
Industry: Energy infrastructure and engineered systems
Why It Made My List
Enerflex jumped to the top of this list because of what happened on October 1. The company announced a contract to design, engineer, fabricate and assemble approximately 450 MW of behind-the-meter natural-gas power generation for a North American data-centre developer. Deliveries are scheduled for 2027 and 2028. Even more interesting, management previously said its pipeline of distributed-power opportunities had grown beyond seven gigawatts across data centres and other power-generation applications. That’s what caught my attention. The 450 MW order may not be a one-off.
What I Like
Enerflex was already improving financially before this contract arrived. At the end of Q2, Engineered Systems backlog was approximately US$1.5 billion. The company generated US$32 million of free cash flow, while net debt had fallen to US$455 million and its bank-adjusted net-debt-to-EBITDA ratio dropped to about 0.8 times from 1.3 times a year earlier. That improving balance sheet matters because Enerflex now has an opportunity to invest into a potentially much larger power-generation business without starting from a weak financial position.
One Thing I’d Watch
The market has already noticed. Enerflex jumped almost 12% on October 1 after the announcement and has moved significantly higher this year. I’d also want to see additional data-centre orders before assuming that the seven-gigawatt opportunity pipeline automatically becomes revenue.
Would I Buy Today?
Enerflex is probably the stock I’d be watching most closely from this list, but I wouldn’t chase it immediately after a major announcement. If I were buying today, I’d probably wait for things to settle down and look for a more attractive entry. The opportunity is exciting. That doesn’t mean I need to pay any price for it.
Final Thoughts
What makes this theme interesting to me is that AI is becoming much more than a technology story. Data centres need power plants. Power plants need natural gas. Electricity needs transformers. Facilities need construction. And somebody has to finance and operate all of it. Over the next year, I’ll be watching for additional hyperscale data-centre announcements in Canada, new behind-the-meter generation projects, transformer demand, project financing and—most importantly—whether proposed projects actually reach final investment decisions.
This list probably makes the most sense for investors who like industrial, utility and energy businesses but still want exposure to AI growth without owning another expensive semiconductor or software stock. There will undoubtedly be projects announced that never get built, and some investors will probably get carried away with the theme. That’s why I want real contracts, real cash flow and businesses that can survive even if the AI buildout develops more slowly than expected. Because despite everything we hear about artificial intelligence being digital, the biggest bottleneck may end up being surprisingly physical: AI can’t run without electricity.
Ready to start investing? Click the Qtrade banner below to learn more and open a Qtrade Direct Investing account.