Onto Innovation Inc. ONTO

Key data

The business

ONTO Innovation makes the cameras and gauges that watch chips get built. Its systems sit inside fabs and inspect wafers for defects, measure film thickness down to the atom, and check the increasingly complicated ways chips get packaged together. That last part is the story. As Moore's Law slows, the industry buys performance by stacking chips, gluing memory on top of logic and fanning connections out sideways, and every one of those steps needs a machine to confirm it worked. ONTO's Dragonfly inspection platform is the tool many packaging lines reach for. The moat is qualification lock-in: once a metrology recipe is dialed into a customer's production line, ripping it out to save a few dollars risks yield on a billion-dollar wafer run, so nobody does.

The customer book is narrow and getting narrower at the top. Four customers were 57% of first-half revenue, the largest at 16.2%, down from 21.6% a year ago. The money follows the leading-edge fabs: Taiwan $190M, the US $168M and South Korea $167M split the first half almost evenly, while China is a rounding error at $32M, which matters as export rules keep tightening. The US line is the tell, tripling from $53M a year ago as new American packaging capacity switches on.

The numbers

The Q2 print (period ended June 30) is a genuine inflection, not a drift. Revenue of $343M is the largest quarter ONTO has ever booked.

Revenue rose 35% against the year-ago quarter and 18% against Q1. Net income jumped 77% year over year as gross margin widened to 53.4% from 48.2%. The non-GAAP $1.93 beat the $1.69 analysts modeled. What the table hides is how bumpy the path was: the $218M trough in Q3 2025 is the cyclical low the business is climbing out of.

Here is the trap the annual table sets. Revenue sat in an $800M-to-$1B band for four straight years, a business that looked cyclically capped, and FY2025 earnings actually fell, EPS $2.78 against $4.06, as margins absorbed cost and mix. But the first half already booked $635M, a pace annualizing near $1.27B, and the forward consensus of $2.19B for FY2028 assumes the pending Rigaku deal folds in. That is the whole argument: is the $343M quarter the new floor or a spike back toward the old ceiling? The variant view is that the packaging cycle has structurally lifted ONTO's demand, and the print that settles it is whether second-half revenue holds the $340M quarterly pace rather than fading to the $250M band it lived in through 2025.

Management

Insiders are quiet, which is neither comfort nor alarm at this size. The only open-market activity in a year was Susan Lynch selling $302,300 in August; plan status is not disclosed, so read it as a single officer trimming, not a signal. The company repurchased $75M of stock in FY2025 and pays no dividend, funding itself through a cash pile that ballooned to $1.25B, plus $629M of short-term securities, after ONTO raised roughly $1.5B of debt to arm the Rigaku transaction. That debt is new for a company that ran essentially unlevered for a decade, and repaying it now sits explicitly on management's own stated worry list.

How it fails or surprises you

Customer concentration cracks. Four names are 57% of revenue and the top one already slid from 21.6% to 16.2% in a year. A single leading-edge customer pushing out a packaging line, common in this industry, can knock a quarter back toward the $250M band overnight, and a 100x trailing multiple gives that miss nowhere to hide.

Rigaku disappoints or slips (downside). The filing repeatedly flags the deal may not close "on the timing expected or at all," and $1.5B of fresh debt rides on it. An integration stumble or regulatory delay leaves the balance sheet levered against a promise consensus has already booked into FY2028 numbers.

Advanced-packaging demand compounds (right tail). US revenue tripled to $168M as American fabs light up, and HBM stacking for AI accelerators is early. If the second half holds the $340M pace, the business earns its way from 100x trailing to under 20x forward, while the market is still pricing the old cyclical ceiling.

Closing thoughts

If second-half 2026 holds near $340M a quarter, the trailing 100x GAAP multiple was an accounting mirage thrown off by the Rigaku build-out, and the stock is cheaper than it looks. If it fades toward $250M, the four-year revenue ceiling reasserts and the multiple has nowhere to go. The re-rate flagged three sessions ago, forward earnings mattering far more than the ≈100x trailing print, has not been tested by any new filing since, and the stock has simply drifted toward its 200-day while everyone waits. The fatter tail, on tripling US demand and widening margins, points up, but the debt taken on for Rigaku is what turns a stumble into real damage rather than a soft quarter.

The bet is still that chipmakers keep needing ONTO's machines to inspect the stacked, packaged chips that AI runs on, and the handful of big customers keep ordering. What breaks it is customer concentration meeting a Rigaku that slips. The one pair of numbers that tells you first is second-half quarterly revenue against the $340M line and the top customer's share against last year's 16.2%. The view holds only as long as the quarters stay above the old ceiling.