Open in interactive viewer → charts, metric popovers & call review

Finning Breaks $4 Billion in Backlog — Just as the AI Power Tape Stops Believing

A record order book, the first data center genset order in Canada, and a group EBIT margin down 50bp: the growth-state bill comes due
FTT.TO · Earnings Call · 2026-11-10

The bill for growth arrives before the revenue does

Finning International posted a third quarter that reads like two different companies. On one page: revenue above $3 billion for a second consecutive quarter, a record $4.1 billion equipment backlog up 42% year-over-year, product support up 7% for a tenth straight quarter, rental up 27%. On the other: EBIT of $243 million that grew just 1%, a group EBIT margin of 8% down 50 basis points, and EPS of $1.20 versus $1.17. Kevin Parkes did not hide it: “For the second quarter in a row, our revenues exceeded $3 billion... our backlog exceeded $4 billion for the first time ever, driven by growth across all 3 market segments.” — Kevin Parkes · 2026-11-10 The interesting question is not whether Finning is growing — it demonstrably is — but whether the market will pay for a dealer that is deliberately converting operating leverage into invested capital ahead of a revenue curve it cannot yet date. Adjusted ROIC of 18.8% sits at the bottom of the 18–25% target range; invested capital turns of 2.36x just clear the threshold. Parkes was unusually blunt about why: “The good thing is we are not in a mature state right now. We are in a growth state.” — Kevin Parkes · 2026-11-10 That is either a patient investor's dream or a warning that the next four quarters look like this one.

The promise kept — and the tape that stopped believing

Last May, asked about Alberta data centers, Parkes said the balance was weighted to backup power and that “we'd like to see some things hit our backlog here in the second half of the year.” — Kevin Parkes, President and CEO · 2026-05-13 This quarter he delivered: a major data center order landed in the Canadian backlog, Caterpillar gensets packaged at the repurposed Red Deer facility, with a product support contract attached. That is a promise kept in three quarters — rare enough to be worth saying out loud. The delivery mechanism is the genuinely new idea in this transcript. Finning has stopped talking about data center power as a project pipeline and started talking about it as a bridging opportunity — modular, standardised, transferable capacity rented to hyperscalers while their permanent generation gets built. “It's not rental really, it's more a medium-term bridging solution to a full-time power generation solution... we don't think customers give up power capacity very easily once they have it.” — Kevin Parkes · 2026-11-10 The Q&A produced the cleanest segmentation yet: backup is a standard template Finning has run in Dublin for twenty years; prime power is bespoke and project-by-project; bridging is the productised middle they are trying to build once and sell many times. Here is the tension. Finning is leaning into data center opportunity precisely as the global tape has turned against the complex that feeds it. Over the last 30 days the market's data center theme sits firmly among the decliners — AI data centers and adjacent infrastructure names (power, cooling, high-bandwidth memory, optical interconnect) are broadly negative, and the 90-day picture is worse still. Finning's exposure is not AI-capex beta; it is Caterpillar reciprocating engines with a 20-year product support annuity, and management was careful to fence it off:

The growth in AI and data center power demand provides meaningful additional upside, but it is not a dependency for future earnings growth.

Kevin Parkes · 2026-11-10
That is the right sentence to say when the tape is cracking. It is also, read cynically, the sentence you say when the order you just booked is a standby template rather than the prime-power crown jewel that speed to market rhetoric has been pointing at for three quarters.

Where the margin went

Product support margin is now the most-asked question on this call for a second consecutive quarter — Product support margins were a top-three theme in Q2 and remain a live irritant. In August, Devin Dodge pressed on the split between mix and transient costs; Kevin Parkes gave three causes. This quarter Ian Gillies asked essentially the same question and got a sharper, quantified answer: roughly two-thirds structural, one-third transient. The structural part is uncomfortable, because it is not really a problem to be fixed. Growth is coming disproportionately from large mining customers whose value propositions carry thinner margin, plus tariffs on heavy steel that bite hardest in mining, plus parts-and-services work Finning previously declined as insufficiently profitable. The transient third is technician ramp — the cost of hiring a technician a day in Canada and waiting for them to be productive. Canada's segment EBIT margin fell 90 basis points to 7.8% while new equipment sales rose 55%. That is what record backlogs look like on the way in. What mattered more was the read-through to capital. ROIC range discipline has been the company's refrain since its 2023 investor day, and the guideposts held: invested capital turns rose to 2.36x, free cash flow was a positive $70 million, and net debt to adjusted EBITDA sat at 1.6x with SG&A at 14.5% of last-twelve-month revenue, down 50 basis points. Devin Dodge asked what it takes to reach the midpoint of the range; the honest answer was better EBIT margins and better capital velocity, neither of which is available while you are pre-building for a growth curve.

Chile exhales, Argentina inhales, and the U.K. quietly compounds

The regional dispersion is the most underrated part of this call. South America printed a 10% EBIT margin, up 30 basis points, on new equipment sales down 35% in functional currency — a fleet recalibration in Chile (retired trucks, weather, labor disruption, deferred awards) that Parkes insisted was timing, not demand, backed by order intake more than four times the prior-year quarter. Chile's tax reform — a phased cut from 27% to 23% by 2029 — costs $3–6 million of noncash deferred-tax remeasurement in Q4 but buys a structurally better investment case. Argentina remains the option: about a tenth of South American revenue, with the first new truck at the Alumbrera mine in Argentina being assembled and contributing nothing to product support yet. Meanwhile the U.K. and Ireland, the market everyone has written off, delivered 6.3% EBIT margin and a 22.6% ROIC up 240 basis points on lower invested capital — and its data center playbook is the template being exported. The announced acquisition of JF Hunt Power is small, but Parkes framed it as a platform, and paired it with adjacent repair services as the two M&A lanes now explicitly in the plan: "M&A is coming into our horizons."

The read

New truck population, rebuild intensity, and technician headcount remain the three levers that actually move Finning's earnings, and all three are pointed the right way — rebuilds in Canada already double all of 2025, remanufacturing heading from 6–8 engines a month toward 14, a night shift coming at Fort McKay. The data center story is the shiny object; the mining population story is the business. What changed this quarter is that Finning converted a talking point into a backlog line and, in doing so, took on the margin and capital drag that always precedes the annuity. The stock's problem, whenever it next prints a tape, will not be demand. It will be proving that an 8% EBIT margin in a growth state is a temporary condition rather than the new normal — and that the first standby genset order in Red Deer is a template, not a trophy.