Camtek and Hawk: Is the AI order wave broadening beyond the first IDM?

Why This Follow-Up Exists

The main article already argued that Hawk is not just another slide deck product. It is the main candidate to turn Camtek from a strong growth story into a bigger scale story. This follow-up isolates one narrower question: does the January and February 2026 evidence already show that the AI order wave is broadening beyond the first account publicly disclosed, or are we still mostly looking at deeper penetration inside one customer that has already cleared qualification?

These three documents create a clean sequence. The 2025 20-F shows that revenue growth came primarily from a higher average selling price for product units, driven by Hawk and Eagle G5. The January 2026 investor presentation then jumps to a much more ambitious framing: 2026 should be a growth year and a milestone on the way to a $750 million target model. Then the February 2026 report adds the most concrete commercialization proof yet, cumulative Hawk orders of $45 million from the same IDM for delivery during 2026.

But those are still three different levels of proof. The first is technological proof, that the product really opens a market and can perform. The second is deep commercial proof, that one customer has already moved from evaluation to repeating orders. The third, and the one that matters most for the 2026 thesis, is breadth: whether the success at the first disclosed account is already starting to show up across more customers, or at least in a broader revenue structure that can justify the $750 million model.

What Hawk Has Already Proven

At the product level, the evidence is strong. Both the presentation and the 20-F describe Hawk as a new platform launched in February 2025 for the most demanding advanced packaging applications, including Chiplets, HBM and Hybrid Bonding. The system is designed for high-end use cases, supports 150 nanometer defect detection, and enables inspection and metrology of up to 500 million micro bumps at tight pitch. The January presentation adds an important commercial layer: Hawk is meant to operate at twice the throughput of Eagle G5 and expand Camtek's addressable market.

At the level of 2025 economics, Hawk already mattered, but the exact way it mattered is important. Camtek finished 2025 with revenue of $496.1 million, up 16% from 2024. Management's explanation matters a lot: the increase came primarily from a higher average selling price of product units as a result of Hawk and Eagle G5. That means the new generation already improved revenue quality and pricing, but it still does not mean Hawk alone has already created broad expansion in customer count or unit volume.

The February 2026 report adds the piece that had been missing until then, proof of depth at one specific account. Camtek disclosed a new Hawk order of approximately $25 million from a tier-1 IDM for AI applications, and said this followed a series of smaller repeating orders from the same IDM, bringing cumulative orders to $45 million for delivery during 2026. That no longer sounds like a pilot, a demo tool, or a one-off placement. It sounds like a customer that has moved beyond qualification and into repeat purchasing behavior.

The 20-F helps interpret why this matters. Camtek explicitly says that selling a new product, especially to a new customer or into a new market, involves demonstrations, technical presentations, site installation, and side-by-side evaluations for about six months, with revenue recognition sometimes taking several quarters. By contrast, repeat orders from existing customers should move faster. That makes the February order sequence exactly what you would expect from a product that has already passed qualification at one account. It is proof of depth. It is not yet proof of breadth.

Where The Proof Still Stops

The gap between January and February is the heart of the story. In January, management was already presenting a new $750 million target model, assuming substantial organic growth, and saying that 2026 should be a growth year and a milestone on the way there. The presentation also lays out the main drivers: higher AI demand, more inspection and metrology volume and steps, more software and services revenue, and a new product generation. That is already a jump from product proof to broad engine framing.

The problem is that direct Hawk disclosure is still much narrower than that framing. The presentation says Camtek has more than 300 customers and that 70% of its business is Tier 1. That absolutely supports the idea that the company has the customer base and competitive position to scale Hawk faster than a smaller entrant could. But it is still disclosure about Camtek as a whole, not Hawk specifically. There is no detail on how many customers have already bought Hawk, how many are still evaluating it, or how much of the AI wave has already translated into repeat orders beyond the IDM disclosed in February.

There is a second limit that matters just as much. The 20-F says one customer represented 11% of 2025 revenue, compared with three material customers in 2024, but it does not identify that customer and does not link it to Hawk. So there is no way to know whether the IDM from the February report is already the major 2025 customer, a new growth layer, or an account that was still smaller in 2025 and becomes more meaningful only in 2026. That is a critical distinction, because without it, it is hard to separate true market broadening from faster deepening at a customer that may already have been important.

This chart does not create an accounting bridge from order to revenue, and it does not assume the full $45 million will be recognized on any exact timeline. What it does show is scale. The cumulative Hawk order from the same IDM is meaningful, but it is still small relative to the gap between actual 2025 revenue and the $750 million target model. Anyone treating the February disclosure as full proof of the January model is moving too fast.

The $750 Million Model Needs More Than An Expensive Hardware Box

The January presentation does not build the target model only on shipping more Hawk systems. It explicitly says that the move in gross margin from 51% to 52% toward 54% to 55% is also supposed to come from higher software and services revenue, alongside new-generation products. That matters because it means full Hawk commercialization is supposed to show up not only in more expensive systems, but also in a broader monetization layer around them.

In 2025, the revenue base still looks overwhelmingly hardware-heavy. Product sales were $468.5 million, while service fees were only $27.6 million, about 5.6% of total revenue.

That means the $750 million model needs more than two things that are already proven, a premium product and one repeat customer. It also needs the company to expand software and services from a still small base, and to show that the new generation is not staying concentrated in one account. In that context, the annual report includes one more useful clue: during 2026 Camtek expects to offer AI technology that includes inspection and automatic defects classification. That is a concrete commercialization bridge, because if that AI layer starts to be sold or improves the services mix, it can help turn Hawk from a high-ASP hardware tool into a broader economic engine. If not, the target model remains heavily dependent on continued system shipments.

Why The Market Should Be Careful With Depth Disguised As Breadth

This debate matters now because Camtek is already operating as if AI demand is not a side story but a planning base. In the 20-F the company says that to support current sales volume and expected demand it expanded workforce, increased production capacity, raised capital expenditures to $22.5 million, and often pre-orders components and subsystems based on internal forecasts rather than firm purchase orders. That makes perfect sense if the market is broadening. It also raises the cost of being wrong if depth at one account does not roll into broader commercial adoption.

That point is especially important because 2025 already proved that Camtek can monetize a richer mix. The missing proof is different: whether that same economic benefit becomes repeatable across several customers, or remains a strong but concentrated jump. February 2026 moves the answer forward because it shows the first account is no longer stuck at a single order. But it still does not answer the bigger question that January was already trying to pre-price, how wide the adoption wave really is.

Conclusion

Hawk has already passed the launch test. It proved technological value, helped lift ASP in 2025, and reached the stage of cumulative repeat orders of $45 million from the same IDM for delivery during 2026. That is real progress, and it is enough to say Camtek is no longer selling only a product dream.

But the sharper conclusion is that the evidence still supports depth at the first publicly disclosed account more than it supports already-proven commercial breadth. The January presentation speaks in the language of a $750 million model, more software and services, and 2026 as a growth year. The February report proves that part of the road there is real. It still does not prove that the whole road is already paved. Put differently, 2026 currently looks more like the proof year for broad Hawk commercialization than the year in which that proof is already complete.