Canada’s 6 Construction and Engineering Stocks, Ranked: Is the Market Wrong About AI?

At the time of writing, I own WSP Global. This article was created using a transcript from the video below, with the assistance of AI.

This is probably the most polarizing tier list I’ve done. Three of these stocks are in the gutter. Three are through the roof. And artificial intelligence is the reason for both.

I ranked the six largest construction and engineering stocks in Canada from S to E: WSP Global, Stantec, AtkinsRéalis, Badger Infrastructure, Bird Construction and Aecon Group. They split neatly into three camps. The engineers sell expert hours and design work. The builders pour the concrete. And Badger is the outsider, digging up buried lines with hydrovac trucks.

Here is the split in one chart. Over the last year the builders have rocketed on nuclear refurbishments and AI data-centre power demand, while the engineers have been sold off on fears that AI will eat their billable hours.

Disclosure up front: I own WSP and have been buying it aggressively. I owned Aecon in the past and sold it. More on both below.

Badger Infrastructure (TSX: BDGI): C Tier

Before anyone digs near a buried gas line, power cable or fibre line, someone has to expose it safely. That is Badger. Its hydrovac trucks blast the soil loose with pressurized water and vacuum it away, and with 1,822 trucks it is the largest operator in North America. About 90% of revenue came from the US when it last disclosed the split.

The edge is the factory. Badger builds its own trucks in Red Deer, Alberta, at roughly US$411,000 each, and no competitor of similar size does that. Each truck brings in around $500,000 a year and pays for itself in four to five years on an EBIT basis, which is why Badger runs some of the best margins on this board. Revenue per truck rose 14% last quarter, management says it is raising prices in every market, net debt is a modest 1.5 times EBITDA, and the dividend (0.8%) has grown about 4% a year.

The problems are cyclicality and price. Construction is 45% of revenue, and when activity froze in 2021 Badger lost money. A 25% US tariff applies to non-US parts in trucks shipped south from Red Deer. And at 25 times forward earnings, Badger is in roughly the 90th percentile of its own decade: it has only been cheaper than today about 10% of the time.

A great company at what is probably a D-tier price. That levels it out in C.

Aecon Group (TSX: ARE): E Tier

Aecon is one of Canada’s largest builders: highways, transit, utilities, and every Ontario nuclear refurbishment. The stock is up more than 100% in a year. Nuclear is roughly 30% of the business, with a seat on Canada’s first small modular reactor at Darlington, and in September Aecon’s teams won about C$3 billion of Pickering refurbishment contracts: a C$1.7 billion 50/50 joint venture with AtkinsRéalis plus a C$1.3 billion turbine contract alongside Siemens Energy. It also leads the consortium building Greenlight, a 932 MW gas-fired plant in Alberta that will power a data centre. The backlog sits at C$10.5 billion.

So why last? Margins. Aecon’s adjusted EBITDA margin went from 1.9% in 2024 to 4.3% in 2025 and about 5.1% in 2026. On razor-thin margins, one bad project wipes out a year, and I know that firsthand: I owned Aecon and sold it after Coastal GasLink turned into years of cost overruns and court fights. Today 75% of the backlog is not fixed-price, up from about 30% in 2021, which helps. But roughly 85% of Greenlight’s project costs sit under fixed-price agreements, net debt is 2.2 times EBITDA, and much of the cash is tied up in joint operations.

Aecon earned $1.03 a share in 2015 and $0.23 in 2025. At 30 times forward earnings, a slowdown or one more overrun could crush that multiple fast. An outstanding company, but one you have to time, buying the bottom of the cycle and selling the top. And Bird does the same job better.

AtkinsRéalis (TSX: ATRL): A Tier

You might remember this company as SNC-Lavalin, one of the most notable corporate collapses in Canadian history after the bribery scandal. It rebranded as AtkinsRéalis and is doing quite well.

The engineers are a different business from the builders. They don’t sell products; they sell people’s hours, which is why their margins are far higher. AtkinsRéalis has something no one else has: through Candu Energy it holds the exclusive licence to the CANDU reactor, the design used by every operating nuclear plant in Canada. Nuclear is about a fifth of revenue but nearly a third of profit, and nuclear revenue is expected to grow from about C$1 billion in 2023 to C$2.7 billion in 2026. The Darlington refurbishment finished early and under budget, management has raised its nuclear revenue target three times in two years, and the balance sheet is roughly net cash with a BBB rating.

The bear case is the consulting side: about 70% of profit, growing just 1-2% organically, with the growth target cut to 5-7% and total backlog slipping from C$21.2 billion to C$20.2 billion. At 19 times forward earnings you are betting on nuclear. I’m fairly bullish on nuclear power, and that exposure is why AtkinsRéalis is down only about 18% from its highs while WSP and Stantec are down 40% or more. A tier.

Bird Construction (TSX: BDT): D Tier

Bird is up about 725% in five years. It builds hospitals, schools, industrial plants, military facilities and data centres, and it is simply the better-run builder: higher margins (7.1% vs about 5%), a much stronger balance sheet (0.96 times EBITDA vs 2.2), a cheaper multiple (26x vs 30x) and more than 80% of its backlog on client-shared cost risk rather than fixed-price bids. Pre-pandemic it was barely profitable at a 0.8% margin.

The growth story is real. Bell picked Bird as its preferred builder for its AI Fabric data-centre build-out, with warrants for up to 2.6 million Bird shares. Defence backlog tops C$1.5 billion. Contracted backlog of C$6.1 billion plus about C$6 billion of pending work covers roughly 3.3 years of revenue, and if Bird hits 8% margins on its 2027 revenue target, EBITDA would be 51-67% above the last twelve months.

But the re-rating already happened. Bird trades near 19 times EV/EBITDA versus 5.5-8x as recently as 2023, revenue was flat in 2025, and there are no buybacks and a frozen dividend. I thought it was too expensive at $30-35 a share and it has only gone up since, so take that with a grain of salt. Still, any delay, any overrun or any slowdown in data-centre builds and these stocks have gigantic downside. D tier, not because it is a bad company, but because the price assumes everything goes right.

Stantec (TSX: STN): B Tier

Stantec’s 34,000 people design water plants, infrastructure, buildings and data centres, with about half of revenue from the US. Many of its fees are a percentage of construction cost rather than hours worked, which is less exposed to AI than pure hourly billing.

The numbers are strong. Backlog is at a record, organic backlog grew 7%, adjusted EBITDA margins hit a record 18.7%, net debt is just 1.3 times EBITDA, and Stantec raised its buyback limit from 2% to 5% of shares. Its CFO said it is “increasingly hard to ignore” buying back stock at these valuations. At about 15 times expected earnings, Stantec sits in the 3rd percentile of its own decade.

The bear case: US organic growth has stalled at 0%, Stantec is likely to miss its 7%+ organic growth plan, there is a new CEO, and free cash flow was negative in the first half. A premier company that I was very close to ranking higher. B tier, only because I like AtkinsRéalis’ nuclear exposure a bit more.

WSP Global (TSX: WSP): S Tier

WSP is the largest engineering firm in the United States after buying TRC in February, with 84,000 engineers, scientists and planners. Revenue is 58% private and 42% public, so it holds up across the cycle, and it has been skating to where the puck is going: the power grid. After POWER Engineers, TRC and Ricardo, power and energy make up about 30% of US revenue, and the CEO says roughly a third of WSP benefits from AI spending growing at double digits. Backlog is a record C$20.1 billion, adjusted EBITDA margins have climbed to 19.1% on a path to 20%, and first-half adjusted earnings rose 24% while the stock fell about 44% from its highs. It trades around 15 times earnings, in the 5th percentile of its decade.

The balance sheet is the debate. Net debt is nearly C$7 billion, about 2.3 times adjusted EBITDA as WSP reports it, above its 1-2x target. But that is the pattern: every big acquisition spikes leverage, and every time it comes back down.

WSP also sold nearly C$1 billion of shares at $232 in December 2025, about 30% above today’s price. The market clearly wants debt paid down before the next deal; the stock fell when the Arcadis bid surfaced and rose when WSP walked away.

Here is why I think the market has the AI story backwards. More than 60% of WSP’s work is fixed-price. One estimate puts each 1% of labour saved at about +0.2 points of margin. If AI trims the mundane work, WSP can share some savings with clients, see revenue dip and still earn more profit. And WSP is short on people, not work, with about 700 open technical roles. If AI pushes more backlog through the same headcount, you get revenue growth and margin expansion at once.

The risk is real: organic growth is about 3.5%, the Americas backlog is soft, and the CEO concedes AI hasn’t lifted productivity yet. If clients push back on price at rebid, it would be a slow bleed. But at a 50%+ discount to its historical valuation, with earnings still growing, I think the risk/reward is excellent. WSP has been my heaviest buy over the last few months, under $200 a share. S tier.

The Final Board

S: WSP. A: AtkinsRéalis. B: Stantec. C: Badger. D: Bird. E: Aecon.

If I had made this list two years ago, it would have been completely reversed: engineers at the bottom, builders on top. The market is pricing anything that could be disrupted by AI for the worst case, and I think it is missing the forest for the trees. It sees billable hours falling; it isn’t pricing in margins rising or AI solving the staffing shortage. This is also the first board I have made where I view all six companies as solid businesses. The ranking is about price.

If you want the written version of every tier board from this series, with the reasoning and the returns since each call, you can grab the complete tier list collection for free in the video description. And if you want the stock reports and tools behind the numbers in this article, that is Stocktrades Premium.

None of this is investment advice. I own WSP Global and previously owned Aecon. Do your own research before buying any stock.