Camtek Sells 2027, Then Sends the Bill to the Gross Margin Line

Camtek's third quarter was, on the numbers, the best in its history: revenue of $161.3 million, up 28% year over year and 21% sequentially, operating income of $47.5 million at a 29.5% margin, and a fourth-quarter guide of roughly $180 million that implies 40% growth. Management then did something it has never done — it pointed at 2027 and put a number on it.

“For the first time in Camtek history, we have visibility extending approximately one year ahead.” — Rafi Amit, CEO · 2026-11-04

That is the headline. The story is what it cost to get there, and what management quietly stopped promising along the way.

The order book nobody wants to quantify

Since January, Camtek has booked more than $800 million of orders — against full-year 2026 revenue of roughly $595 million. About 80% is advanced packaging applications, and slightly more than half comes from OSATs, the outsourced assembly houses where Camtek holds, in Ramy Langer's telling, a dominant position. Management attributes the historic order momentum to customers reserving future floor capacity: “customers are willing to give us today orders a year ahead, and this is not that they necessarily need the machines today. They just want to secure capacity on our floor.” — Ramy Langer, Executive · 2026-11-04

Yet when analysts pushed for the split between 2027 revenue already in backlog and revenue still to be booked, the answer was a wall. Langer, on the initial $800 million outlook: “we cannot give more information at this stage.” — Ramy Langer, Executive · 2026-11-04 On share gains: too early. On hybrid bonding revenue: not substantial until 2028. On whether the new increase in prices is embedded in the $800 million or incremental: “it's too early to say. I do hope that there is some upside to these numbers.” — Ramy Langer, Executive · 2026-11-04 For a company that just told investors it can see twelve months out, the refusal to decompose the number is conspicuous — and consistent with prior quarters, where Langer repeatedly declined to split HBM from CoWoS or 2026 from 2027.

The margin promise that moved

Here is where the report diverges from the pitch. Gross margin landed at 51.1%, below internal expectation, because of what Moshe Eisenberg called industry-wide supply chain pressure that “resulted in price pressure on parts, components, and material used in our system, affecting our COGS.” — Moshe Eisenberg, CFO · 2026-11-04

This was not a surprise risk. In May, an analyst asked directly whether rising component pricing — DRAM, general parts — was a margin headwind. Langer acknowledged “some pressure from the supply chain” — Ramy Langer, Chief Executive Officer · 2026-05-12 and said the second half would still show gross margin improvement. In February, Eisenberg had gone further, claiming “we do have the capacity, we have the inventory and all the supply chain ready for the growth.” — Moshe Eisenberg · 2026-02-18 And in August, the CFO set an explicit target: “We certainly hope that we will get anywhere between 52.5% to 53% exiting the year.” — Moshe Eisenberg, Chief Financial Officer · 2026-08-10

As of this call, that aspiration is gone. The new language is gradual improvement starting from the first quarter of 2027. Same company, same demand, one quarter later — and the margin exit-rate target has slid a full calendar year. That is the real change in this report, and it is not in the press release.

I expect these topics to be resolved, I would say, within three to six months. I think it will balance off. But this is really my feeling. We're not sure about it.

The remedy has two legs: pushing component cost through to customers, and adding capability to justify higher average selling prices. Tellingly, the company is no longer describing margin expansion purely as a mix story about Hawk and Eagle G5 — it now requires price. And the mitigation plan leans on second source qualification for memory, CPUs and machine handling components, which is a procurement fix, not a structural one.

The software promise, deferred twice

The second deferred item is software. Camtek bought Visual Layer to build an AI-driven defect classification and metrology platform, and in May Langer said: “I do believe that we will see the contribution of revenues from these capabilities in the second half of this year.” — Ramy Langer, Chief Executive Officer · 2026-05-12

Today: “At this point, we have pretty much zero software offerings.” — Moshe Eisenberg, CFO · 2026-11-04 First revenue is now expected in the first quarter of 2027, double-digit millions across 2027, and roughly $40 million within the three-year plan toward $1.2 billion of revenue at 34%–36% operating margins. The software offering is real and now formally underwritten in the target model — but it arrives two-plus quarters later than advertised, which is exactly the kind of quiet slippage analysts should price with skepticism, not enthusiasm.

Similarly, hybrid bonding keeps being pushed: it was a declining theme in prior quarters and now management is explicit that 2027 will contain no substantial hybrid-bonding business. And CoWoS — a multi-quarter obsession — has faded from the narrative; the incremental upside management now points to is share gain in 2D inspection, not the leading-edge interposer ramp.

The tape disagrees with the order book

Camtek's demand signal is accelerating. Its end-market tape is not. Over the last thirty days the wider AI hardware complex has been a bloodbath: the AI data center theme is down more than 20% across some 56 tickers, high bandwidth memory down roughly 36%, co package optics down about 36%, data center capacity down nearly 40%. Camtek's customers' customers are being de-rated precisely as Camtek's order book reaches a record.

That divergence is the whole investment question. Either the selloff in AI infrastructure equities is a multiple reset that leaves an inspection and metrology supplier with a one-year backlog untouched, or the 2027 slot reservations are a double-ordering artifact that will be re-timed. Management sides emphatically with the former, noting the install base of roughly 4,000 systems and the sticky, customization-heavy nature of the OSAT relationship against bigger competitors. On the competitive question — the large players and Chinese entrants — Langer was direct: “We do not see any competitive threat from those small competitors in China.” — Ramy Langer, Executive · 2026-11-04

Bottom line

Camtek is a genuinely rare setup: a $5.2 billion company with $877 million of cash, a record $161 million quarter, orders exceeding 1.3x forward revenue, and a credible path to doubling revenue to $1.2 billion in three years. But the report's honest content is that growth is now being bought with price increases and component scrambling, that the 52.5%–53% gross margin exit rate promised in August has become a 2027 aspiration, and that the software annuity that was supposed to start this year starts next. Watch the first-quarter 2027 gross margin print: if it does not inflect, the $1.2 billion target's 34%–36% operating margin math becomes arithmetic that does not close.