Onto Innovation | Earnings Review: High Bandwidth Mogging

Jestermaxxed in CoWoS but HBMogged Camtek at Micron

Opinions are my own and do not represent past, present, and/or future employers. All content is based on public information and independent research. This newsletter is not financial advice, and readers should always do their own research before investing in any security. I am invested in the semiconductor industry. As of the date of this publication, I may hold long or short positions in securities discussed in this article.

---

Pre-Earnings Vibe Check

(This section was written prior to the earnings call to communicate the prevailing narratives and provide a contrast to post-earnings sentiment)

I’ll keep this section short for today because the thesis I published yesterday says most of what’s necessary already. Required reading before you continue!

Onto trades at $220, up 30-ish percent YTD, driven mostly by very very robust strength in the semicap factor. This is all beta no alpha IMO, so the idiosyncratic situation hasn’t changed much.

Consensus Expectations

Expectations for Q4 2025 revenue is $266mm, gross margin 54.5%, EPS $1.27.

Expectations for Q1 2026 revenue is $265mm, gross margin 55%, EPS $1.25.

In addition, they revised their 1H26 over 2H25 growth from 5% to “at least 10%” at a recent conference. They reiterated expectations of 20% advanced packaging growth.

Therefore, they probably need a moderate beat on these to stay flat. Big beat is big upside because something unpredictable happened, in-line will be downside. Camtek only beat slightly (1-3%) and was down to flat today.

CoWoS Share Recapture at TSMC

The most important indicator today is commentary around TSMC qualification. Based on recent sellside reports, they seem to be “far along” and could be within 3-6 months. They have passed “optical validation” which is NOT qualification but is a very good sign.

After that, Onto has specifically stated aspirations to regain 50% of lost share in 2026, which was also my assumption in my writeup. Watch for if that changes.

Share Gains at Micron

Onto has fully qualified for Micron. They are in and orders are flowing. The only question is how much.

There is some interesting Camtek commentary I wanted to address. They have flatly denied any share loss.

• “We have not lost any market share to competitors. We also estimate that we will be able to increase our market share this year.”

• Called HBM4 a “major opportunity” — claims their Hawk platform handles the higher density/accuracy requirements

• Got a $45M order from an “IDM customer” expanding capacity for AI components — sounds like a memory maker

• Says they’ll penetrate “additional production steps and expand our total available market”

However, there was also zero mention of the infrared integration problem.

I doubt this means that they would never lose share. It is expected that the incumbent would not admit to share loss. If Onto’s management specifically calls out infrared integration as an important competitive advantage and backs it up with tens of millions in new orders, Camtek’s defense is disproven.

Share Gains for Front-End Metrology (OCD and Thin Film)

This is the only theme I did not mention in my writeup. Onto also plays in front-end metrology and that segment is deserving of its own article (which will be Onto part 2).

Onto has two front-end metrology tools: OCD and thin-film. Right now, Onto and KLA sort of split front-end metrology 50/50. Onto has more share in OCD while KLA has more share in thin-film.

Any upside bias to this split means Onto go up, downside means Onto go down.

WFE Demand Upside

For advanced packaging, last quarter, customer conversations suggested the need for as much as 20% more tools in CY26 (according to Oppenheimer) which also matches what Onto themselves stated at their recent conference. My assumption for packaging growth this year is 25%.

Upside here comes from CoWoS expansion, HBM demand, and Intel’s EMIB.

In the front-end, all that matters is TSMC’s expansion of the N2 node. I don’t think we are yet at the point where Intel 18A-P and 14A can be considered upside, but any long-term commentary there can also be a source of strength.

The leading edge is in a massive shortage so let’s see if that shows up today.

The Print

Behold my Claude Code terminal tables once again

Q4 2025 was in-line. A bit disappointing, definitely some softness. We’ll see how they explain this on the call.

Moderate beat, above consensus, probably missed buyside bogeys slightly, enough to keep the stock flat I think.

However, the actual numbers aren’t what validates or disproves the thesis. Almost all the information should be related to commentary.

From the PR, we already get the following.

“Next generation Dragonfly shipped to multiple customers with evaluation periods now ongoing.”

That’s Dragonfly 5. Nothing new here. Looks like no major qualification news for TSMC before the call starts.

Hold on. Hold on. WOAH

“$240M+ volume purchase agreement with a leading HBM manufacturer for Dragonfly 2D inspection and 3D bump metrology through 2027.”

That’s almost certainly Micron. $240M through 2027 is VERY BIG. I assumed $30M in 2026, seems like I undershot by a lot. Let’s see how much of this ships in 2026.

Soft numbers + no TSMC news + BIG HBM order. Interesting…………

Excited for the call.

Price Action

flat ahh price action. nothing to write home about here.

The Call

Key Positive Themes

“Like Taking Candy From a Camtek”

“We have not lost any market share to competitors. We also estimate that we will be able to increase our market share this year.”

Sure, Camtek. A $240 million VPA from a competitor accelerating into 2026 at a customer where you have near-monopoly is a strange way to not lose share. Lol

Onto closed a volume purchase agreement worth over $240 million with a “leading HBM manufacturer” for Dragonfly 2D inspection and 3D bump metrology, covering expansion plans through 2027. $60 million of that is specifically for 3D bump metrology systems. They didn’t name the customer, but this is almost certainly Micron. Onto was already qualified there, we knew orders were flowing, and the language matches prior commentary perfectly.

I assumed $30 million in Micron HBM revenue for 2026 and I was off by a factor of 3-4x. About the delivery cadence: management says originally planned as 1/3 in 2026 and 2/3 in 2027, but demand is accelerating and the split is moving toward 50/50. That implies roughly $80-$120 million shipping in 2026 alone.

Even better, management said they “for sure expect additional VPAs” and are “in conversation with other customers.”

Let’s use our deductive logic here. There are 3 manufacturers of HBM in the world. Micron placed today’s order. Samsung already uses Onto, so it wouldn’t be new.

That leaves SK hynix.

Very good. Lots of upside here.

WFE Demand Go Brrrr

The demand environment is unambiguously strong and accelerating. A few numbers that jumped out:

• Advanced packaging revenue guided to grow over 30% in 2026. This is up from the 20% they cited at a recent conference.

• H1 2026 over H2 2025 growth was further raised from “at least 10%” (from recent conference) to 12-14%. This is the third consecutive raise on this metric across recent communications.

• Backlog nearly doubled to roughly two quarters of coverage. This matches the broader semicap theme of extending lead times.

Gross margin is expected to expand sequentially in each quarter this year. Combined with higher revenue, this means operating leverage is just getting started.

On capacity, Onto can serve up to a $2 billion annual run rate with existing factories and is ramping extended facilities. For a company doing $1 billion today with $2b of capacity and building more tells you where management thinks this is headed.

However, here’s what I think is most important. Onto guided advanced packaging revenue to grow over 30% in 2026 with explicitly zero contribution from Dragonfly G5 qualification and ramp at TSMC. They said the “over 30%” language is them trying to be conservative. The 30% growth is almost entirely driven by Micron HBM orders.

So any TSMC CoWoS share recovery via Dragonfly G5 would be incremental upside to that number. Management noted that the current mix of revenue additions (growth) is about evenly split between HBM and CoWoS but if G5 qualifies, the growth mix would “lean towards CoWoS.”

Key Negative Themes

Dragonfly 5 CoWoS Purgatory

No major qualification news. “While still early, initial impressions from customers are positive” is the same vague language we’ve been hearing. The next-generation Dragonfly (G5) has shipped to multiple customers with evaluations ongoing, but there is no change to previously communicated timelines.

They also sounded a bit dodgy around this question on the call. This probably either pressures the stock or keeps it from exploding from the massive HBM upside.

Front-End Metrology Cleanroom Constraints

We didn’t really get much from management with respect to OCD and thin film front-end growth.

Management says business is being driven by “the hardest and most difficult” nodes, meaning GAA. Nearly all the demand is concentrated there. The attach rate for OCD metrology at 5nm and above is much lower. This is obviously good because it creates a new market for Onto, but we already know this.

Onto expects two logic customers to ramp in 2026, which almost certainly means TSMC N2 and either Samsung SF2/Intel 18A.

They mentioned uncertainty in the timing, which they will refine when closing additional VPAs. Will grow “at least 15%.”

I was hoping for either higher visibility or more bullish commentary. However, I think there’s a major silver lining.

KLA barely discussed metrology at all on their prior 2 earnings calls. The product-level color they gave was:

• Inspection (broadband plasma / BBP): grew 25% in CY2025

• Patterning: grew 12% in CY2025, solid but lagging inspection

• Reticle inspection: “going to be very strong” tied to design starts

• E-beam: “positive momentum moving forward”

For metrology: Rick’s only comment was that it’s “historically a little closer tied to capacity” and they’ll “see some relative recovery.” “Relative recovery” implies KLA’s metrology business underperformed in 2025 and they expect it to catch up as capacity comes online.

So KLA and Onto are both seeing low visibility in front-end metrology. This is consistent with cleanroom constraints and timing of orders, not share loss on either side. Metrology tools go in later when fabs are actually being built out and need to calibrate processes, whereas inspection tools are needed earlier in R&D and process development.

Also some increasing contribution from some new products: New products across the portfolio currently represent about 10% of revenue, growing into 2027. This is still early-stage, but the insertion and initial ramp happening this year sets up a broader revenue contribution next year.

Conclusion

This was a mixed quarter with strong upside bias. Overall, the hair on the CoWoS and front-end metrology sides will prevent the huge HBM order from delivering massive upside for the stock. Overall, don’t think the stock makes a huge move in either direction.

However, I think the setup for the next few quarters got more positive today. Why?

The negative themes have a high probability of reversing while the positive ones will either stay or get better.

For the negative themes, in H2 we’ll see more color on front-end metrology as cleanroom constraints subside and we have the massive call option of Dragonfly G5 qualification at TSMC.

For the positive themes, the Micron HBM order ain’t gonna reverse. Instead we can see it get bigger or even see an order from SK hynix.