Ryerson Beats Everything, Then Guides Like It Doesn't Quite Believe It

A record $2.2B quarter, a transactional-program spread that blew out to 1,000 basis points, and a data-center number that quietly became 9% of revenue

RYZ · Earnings Call · 2026-10-29

Beating the quarter, hedging the year

Ryerson did not just beat its third quarter — it beat it in the direction opposite to its own guidance. Management had told investors to expect a seasonal shipment decline of 3%–5%; instead shipments rose 1.2% sequentially to 814,000 tons, same-store shipments climbed 1.8% to their highest level in more than five years, and revenue hit a record $2.2 billion. Adjusted EBITDA excluding LIFO came in at $115 million against a guide of $88–92 million, with the newly acquired Olympic Steel contributing $30.2 million. Diluted EPS of $0.61 roughly doubled the prior quarter's $0.30. Then came the fourth-quarter outlook, and the tone changed. Shipments are guided down 6%–8% sequentially, revenue to $1.96–2.04 billion, adjusted EBITDA excluding LIFO to $100–104 million. KeyBanc's Samuel McKinney asked the obvious question: with October quoting and ordering described as strong, why would the combined company merely meet normal seasonality? Eddie Lehner's answer was disarmingly candid about the limits of his own visibility: “we are going to have to wait until February to answer that” — Eddie Lehner, Chief Executive Officer · 2026-10-29. That gap between a blowout print and a conservative forward frame is the central tension of this call.

The data-center number that finally got a number

What genuinely changed on this call is quantification. A year ago, when Katja Jancic of BMO asked how much exposure Ryerson had to data centers, Lehner waved it away: “It's a subvertical, Katja. It's really hard to get an exact fix on that.” — Edward J. Lehner, President and Chief Executive Officer · 2025-07-30 Two quarters ago, data centers was a qualitative tailwind buried in the end market discussion. On Thursday, Rick Marabito put a ruler on it: data center and power generation projects are an estimated 9% of revenues, up from 6% in the second quarter, with data-center volumes up roughly 65% since the first quarter of 2026. That is a company that has stopped describing a theme and started reporting it. It is also a stark contrast with the tape. In the Know Trend global price history, the AI data center cluster is a 360-day winner — 44 advancers against 12 decliners — but is down 5.3% over 90 days and 30 days, with 49 names negative. The market has already begun fading the theme Ryerson is accelerating into. That divergence cuts both ways: either Ryerson's order book is a lagging confirmation screen for physical metal demand (copper busbar, enclosures, cooling, HVAC) that equity investors are pricing too early, or the company is leaning into the most crowded part of the industrial cycle just as it turns. Lehner's own framing acknowledges a bifurcated economy — a “two-speed economy with AI infrastructure, energy, defense” — Eddie Lehner, Chief Executive Officer · 2026-10-29 growing while agriculture, consumer, and rate-sensitive markets stay subdued.

The spread that is widening, not closing

The most interesting analytical thread is one management did not volunteer. On the July call, Lehner sized the margin gap between a transactional ton and a program ton at “between 700 and 800 basis points” — Edward Lehner, Chief Executive Officer · 2026-07-30, and noted the historical norm was “about 600 to 700 basis points difference” — Edward Lehner, Chief Executive Officer · 2026-07-30. On this call, asked by the same analyst whether program margins were inflecting, he supplied a number that moved the wrong way:

Q4 is really off to a very promising start, and we've seen the margin differential between a transactional pound and a program pound, even though both of them are moving higher. That spread has gotten up to about 1,000 basis points.
Eddie Lehner, Chief Executive Officer · 2026-10-29

Both sides are rising — but the spread blew out by 200–300 basis points, not in by 100. The mechanics are straightforward and were explained well: most of the program book resets on a quarterly lag, so when carbon pricing grinds higher month after month, COGS outruns the program customers price reset. The tell is the LIFO line — $30 million of expense against guidance of $16–18 million, a near-double miss that is the accounting fingerprint of input costs rising faster than management expected. The LIFO expense guidance remains elevated in Q4 at $28–32 million. So the honest read is that the program side business is healing on volume but still losing on price — and the spectacular Q3 was carried by spot, not by the contract book. That is a lower-quality earnings mix than the headline suggests, and management's pivot in every answer toward the transactional book is itself the concession.

Synergies are real; cash flow is the invoice

The integration story, by contrast, is delivered with rare specificity. About $13 million of synergies landed in Q3, and the company exited ahead of its first-year $40 million target before the February merger anniversary, guiding to a $56–60 million annualized run rate by year-end against a two-year $120 million goal. The commercial synergy line is closing on its own two-year target early. Lehner offered a genuinely useful operating stat — productivity of 2.25 tons per employee per day — and cited expense leverage as the reason WDSG&A fell from 16% to 15% of sales. The combined network is no longer a slide; roughly 90% of Olympic's inventory and transaction data now sits inside Ryerson's system. But the balance sheet paid for the growth. Operations consumed $18.2 million of cash in the quarter as working capital expanded, net debt rose $53 million to $966 million, and inventory days held at 73. The effective net cash line was already negative $912 million at the June quarter-end, and it will look worse before it looks better. Leverage fell to 3.2x from 4.0x, and management reiterated a path below 3.0x by year-end — the same deleveraging priority that Rick Marabito called “keep the main thing the main thing” — Richard Marabito, President and Chief Operating Officer · 2026-02-20 back in February. Gross margin was 17.7% in the June quarter before slipping 50 basis points in Q3. That puts Ryerson roughly in the second quartile of its own historical band — a long way from the 23.7% peak of 2022 but far from the 13% trough in 2022Q4. The stock knows this: RYZ is up 16.8% over the last 90 trading days, yet still sits 29.7% below its July 2023 peak of $44.08 on a $1.6 billion market cap.

The verdict

What changed is not the story — it is the arithmetic. Data centers went from an unquantifiable subvertical to 9% of revenue, the transactional-program spread went from 700 basis points to 1,000, and the merger's first-year synergy target fell a quarter early. What did not change is the thing investors have waited through the entire merger cycle for: the contract book catching up on price. Until annualized run rate savings convert into margin rather than into working capital, the record revenue line will keep flattering a company that is still, operationally, playing catch-up.