Aecon's Record Backlog Is Now a Nuclear Fabrication Story

Pickering flipped the narrative from legacy wounds to 400,000 square feet of nuclear-qualified shop floor — but the CAD 645 million of free cash flow is mostly a one-time working-capital event, not a new run-rate.

ARE.TO · Earnings Call · 2026-10-30

What actually changed

For two years the Aecon call was a ritual of penance: Western Civil, legacy projects, substantial completion, pain share. Those phrases have fallen out of the company's own keyword stack. The freshest, highest-conviction language in the Q3 transcript is fabrication capacity, sitting alongside nuclear builder and reported backlog. That is not a marketing tweak. It is a different company describing itself. The numbers back it. Revenue of CAD 1.9 billion rose 26%, of which 71% was organic. Adjusted EBITDA of CAD 130 million rose 40%, and construction segment margin expanded to 6.8% from 5.8%. Backlog hit CAD 14.6 billion versus CAD 10.8 billion a year earlier, on CAD 6 billion of awards booked in the quarter. Crucially, the legacy drag has collapsed: negative gross profit from those jobs was CAD 9 million in the quarter, against the roughly CAD 94 million the same wound cost across 2025. The final legacy project reached substantial completion in June. Jean-Louis Servranckx was not shy about the framing:

When I read, Aecon, Canada nuclear builder, I tend to say, Canada, the nuclear builder is Aecon.
Jean-Louis Servranckx, President and CEO · 2026-10-30

Capacity as moat, not capex

The single most interesting disclosure was buried in the Q&A, not the prepared remarks. Aecon agreed to buy additional nuclear fabrication capacity in Ontario for roughly CAD 50 million, a footprint about 40% larger than its existing Cambridge facility, with adjacent acreage for expansion. Pro forma, the company goes from about 300,000 square feet of fabrication to just over 400,000, across four nuclear facilities in Canada and the United States. CFO Jerome Julier drew the line precisely where it matters: “The big differentiation is the ability to do it under nuclear quality programs, right? That is the barrier to entry. It's not easy. It requires systems, processes, experience, resume.” — Jerome Julier, Executive Vice President and CFO · 2026-10-30 The strategic logic is that modularisation is the release valve on a labour-constrained market — “You don't have the same people renovating tracks or building industrial facilities or working in the nuclear or laying part of electrical grids” — Jean-Louis Servranckx, President and CEO · 2026-10-30 — and Servranckx flagged the underappreciated submarket: extended power uprates and life extensions, worth something like CAD 1 billion per reactor for the utilities, with Aecon already carrying more than 1,500 nuclear workers in the United States.

The bridge management wouldn't give in July

The sharpest contrast with the prior call is disclosure posture. In July, Julier closed the door: “With regards to 2027, we are not providing an outlook on that 1, so we are not going to front run it.” — Jerome Julier, Executive Vice President and CFO · 2026-07-31 Three months later he built the bridge in public — roughly CAD 5 billion of backlog executing in the next twelve months, plus CAD 1.1 billion of recurring revenue, implying about CAD 6.1 billion, with a CAD 750 million-to-CAD 1.2 billion band of book-and-burn work on top that could “crest the CAD 7 billion mark” — Jerome Julier, Executive Vice President and CFO · 2026-10-30. That is not guidance, but it is a materially more confident conversation with the street. Servranckx's macro line escalated too: “Probably in my life, I have not seen such a dynamic situation for infrastructure construction in the five years to come.” — Jean-Louis Servranckx, President and CEO · 2026-10-30 Note what he did not do: promise margins back to the old lump-sum peaks. The contract mix now runs roughly 64% variable price and 81% non-fixed-price or collaboratively developed, and the recurring utility book (CAD 987 million, up 42% on nearly three-year MSAs with close to 100% renewal) is a deliberately lower-beta earnings stream. The margin improvement is real but capped by design.

The cash line is a level shift, not a trend

The headline that will travel furthest is CAD 645 million of trailing-twelve-month free cash flow versus negative CAD 41 million a year ago. It deserves a caveat. Alistair MacCallum was explicit on seasonality: “Q4 will be slightly positive, but it will not be what we saw in Q3.” — Alistair MacCallum, Senior Vice President, Finance · 2026-10-30 And on next year: “there's still a few opportunities to improve working capital next year, but it won't be at the size of this year.” — Alistair MacCallum, Senior Vice President, Finance · 2026-10-30 Q3 was unbilled reduction, claim resolution and milestone billing landing at once. Management's framing — operate at "this new level" — is the story to test, not the quarter's number. One more quiet shift: net debt of CAD 338 million includes a CAD 320 million repayment agreement for the preferred shares of Aecon Utilities, closing in Q4. In March, when asked about eventually listing that business, Julier demurred: “Our focus with the business is to continue to grow and build out the recurring revenue programs and expand its diversification... First things first.” — Jerome Julier, Executive Vice President and CFO · 2026-03-06 Six months later they are buying the minority in instead. That is a very different use of a very strong balance sheet, and it happens just as Maple Fibre — a majority-controlled fibre network alliance with a Tier-1 telecom — begins construction.

The tape is voting the other way

Here is the tension worth carrying forward. Over the last 90 days, the market's own theme price movers have de-rated precisely the end-markets Aecon just sold: small modular reactor equities are down about 5%, low enriched uranium down nearly 5%, and the AI-datacentre complex — AI data center down 5.3% with 49 decliners against 7 advancers, data center capacity down 8% — has been the weakest cluster in the tape. That is not necessarily a contradiction: Aecon sells refurbishment and transmission labour under cost-plus and collaborative structures, not merchant reactor equity or hyperscale construction risk. But it does mean the backlog was largely anticipated, and the tape is now asking whether power-demand headlines convert into cash before they convert into multiple. For the first time in years, Aecon's problem is no longer the past. It is proving that a CAD 14.6 billion book at a deliberately thinner margin can compound.