COHU (COHU) Q1 2026 earnings call transcript

I would now like to hand the conference over to Jeff Jones, Chief Financial Officer. Please go ahead. Good afternoon and welcome to our conference call discussing COHU's first quarter 2026 financial results and our outlook for the second quarter of 2026. I'm joined today by Luis Mueller, COHU's President and CEO, and Matt Hutton, COHU's VP of Strategy and Investor Relations. If you need a copy of our earnings release, it can be found on our website at cohu. com or by contacting COHU Investor Relations.

A slide presentation accompanying today's call is also available in the Investor Relations section of the website. Replays of this call will be accessible via the same page after the conclusion of the call. During this call, we will be making forward-looking statements that reflect management's current expectations concerning COHU's future business. These statements are based on information available to us at this time, but they are subject to rapid and sometimes abrupt changes.

We encourage everyone to review the forward-looking statement section of our slide presentation and the earnings release, as well as COHU's filings with the SEC, including the most recently filed Form 10-K and Form 10-Q. Our comments are current as of today, April 30, 2026, and COHU does not assume any obligation to update these statements for events occurring after this call. Additionally, we will discuss certain non-GAAP financial measures during this call. Please refer to our earnings release and slide presentation for the reconciliation to the most comparable GAAP measures.

Now I'd like to turn the call over to Luis Mueller, COHU's President and CEO. Luis?

Good day, everyone. Thank you for joining our Q1 2026 earnings call. We started the year with strong momentum across multiple product lines with orders up 57% year-over-year, reflecting both improved semiconductor market conditions and the increasing relevance of our technology portfolio across AI and high-performance compute applications. An important driver of this momentum is the expansion of AI workloads and inference processing, driving greater computing power density that has become a primary bottleneck.

AI accelerators and HPC processors generate immense amounts of heat during operation. Testing these chips requires maintaining precise temperature environments to ensure functional accuracy and long-term reliability. If a chip is tested at the wrong temperature, its performance metrics may be skewed, leading to lower yields or worse, latent field failures. As a result, COHU's proprietary and industry-leading thermal capabilities are highly valued by customers. Based on current engagements and design activity, we now see a computing segment opportunity pipeline of approximately $750 million, including roughly $650 million in test handlers and an additional $100 million from HBM inspection, and both growing at rapid rates. For fiscal 2026, we're now increasing our high-performance computing revenue outlook to approximately $80 to $100 million.

We are emboldened by the opportunity pipeline across 12 customers, with five customers in qualification stage and another seven in early engagement stage. During the first quarter, we continue to benefit from rising device complexity, higher power density, and accelerating AI adoption, trends that are reshaping test, inspection, and manufacturing requirements across the semiconductor value chain. In fact, semiconductor value is moving to the mid and the back-end manufacturing, driving substantial growth in the test arena. The estimated semiconductor test serialization also increased sequentially to 78% at the end of the first quarter. Automotive and industrial markets are gradually improving again as customers started investing in test capital. Many of our customers are broadening their product portfolio to serve AI data centers, as these transition to 800-volt DC infrastructure and more power management-efficient solutions with Gallium Nitri technology at RackScale's server board, such as the new Vera Rubin platform.

Across each of these applications, our customers are prioritizing quality, performance, and scalability. At the same time, software platform gained traction as analytics moved from pilot deployments into broader production environments. These wins validate both the technical performance of our solution and the growing appetite for software-enabled yield and productivity investments. There's a significant SAM opportunity for COHU in this space and a significant lifetime value in software subscription. This is illustrated well in the first quarter, when a $20 million system order came together with $330,000 a year of software subscription, which over the course of the lifetime of these systems could yield approximately $5 million in recurring revenue. The financial implication of this shift is twofold.

First, software subscriptions provide high margin, recurring revenue that is less susceptible to capex cycles. Second, by improving overall equipment efficiency and reducing mean time to repair for customers, we build deep operational stickiness that makes it difficult for competitors to displace our systems. I would now like to highlight a few customer wins in the first quarter. Starting with our test handler business, we've orders up 54% year over year. We secured two major Eclipse orders in the first quarter. The first win supports AI data center applications with a U. S. fabulous customer developing server and inference devices. As power density and mechanical complexity increase, Eclipse, combined with our T-Core active thermal control, enables the customer to standardize on a common handler platform across multiple device generations.

This reduces capital risk while extending the life and value of the install bays. Closed loop junction temperature control was a key differentiator, ensuring consistent temperature test quality, higher yields, and faster production ramps. In addition, the customer is adopting COHU's base prescriptive analytics software to improve equipment efficiency, increasing system value, enabling recurrent revenue for COHU, and strengthening long-term engagement. Again, strategically, this win deepens our computing footprint, embeds Eclipse into the customer's roadmap, and positions us as the platform of record, representing an estimated $100 million incremental revenue opportunity at this account over the next three years. The second order supports data center computing, mobile, and automotive processors, and another EOS-based Fabless customer using the Eclipse platform.

Our solution allows both the customer and their OSATs to address multiple markets while leveraging T-Core thermal control to maximize yield and asset utilization. Together, these strengthen our engagement across high-performance computing and AI markets, driving near-term system revenue and long-term platform, software, and recurring value growth. Our customer engagement for Eclipse expanded in the first quarter with an additional five customers in different stages of qualification, representing an incremental $200 million of revenue opportunity starting late this year and into next year. We are very bullish about the customer traction and the growing opportunities to expand our presence in the $750 million high-performance computing market.

These opportunities are rapidly taking shape as compute power increases and with the need to actively manage silicon junction temperature at higher power and power densities. Now, turning to our inspection and metrology business with orders up 64% year over year. In HBM memory, we continue to see stronger mentor for final inspection of HBM3 and HBM4. We're investing in this market and keeping pace with design requirements to support next-generation HBM5. We're now forecasting revenue growing 80% year-over-year to approximately $20 million with our Neon HBM platform. In the first quarter, we also secured a significant volume repeat order for our EON inspection system from a U. S. headquarter and also from a Korean customer.

Our inspection business is growing fast, and we estimate revenue at approximately $70 million this year. Semiconductor test orders recorded an impressive 163% increase year over year. Headlines around AI infrastructure typically focus on the massive compute devices required to train and run large language models, along with the memory and networking technologies that enable scale across the data center. Less visible, but equally critical, is power delivery. Every AI system depends on precise, efficient power management to sustain peak performance. This is where the DiamondX precision instrumentation becomes decisive.

Our tester was qualified for testing power devices, strategically expanding our footprint in AI-related applications, and embedding it more deeply into our customer's roadmap. As power density increases, customers are implementing GAN-based technology to minimize energy loss and thermal impact. While GAN offers a clean efficiency advantage, it remains less mature than traditional CMOS, creating technical and economic challenges as customers scale production to meet data center demand.

Moving to our interface solutions group, we've seen increased adoption of our higher current contactors for AI power applications at existing customers. We also expanded our product offering and received multi-unit order for a new silicon photonic solution. These photonic switches formed the backbone of cloud and AI Ethernet fabric, and we're now testing them.

In closing, Q1 was a strong start for the year and a clear validation of our strategy. We see momentum rapidly build across AI infrastructure, high-performance compute, power management, and smart manufacturing, driven by rising device complexity and increasing power density. Our expanding presence in thermal handling, advanced inspection, precision test, and high-value software is translating into larger platform wins, recurrent revenue opportunities, and deeper customer engagement. With a million computing segment opportunity in front of us and improving utilization across our core markets we are accelerating our investments to capture new customers and we are expanding production capacity to move confidently through the remainder of this year and into 2027 These secular tailwinds combined with disciplined execution and continued investment in innovation position COHU to deliver durable value for our customers and shareholders Thank you for your continued support.

I will now turn the call over to Jeff for a deeper review of our financial results and forward-looking guidance. Jeff?

Thank you, Luis. Before reviewing the first quarter results and providing second quarter guidance, please note that my comments refer to non-GAAP figures. Details about non-GAAP financial measures, including GAAP to non-GAAP reconciliations and other disclosures, are included in the earnings release and investor presentation on our website. For Q1, 2026, revenue exceeded midpoint of guidance at $125.1 million. Recurring revenue, driven primarily by consumables and typically more stable than systems revenue, represented 60% of total revenue. No customer accounted for more than 10% of total sales during the quarter. Gross margin was 46.5%, above guidance, primarily reflecting a more favorable mix as recurring revenue exceeded our forecast. Operating expenses were higher than guidance at $55 million, reflecting our decision to scale resources to support the rapid increase in high-performance compute opportunities.

This included accelerated spending on design materials as well as incremental engineering and field support to fulfill production orders and complete new opportunity qualifications.

Net interest income after interest expense and a small foreign currency loss was approximately $2.1 million. The Q1 tax provision was lower than guidance at $4.8 million. Now moving to the balance sheet. Cash and investments increased approximately $5 million during Q1 to $489 million, and cash from operations was $10 million. No stock repurchases were completed during the quarter. Total debt is $305 million and includes $288 million from the Q4 2025 convertible debt offering. Capital expenditures were approximately $2 million, mainly for facility improvements and IT equipment. We're targeting total capital expenditures to be about 2% of revenue in 2026.

Looking ahead, we expect Q2 revenue to increase 15% sequentially and 34% year-over-year to approximately $144 million, plus or minus $7 million. The increase is driven by demand tied to the ramp and high-performance compute opportunities and continued recovery in automotive and industrial segments. We're increasing our full year 2026 revenue outlook for growth over last year of 20 to 25%. Q2 gross margin is projected to be approximately 44%. For the full year 2026, we project gross margin in the mid 40% range as we ramp our supply chain and production capacity to support the rapid business expansion in high performance computing customers. Operating expenses are expected to be lower than Q1 at about $53 million. We intend to continue investing in resources to capitalize on the growing list of HPC opportunities, and we expect quarterly operating expenses through the balance of the year to remain in the low $50 million range, consistent with our Q2 guidance.

Net interest income in Q2 after interest expense and foreign currency impacts is projected to be approximately $2 million at current interest rates. The Q2 tax provision is expected to be about $5.3 million, and diluted shares are projected to be approximately $52.6 million, including $4.2 million shares attributable to the convertible debt. And of that amount, 3.3 million shares will be fully offset by the capped call, but are required for U. S. GAAP diluted EPS calculations.

In summary, our operational focus for 2026 is to support R&D investments and production ramp needed to secure multiple design wins in the compute market, including AI data center infrastructure, HBM memory, and physical AI applications, while progressively increasing free cash flow generation. That concludes our prepared remarks, and now we'll open the call to questions. As a reminder, if you would like to ask a question at this time, please press star 11 on your telephone and wait for your name to be announced. To withdraw your question, please press star 11 again.

Yeah, so at least on your first point here, Brian, the quarter-over-quarter increase in HPC systems revenue was about $10 million. So it's just under half of our increase quarter-over-quarter. And that puts us then for HPC, at least systems revenue, in the first half of 2026 at roughly about $30 million.

I think that pretty much answers the second part of the question of what's left for the second half right there.

I can do that math. Thank you.

So I think we're going to hold that for the balance of the year. The system's revenue percentage will increase. Well, the system's revenue is going to increase faster than the recurring.

And so that is why we see, you know, the 46.5 gross margin in Q2 hitting a little bit of a headwind in the second half. And so we think we're going to end the year somewhere in the mid 40% gross margin.

No, the qualified 100 million is sort of this year's spend from these customers. Now, like I said, we're probably going to be getting a portion of that this year, not the entirety of it this year.

Yeah.

Yeah, they're not all equal size, Brian. I mean, we got, you know, kind of a $10 to $40 million spread, depending on the customer here on an annual basis, the way we see it. We expect to be getting some qualifications completed by early Q3. The question is, do we then have an opportunity to get orders and participate on demands still in 2026? Does our lead time support that as well or not? And so it's hard to call right now if it's going to end up hitting revenue in Q4, plus obviously revenue recognition as well. You've got to account for accounting rules, or if this is going to end up spilling more like early 2027 at this point.

That maybe will also, like you said, inform what the revenue could be this year versus what might have to be captured next year.

So we are operating at about a 14 weeks, I should say cycle time instead of saying lead time, on handlers right now, on our thermal handlers. I think a bit of the challenge is, you know, if you get a $30 million order, not all of it is going to ship in 14 weeks, as you can imagine. It's spread over several weeks, several months. And as we start layering additional customers, we are working hard here to open that manufacturing pipeline, both from a supply chain side, meeting regularly now with suppliers and understanding who are the choke points, particularly for our thermal heads. As well as internally, we are hiring resources in Malaysia. We're looking at relay out of the facility in Malaysia to open up more floor space.

So, you know, I can tell you 14-week cycle time, but lead time really largely depends on the size of the backlog we have in front of it.

As of maybe a quarter ago, a little bit more than a quarter ago, that seems to be spreading out more broadly here. As inference is starting to pick up and sort of the realization we need more computing power going along with the GPU power that being deployed That more of what I seeing It sort of that spread out of demand for different types of processors and network processes inclusive Okay.

Right. Right, of all sorts, all have high voltages, create a lot of heat. So all of these in-market customers that you hear about from the custom ASIC guys to the GPU guys to the CPU guys, All of them need some sort of temperature-controlled handling equipment for their processors, correct? That is correct.

And is that the market that you're referring to when you talk about the $750 million TAM? Is that kind of aggregating what most of these customers' thermally controlled temperature handler demand is? How do you come up with that $750 million? Yeah, by the way, we're not calling it necessarily a TAM. We're calling it more like a SAM, to be fair, because we have a pretty defined list of customer and customer device classes that we're telling up to 750 million. I think if we were to talk about a TAM, it's likely a bigger number, and we're not attempting to guess that, so we're not going there. We're being very targeted here to the list of 15 customers that we have tallied and customer applications that we have tallied up that comes up to the $750 million.

That's what it is. It's a very targeted list. We know what these customers have for buying pattern this year, and that's how we come up with that number. We also understand that some of these customers are ramping, so I guess the expectation is that SAM itself could be bigger next year.

But like I said, we're not trying to guess the TAM, the total available market. We're just guessing here from customer information what we see for their spending this year. Okay, and then final question for me is, could you just elaborate a little bit more on the silicon photonics and what exactly the application is you address there and how big a piece of business that could be, let's say, next year? I realize because we're just starting off now, but maybe just elaborate a little bit more on what you're seeing there. Thank you. Sure. That is really, I would call it a beachhead business at this point. We sold a number of interface, you know, we call it contactors, right, interface products here for silicon-photonic application at one of the large accounts.

You know, there's really two major drivers in the industry, I think, today and a few others. But these are interface products. So you're talking about sort of $10,000 or so contactors that we sold several of. We are working to provide solutions that include our handler with the contactors. But I'm not going to venture to guess what kind of revenue opportunity for 2027 that is at this point. It's not included, not really included in our $750 million at the moment.

That is correct.

That is correct. Thank you, Luis. You're welcome, Dave.

Thanks, Greg. Thanks for the question. I think we finished the qualification. We're very successful on the qualification of the Eclipse at one particular account that sort of looked like, okay, we could capture a bigger share of the revenue in 2026. So we qualified, I guess, in time to catch the next round of orders, and that just increased the size of the pipeline for this year. That's just simply that.

Can you help us with some color on where you're seeing that strength? Is there a propondence towards OSAP versus IDM? And do you expect to ship all those systems this year in any color on linearity would be helpful? Yeah. When we look at orders here, it's actually roughly, depending on the market segment you pick, it's about 30%, 40% increase year over year. There's one segment in particular that is driving, not surprisingly given what we're talking about here, it's computing, that it's up about 211% year over year. That's pretty much what's driving the business. Now, I do have to say the industrial segment is picking up a bit as well. That is also strong, came out pretty decently strong in the first quarter.

Yeah, so we see a ramp in Q3, and of course, some of that will fall into Q4 as well.

Got it. And then just going back to the point that the company is making on page seven of the deck where We've got the expanded AI computing pipeline with almost a half a billion in engagement and then 150 million to 200 million in qualification. Can you provide any color how quickly we can move some of that engagement activity into qualification and then through qualification? How much of that is really something that can convert in 2026 versus what you might have your eye on for 2027, guys?

I think at this point, Craig, it would be safe to say that we're working to complete the qualification of the about $200 million opportunity in 2026. As I mentioned earlier on a previous question, we'll see if we can get some of that revenue also in 2026, but largely 2027. On the balance here, the remaining $450 million, $500 million, those engagements are likely to move into QUAL later this year, beginning of 2027.

Is that right, Luis? That is right. That is right. And, you know, a qualification of these things take a good six months timeframe, and then from there, production ramp. I do have to point out a little bit here, too. Largely, the recurring portion of this is going to come out about a year after shipping systems, right? So you've got to remember, our system ship with about a year's warranty. Once that expires, you start getting the spares, the service. These devices typically have 18 months lifetime anyways. Thereafter, you start getting new kit orders. You start getting potentially new ThermoHead orders for upgrades.

It's high-performance computing, so those ThermoHeads are very specific to the application. Maybe you can use it across two generations, but the ThermoHeads themselves eventually you need to replace. So, you know, within a 12-month timeframe, we should start seeing the recurring revenue kicking in. And the recurring revenue, maybe it wasn't really clear on the slide here, is included on that $500 million bucket as well.

Okay. So, you've got a nice one, too, but with the second punch included in the chart. Yeah.

The software right now is very much going kind of hand-in-hand with our sort of test handlers and inspection systems, so basically the automation pieces. We have an element of software we call PACE Inspection, goes in with the inspection platforms. It helps optimize yield of the inspection systems. And then we've got a PACE prescriptive that goes along with both test handlers as well as inspection metrology systems that help optimize overall equipment efficiency, optimize maintenance, predictability, and output of the factory.

So if you think about that software base, we are now currently at an ARR, you know, annual recurring revenue here, of about $1.2 million. So this is what we have sort of in bookings for annual subscription of software. The attachment rate of that subscription, it's still pretty low. It's really about 1.3% of our systems have a software subscription attached to it. So a low number, so plenty of room to grow. So, but as I pointed out here in the script, the value of that software is pretty big because if you get it in, like we got here in the example given, you know, $20 million system order, $330,000 of software annual subscription, through the lifetime of that product, that's about a $5 million of recurring revenue we're going to collect through the lifetime of the product at a pretty high margin, right?

So it's still a small piece of the business. It is a growing piece of the business. It's growing fast. I think we're expecting it to be close to $3 million in revenue this year. That's more than 200% growth year over year. But it does carry a really nice lifetime value recurring component to it that adds to our overall recurring business. Thank you Thank you I will now turn the call over to management for the Q2 We are seeing that pick up right now. Another interesting data point here, the industrial utilization at the end of Q1 was 79%. So it's right there at that capacity by threshold of 80%. Industrial is doing well. It had a good increase in orders quarter over quarter, and about half of the revenue growth quarter over quarter going into Q2.

On the RF side to your question, we're also seeing a bit of a pickup on RF tester orders sales in the second quarter. There is typically a seasonality. That seasonality tends to be late year, like Q4 to early Q1 when RF picks up. It's a little late here. We're going into Q2 and seeing a bit of a pickup in RF. I can't completely explain that to you. Why? And then obviously there are technology transition points that are major drivers in RF, one coming up in the next 18 months or so associated with FR3 or what's commonly known as 6G.

Matt Hutton here. Yeah, so, I mean, we continue to look at opportunities, as you can imagine, from what Luis and Jeff highlighted. They're mostly opportunities in the reoccurring space, our growth areas. You know, we'll continue to be disciplined, look at buy versus build analysis, and look for opportunities. You know, unfortunately, a lot of the tailwinds that some of these companies are receiving, that we're receiving, they're also receiving. So, a lot of valuations remain elevated, but we'll continue to be disciplined and look at opportunities in our growth areas.

We certainly expect growth in 27. I mean, we have that in a qualification bucket there of 150 to 200 million that will add to 2027. Also, pretty encouraged with overall test utilization getting very close to that 80% mark. So all things being equal, yeah, growth in 27. At what rate? We haven't tried to pencil that in yet, so we're going to reserve another quarter or two before we talk about that.

Can you kind of break down that remainder, please? Yeah. Hey, let's back up a little bit. So initially we came out and we said HPC revenue in the 60 to $85 million range for 2026. What we're doing now is increasing that 60 to 85, we're increasing that to 80 to 100 million. Most of that relates to the Eclipse handler. The NEON for HBM inspection, We previously said it was 15 to 20. I think we're at the higher end now of that range.

And we're, you know, we are, as Luis had mentioned, in qualifications or fixed qualifications for our testers, also participating in some HPC revenue.

Does that help clarify?

Yeah, the way we see it now, Dennis, you know, we would expect Q3 to be pretty similar to Q2, somewhere in that 144, 145 range. Q4, you know, we could have some seasonality, so a slightly weaker Q4, maybe down single, single mid, single, single digit. It's a little bit, yeah, quarter over quarter.

Yeah, we were talking about inspection and metrology business here. We saw a big increase in orders in inspection of metrology in the first quarter. In fact, let's see here, I think it's up year over year 64%. We are expecting that business to hit about $70 million in revenue this year. And, you know, what's driving that? One is HBM, which we're now guiding to about $20 million in a year. And the other one is just further demand for inspection products from both a U. S. and a Korean customer with large orders in the Q1 thing.

That's helpful.

Yeah, I think that's a good observation, Vee. However, we are incurring some higher initial costs here to ramp the Eclipse supply chain and production. It's coming at us very quickly. It's a new configuration, and so we're having to spend more money, more costs, again, on supply chain and production. I expect those costs to carry through almost probably through this year. So, 2027, you know, we'll see lower costs, particularly for Eclipse. On top of that, I think similar or in line with other companies, right, there's a small impact from higher energy and freight costs, you know, to the tune of about 10 basis points.

On top of that, we are also seeing higher cost of memory ICs that we use on our products. It's not a large, huge number, but it's about 10 basis points.

Is that like the, you know, 200 bps of a decline that you have? Can you maybe characterize what's cost-driven, what's kind of mix-driven? Well, yeah, it's kind of a combination here. It is definitely cost-driven, as I mentioned, for the Eclipse platform in terms of supply chain and production. And then, you know, to a certain extent, that also relates to mix, right? But I'd say cost first, mix second.

Yeah, you bet. So I'm forecasting Q2 will be lower than Q1, but we're going to still be elevated from the model. So we're going to be about $53 million for Q2 operating expense. And that's because we are going to continue to invest in the resources to capitalize on these opportunities that we have in HPC. So I expect that sort of 53 or call it low, you know, low 50 million range to persist through the second half of this year. For OpEx. For OpEx. Yes, that's OpEx.

These are all Eclipse. These are all Eclipse thermal... I understand. ... thermal handler application to some form or another of a processor device.

Thank you very much. And before we sign off, I'd like to note that we'll be attending the following investor conferences during Q2. And those conferences are the TD Cowan conference on May 27th in New York City, Craig Hallam conference on May 28th in Minneapolis, the Stiefel conference on June 2nd in Boston, and the Evercore conference on June 3rd in San Francisco. If any of you plan on attending these conferences, please reach out to your conference contacts or let us know, and we will arrange for a one-on-one meeting.

So thank you for joining today's call. We look forward to speaking with you again very soon.