Onto Innovation Is Guiding to 45% Growth Into a 45% Drawdown
A $1.3 billion backlog and a 58% gross-margin exit rate — and a stock that has lost nearly half its value since June. The report is about which side the tape is disputing.
ONTO · Earnings Call · 2026-11-09
The numbers refuse to argue
Onto Innovation's fiscal third quarter was, on its own terms, close to flawless. Revenue of $410 million rose 19% sequentially and nearly 90% year-over-year. Gross margin climbed another 60 basis points to 57.6%, operating margin reached 32.6%, and earnings per share of $2.44 beat the high end of guidance. Management then raised the fourth quarter to $430–450 million, which puts the full year near $1.485 billion — growth above 45% — with a gross-margin exit rate slightly above 58% and operating margin near 34%. Because the filed statements lag the call by a quarter, the reported numbers still show the July quarter: $343 million of revenue, up 35% year-over-year, on a 53.4% gross margin and a 20.0% operating margin. In the three months since, gross margin gained four more points and operating margin gained more than twelve. That is not a demand story with a margin kicker; it is a manufacturing story riding a demand wave — the extended factory model pulling production into lower-cost labor markets, shortening freight lanes and, as Brian Roberts put it, eliminating "some of the duties and tariff risk that was certainly impacting the business in 2025." Cash generation is doing the same thing. Free cash flow before stock comp was $48 million in the July quarter, up 204% sequentially, while R&D ran about 11% of revenue. The company also bought back $205 million of stock in the quarter.The tape is voting the other way
Here is the tension that makes this report worth reading. ONTO peaked at $378.45 on June 30, 2026. By November 9 it was down 48.7% from that high, and down 36% in just the last ninety trading days. Against the $8.5 billion market cap and a guided $1.485 billion of 2026 revenue, the shares trade near 5.7x forward sales — a multiple that would have been unthinkable for this business two quarters ago, when price-to-revenue sat at 16.6x trailing. That divergence is not idiosyncratic. The global tape's biggest ninety-day decliners are precisely Onto's customer complex: AI data centers down 23% with 57 names negative, high bandwidth memory down 29%, co-packaged optics down 28%. The market is not questioning whether Onto can ship; it is questioning whether the buildings those tools serve keep getting financed. Onto's entire equity story now sits on the wrong side of that macro argument, which is exactly why the earnings call's most important exchange was not about any product.What is genuinely new: measuring inside the die
Beneath the backlog noise there is a real, durable shift, and it shows up as the company's top keyword of the quarter: capital intensity. Optical critical-dimension metrology is moving off the scribe lines and into the transistor. Michael Plisinski described the mechanics plainly: “There has been a real push to do more in-die metrology, so to directly measure the structures in the die, not the test structures on the streets.” — Michael Plisinski, Chief Executive Officer · 2026-11-09 That matters mechanically. Measuring device structures rather than throwaway test structures requires a smaller spot, greater precision and heavier data modelling — and, critically, it converts metrology from a monitoring function into a process-control function, which lifts sampling rates and therefore tool count per fab. The Atlas G6, which does this, set a record as the fastest OCD platform launch in company history. Advanced nodes are on pace for more than 35% growth this year with an even higher rate expected in 2027; management said confidence is highest in logic and DRAM, with NAND the laggard. Packaging is the louder engine. The Dragonfly platform carried nearly all the growth in specialty and advanced packaging, advanced packaging is tracking to grow north of 80% in 2026, and 2027 purchase orders for the G5 already equal double the tool's entire 2026 revenue. Panel revenue should double this year with backlog supporting over 40% growth next year and orders extending into 2028. Silicon photonics plus power doubled sequentially off a small base.The backlog is a customer statement, not a supply failure
Backlog reached $1.3 billion, up 20% in a single quarter, with visibility "in some cases, into 2028 and beyond." The obvious skeptical read — that lead times are blowing out — was put to the CEO directly, and the answer was firm: “It's not a function of us not being able to deliver to the demand... customers are planning ahead, and they want to secure our tools for their ramps.” — Michael Plisinski, Chief Executive Officer · 2026-11-09 That is consistent with the prior quarter's disclosure of a $1.1 billion backlog with “30%, 40% of it covering 2027” — Craig Ellis, Analyst · 2026-08-06, and with the August claim that “we had the capacity for $2 billion with our in-house factories... we're not capacity constrained.” — Michael P. Plisinski, Chief Executive Officer · 2026-08-06 The narrative has been continuous across three calls; what changed is magnitude, not direction. Which means the burden of proof is now on customers honoring schedules in a market that is actively debating data-center returns.The question that mattered
Melissa Weathers of Deutsche Bank asked the only question that could move the stock — whether Onto has seen any change in order patterns given "the uncertainty around data center spending that we've seen the last couple weeks."The tell, though, is that data center enters Onto's own keyword set for the first time this quarter — and it arrives as a risk to be managed rather than an opportunity to be sold. When a supplier's vocabulary shifts from "AI packaging" to "concerns about data centers," the narrative regime has changed even if the orders have not.No, not at all. Not at all. The demand was already far exceeding capacity, so kind of some slowdown on some data centers that already probably couldn't get enough chips... I don't see any slowing down from our customers at this time, that's for sure.
What went quiet
Rigaku. In May the 27% stake was a strategic pillar with a licensing revenue stream, a hybrid metrology roadmap, and a stated plan to recover the forgone interest income. Brian Roberts, then:Today the only mention is a unrealized loss of roughly $40 million, marked quarterly against Rigaku's JPY 1,740 share price, excluded from non-GAAP, alongside $720 million of cash paid in August. Hybrid metrology was Onto's number-two company keyword in the June quarter; it has since dropped out of the conversation entirely. The Rigaku transaction has been demoted from a three-to-six-year strategic option to an accounting line, and that silence is the most informative thing on the call.We feel very confident that we will more than pick that back up through the combination of the licensing revenue that we have talked about as the primary revenue stream, plus the dividend income.