ADF Group: The easy part of the rally is done — now look at the cash

ADF Group's second-quarter report read like a stumble on paper and a clean step underneath, and the market chose the paper. The Quebec steel fabricator posted net income of C$3.0 million, or C$0.10 a share, for the three months ended July 31 — a long way short of the roughly C$0.32 the street expected — and the shares fell about 7% on the news. Yet the same release carried revenue of C$95 million, up 79% from a year earlier and ahead of the consensus estimate. When a company misses badly on profit but beats on sales, the first question is what actually happened inside the quarter. Here, the answer is more accounting than it is steel.

The "miss" was mostly an accounting echo

The shortfall did not come from the business losing money on its work. Roughly C$4.3 million of the quarter, or about C$0.15 a share, was a non-cash mark-to-market charge on ADF's deferred and restricted share units — stock-based compensation that gets expensed when the share price rises. ADF's stock was up around 77% for the year heading into the report, and that climb is what created the expense. The chief financial officer put it plainly: the company was a victim of its own success. Another C$1.8 million, or C$0.06 a share, went to foreign-exchange losses on the U.S. dollar. Strip those two line items out and the underlying operation looked stronger, not weaker: adjusted EBITDA of C$8.4 million roughly doubled from C$3.7 million a year earlier.

This is the trap in reading a headline EPS number for a company whose stock has already re-rated. The same arithmetic that made shareholders happier on the way up shows up on the income statement as a cost. It is real accounting, but it is not a deterioration in the ability to fabricate steel and get paid for it.

The proof is the backlog and the cash, not the print

The reasons to be interested in ADF were never in the quarterly EPS line anyway; they sit in forward bookings and cash generation. At the end of July the order backlog stood at a record C$693.7 million — roughly two years' worth of revenue at current run rates — with about C$243 million of it coming from Groupe LAR, a structural-steel maker ADF bought in September 2025. A record backlog is the concrete bridge the market keeps underestimating: it does not rely on hope for next year, it relies on work already signed.

The cash side supports the same read. In the first half ADF generated C$47.1 million of operating cash flow, and its cash balance climbed C$28.7 million since the start of the fiscal year to C$91.4 million against basically no net debt. One honest caveat: that cash tally includes a C$25 million customer-claim settlement, so the underlying run-rate is below the headline number. But even discounted for that, a business throwing off cash while holding a record backlog is generating the free-cash-flow proof this kind of setup depends on. The market is still pricing the single-quarter accounting miss while the operating setup keeps getting cleaner.

The honest caveat: this is no longer beaten down

Here is what I will not pretend. ADF is not a bottomed-out value stock waiting for a lifeline — it is up roughly 77% year to date and trades near 14 times trailing earnings. The market has already paid up for the backlog growth. The whole argument from here rests on one variable: margin.

Second-quarter gross margin slipped to 18.7% of revenue from 20.7% a year earlier, squeezed by higher steel prices, tariff-related costs on Canadian fabrication shipped into the U.S., and the still-low-margin legacy Groupe LAR backlog booked with the acquisition. That is the strongest bear point in the report, and it is a real one. Management's retort is that the first-half gross margin of 21.5% is a better guide to the normalized business and that margins should climb as the low-margin legacy LAR work rolls off. That is a testable claim, not a promise: the LAR backlog converts and fades, and the margin line either holds near 21% or it does not.

So the break condition is specific. If the converting backlog holds margins near the 21% level while the LAR noise fades, the re-rating has a financial floor under it. If tariffs and steel costs keep eating the spread and margins ratchet down instead, then the "improving underneath" story fails and the stock is just expensive on a handful of good quarters. Buyers are paying for the margin recovery — watch that line, and let the cash and the backlog tell you whether the rest of the year matches the operating path rather than the EPS echo. The numbers you can trust are the ones that show up in cash and bookings, not the ones that showed up in this quarter's headline.