Aehr Test Systems

Q4 2026 Earnings Call — July 14, 2026

Analyst (Craig Hallam Capital Group): Thank you. Congratulations on the fantastic outlook. Gayn, I know you gave some description of revenue by application for the last fiscal year, but as we're thinking about the $130 to $150 billion, can you just give a rough idea of the assumption as far as product mix, you know, between AI chips, silicon photonics, power, and if there'll be any memory revenue?

Executive (Title): So we will. And the only challenge, of course, is, you know, we've never done that before. In reality, our ability to always predict, but in our current kind of roll up right now, it's pretty similar to last year in terms of AI, 70-ish percent, the silicon photonics, maybe 15, 20 percent, maybe in that same range. And then the power semiconductor and miscellaneous is kind of the rest. But there's some pretty good, and by the way, and I will state, we currently do not anticipate any memory revenue in even the 150 number. So upside any potential revenue from memory would be in addition to that, but right now we're not assuming any of that. And the bulk of that, if not the majority of all of that range of 30 to 150 is coming from current customers. So not really even including, I want to give myself a little leeway of this new benchmark customer that we just completed with amazing results. That would be upside to that as well. So I think last year I talked about the silicon carbide customers a year ago were all talking about how great it's going to be and it's going to be a turnaround year. They didn't really give us really detailed forecasts.

So I think I shared a year ago, I'm kind of in a wait-and-see mode. I'll just believe it when I see the orders. This year seems a little bit more specific. We have some people with real numbers, some real forecasts, some real targets. We're seeing the population of our installed base be at pretty well full capacity. So it does feel more real this year, and so I'm more optimistic. But we still have a pretty small number. Yeah. Yeah. A pretty small number. Actually, I think we might have just exceeded it. So that's probably what we want to do. Chris is looking at me like I'm sharing too much. So that order we just got, those orders we just got in, I think exceed what we were already planning. So we'll see how this goes. There's just a lot of opportunity. But AI is going to be a big chunk of our business this year, both in dollars and percentage.

Analyst: Just on the memory, it sounds like you feel more confident in the development program. Last quarter you did mention kind of roughly the same words that it could lead to orders in 27 and revenue in 28, but it sounded like you were more confident that you will get orders at some point in fiscal 27 for revenue in 28. Did I hear that right?

Executive (Title): I don't know if I'm more confident than last. I'm not sure I'm more confident than last quarter. I mean, I was moderately confident last quarter as well. We have more detail. We've actually started to spend money on it. So we've hired some people and redirected some people and we're doing architectural schematics and assessment. There's some software that started, you know, physical board layouts as part of the proposals to the customers. And, you know, but we really believe you need a committed customer to a project. You know, there's a sort of, whether it be a PO or a sponsorship, but you need to make sure you're aligned with what they want and what they're going to buy at the price point you're willing to sell them at. And that's the piece that we still need to get ourselves resolved to. There's been some dynamics during the last year, if you look in the last 12 months, what's kind of changed. The introduction of high bandwidth flash is a doozy, right? That has really put at least a couple of them, it's publicly known, on their heels in terms of what they were going to do, how they could test it, and it broke a lot of things.

The other piece is the new HBM, which is a DRAM. High Bandwidth Memory is DRAM. High Bandwidth Flash is, of course, Flash. But HBM, new standards that are coming out, and I don't want to get too carried away here, some of which driven by specific end users, have embedded BIST capabilities in it, which means that the testability changes dramatically. and we think could be in our favor. So it's given me some additional optimism with respect to our ability to maybe do HBM sooner than we were thinking, you know, certainly six months ago. But I'm going to just keep a optimistic attitude that we're going after this business. But I, you know, I would be hesitant for you to build models with a lot of revenue in that yet. Certainly not next year.

Analyst: Great, and then I'll just slip in one more if I could. There was a lot of different discussions regarding capacity, and I guess I couldn't write all the answers down quick enough to kind of do the math. But as you sit here today, you know, as far as top-line revenue capacity, what do you believe it is?

Executive (Title): Yeah, so we've talked about some different pinch points, okay? So people that have visited us know that we have done the, until recently, all of our systems and our consumable wafer packs came out of our Fremont facility. But the supply chain of our chambers, our blades, and all the subsystems and printed circuit boards are of course built by contract manufacturers. For the 150 employees or thereabouts that AIR has, we have probably another 250 people in direct contract manufacturers building stuff for us in low cost regions around the world. We have increased our capability in our facility, you know, maybe tenfold with the remodel and the other things that we've done but we still are only running one shift from an operational perspective. We don't have the people to chip tenfold but the capacity is here to be able to do that. People go well that's crazy why would you do that? Well as some of the deals we're talking about if a customer comes to us and said I want a hundred systems you have to be able to say I can do that and it won't take me three years to build it for you. In parallel, we talked about this in more detail last quarter.

We pulled the trigger last September with one of the contract manufacturers in Asia that does assemblies of one of our chamber suppliers. And they upgraded their facility to near clean room space in anticipation of some of the projects that we wanted to outsource to them. We then initiated and gave them orders and worked with them to build our Sonoma systems. And then the last quarter we shipped our first Sonoma systems directly to a customer out of that facility. Our team was there working with them, bringing them up, qualifying everything, demonstrating, then observing, then qualifying the tools. But those systems are actually being installed I think last week or this week at our large hyperscaler customer. and their entire backlog is expected to ship directly out of there. We think the capacity out of that is probably upwards of 20 Sonoma systems a month. And that would basically not impact any of our capacity here. So our capacity is significant. I mean, measured in hundreds of millions of dollars for sure. and we're doing some things to even potentially increase that.

Not that we would then ship a billion dollars a year, but what happens if we have to ship a couple hundred million and a quarter? Think of it that way. So one of the challenges in our business is that it's always going to be lumpy and we do not want to put money in place and infrastructure that's permanent because the business is cyclical and, you know, five, eight years from now, what's this thing look like? We don't want to have, you know, built up into a, you know, an enormous business. And if there's a softer year that we start losing money or not making as much profit. So I know that's a negative tone on this thing. But the discipline is how do you build up massive amounts of capacity without building up massive amounts of just fixed infrastructure?

Analyst: Got it. That's fantastically clear. Thanks. No other questions. Thanks, Christian.

Analyst (William Blair): Hey, thanks. Thanks for taking my question. And big congratulations to you, Gayn and Chris. I know how long it's been coming in terms of this inflection. So a few questions for you. I guess first one, maybe just piggyback off your last response, Gayn, to Christian there. Did I hear you correctly, 20 Sonoma systems per month? And if that is the case, that's about 100 million of Sonoma. Is that the right way to think about that in terms of the 140 midpoint of guide?

Executive (Title): So let me make sure I understand. So at our 140 midpoint, that's nowhere near 20 Sonoma systems a month. The production Sonoma systems that we have in backlog, we've shared this with different people, including the consumable elements are somewhere in the $600,000, $700,000 a piece. They're actually smaller configurations than times quantity. And I think that we've done a pretty good job of announcing orders along the way. So our package level burden Sonoma revenues are, let me make sure I do this. Thank you for having me. That's going to be one of the challenges. We're actually building them right now. We're going to start shipping them here during this quarter, and then the bulk of them actually get delivered in the second quarter.

Analyst: That is helpful. And then on the wafer level, you know, if we took that midpoint, what are the, are you mixing in, I'm trying to separate consumables from, you know, package and wafer level. Can you give some framework in terms of that breakdown?

Executive (Title): Well, let me answer that question, Jet. So, if you heard me, I mentioned about, you know, we're roughly in the 30% kind of distribution. Yeah, in revenue.

Analyst: And that's wafer packs and BIMs?

Executive (Title): Yeah. Oh, I think that's a good assumption. It's about 30%. That sounds right to me as well.

Analyst: Okay. That's super helpful. And then two more from me. Just along the lines, or maybe three more, sorry, but your largest processor, GPU, TPU, CPU customer, your AI customer I think is how you describe it, do you, in addition to the benchmark testing, I'm just curious, they're going with package level, but could you update on the activity or discussions around wafer level?

Executive (Title): It's a bit of a lookup table. So I think what you're talking about is our lead AI production package level burning customer who we've identified as one of the big hyperscalers. Okay. All right. They use Sonoma for all their qualifications. They had never done production burn-in before and they moved to production burn-in as we acquired NCAL. And that's been a great win for everybody because of our capabilities to support them and obviously to meet their capacity needs. That customer, the first device that ever went to production was on Sonoma. The second device is also are going to be on Sonoma. They already bought some of the systems for it. That particular device has been a little bit delayed in terms of its timing, but is still expected to ramp, you know, or at least start the ramp during our fiscal year at this point. That's the one I described that has twice the power, and they're expecting higher volumes off of it.

And then that same customer talked to us maybe six months ago the first time the third device they really want to consider wafer level burden and one way you might want to think about it is if there's a progression of number of compute chips in a single package as they get higher wafer level burden has a huge advantage because the devices fail during burden and if you fail one of the compute chips you throw away the other compute chips and all of the memory. And so at that point, it gets extremely expensive from a yield loss to continue to do package level burden. So that would be where you'd want to cut in at wafer level burden. So they've actually recently, again, this quarter, you know, again, they're just bringing up their second device, but the third one, they've been talking to us about, okay, what DFT do we need to do? What do we need to put in? How do we optimize it around the Fox system? So it's pretty encouraging.

Analyst: Got it. That's helpful. And just shifting gears, silicon photonics and co-packaged optics, I know you mentioned activities. The gallium nitride really struggled with reliability in the auto. Auto requirements are far lower than in data center by about 5x in terms of mean time failure. So I think that would be a good thing for you guys, but I'm just curious how you're thinking about market adoption, timing, and then role in GAN specifically. I know what you've got for silicon carbide. Just curious in terms of activity around GAN and the higher failure rates there.

Executive (Title): You actually mentioned silicon photonics. You mentioned silicon carbide and GAN. Let me just think through that window. Gayn has been interesting. If you remember a year and a half ago or so when we were engaged with the first Gayn customer, we didn't know if they were going to go to production, Vernon, or not. Since then, we've been able to verify they are, and they're going through a major qualification with a bunch of industrial, automotive, and infrastructure and data center applications. It's unlike some carbide that's like two or three big applications. Gayn, we've done like 12 different types of devices. So they're different voltages and power requirements that they go through all from solar to data center to electric switches in infrastructure and homes to automotive. They're all over the place. So they all have different kind of requirements, but so far as we can tell, every one of them needs a production burn-in screening to meet the reliability requirements. and it's been a learning process for us. So again, it's extremely difficult to actually test and burn in at wafer level.

We've learned, but it's now been debugged and fully working through and things are going really well for us right now. So we think that that will grow. Now the GAN market by the market forecasters is just dollar wise is measured in single to tens of billions of dollars in this decade. Whereas AI or Memory is 10x that. So they simply could never spend as much money as some of the other markets. But from a reliability perspective, it's very clear the value of burn-in. And we've been proving and validating the value of wafer-level burn-in of GaN similar to silicon carbide. And then silicon photonics, similarly, that's been the big debate is as the data centers really need to go to burn-in or to... Silicon Photonics, optical transceivers for fiber optic communication basically. The burn-in requirements are kind of very clear in front and center. So the hyperscalers, the big data center guys are all talking about this and it's sort of cool to be in Vernon, if you will, right now. That sounds kind of funny, a plan words. I've actually, Chris is laughing. I've never said that before, but it is. It's just, right now, it's a lot of people are talking about Vernon as kind of the place.

I believe that Vernon is by far the fastest growing segment in all the semiconductor tests right now, and wafer-level Vernon in particular.

Analyst: Awesome. Last question for me, I promise. Chris, can you just help with cadence in terms as you look at the year? How are you thinking? I mean, Gayn mentioned second quarter in terms of some concentration on the package level burn-in systems. Is there some way to think about, you know, this is a pretty big shift. So if you know, historically, we kind of said, hey, our first half is softer than the second half, but I would say this year is hard to say that because, you know, as Gay mentioned, we're going to ship a lot of the package level, you know, Sonoma systems in the second quarter. So second quarter is going to be a big quarter. So you could... First quarter is going to be pretty good. Yeah, pretty good. Maybe third quarter is flat, fourth quarter is up from there. So there's a high likelihood that the first half might be the same or even better than second half, but, you know, it... Second quarter is going to be a very strong quarter for us.

Analyst: Got it. I'll jump back in queue. Thanks, guys.

Analyst (Lake Street Capital Markets): Hey, guys. A few questions for me. One is a clarification question. So that April order you put out, I think it was a $41 million production follow-on order with a lead hyperscaler. I think that was a production order for the Sonoma Systems. Just to be clear, they aren't transferring over to wafer-level burden. This is a completely different customer, correct?

Executive (Title): No, that's the one we're talking about that is currently on package. Then the second part is package, and the third part is wafer-level or potentially being evaluated for wafer-level. It gets kind of interesting because they want to make sure we can also do the package for the third one. So we have to, you know, we kind of, I guess, we have two dogs in that race, and it's a two-dog race. So, yeah.

Analyst: Okay. But by the way, that third device we think is not even this fiscal year.

Analyst: And I guess my follow-up to that would be expect sort of a similar order size, I mean, on the wafer level side for them?

Executive (Title): Good question. If they do the same capacity, it would be higher. Our wafer level systems do in fact cost more per dot than the package level today. Margins are better too. but the value proposition is different because you get the yield advantage which more than makes up for the price of the machine. So just from a product positioning and a pricing, I mean we have a lot more IP in it, there's a lot more R&D people, we have enormous investments that we recover through the wafer level and we have all this IP that is worth it. So customers come to us not just to, you know, cost-effectively do burn-in, but at wafer level, it's because the device doesn't even exist in another form, or they put the device in with something else, and then if they burn it in and it fails, then the something else gets thrown away. And so the wafer-level burn-in value proposition has more to do with the yield improvement, which can be measured in, you know, 1% of all products if they have eight devices in a package, it's equivalent of 8% yield advantage. The cost of test might be 0.1% or 0.101%. So the price doesn't really come into effect.

Analyst: And then last one for me, and I don't know, you might have mentioned it, but that second major AI processor that just completed the benchmark testing, I mean, what's sort of the ramp up there? I mean, I mean, I know you kind of left it open-ended with the guidance range of being $130 to $150 million. Could you take it higher? I mean, could you see a major order come in in fiscal year 27 to be able to complete that?

Executive (Title): Yes. Pretty brief. Yeah, we can. We can. I mean we've as we've been adding capacity and you know we have multiple AI customers that are engaged with to some extent we can build to forecast you know we are having discussions about what if you throw a party and everybody comes right what does that look like and we're trying to meet at very high levels with the customers and potential customers and make sure we get an idea of what kind of capacity they look like. These systems have a lot of capacity, though. I mean, when you think about a tester that in the footprint of a competitive ATE machine, like a 93K from Advantest, I'm testing nine AI wafers in parallel with, say, a quarter of the 60 devices that are running. This is a lot of capacity that you put in place. So if we say, oh, I'm going to go ship 20 machines it's 180 blades, 180 wafers of capacity. 180 wafer test floor is a big test floor. And we could put 18, we put 20 of these and it's in the size of 20 parking stalls in the garage and out in your parking lot. Think of 20 cars out there that's more than enough to support 20 of our machines. Now think of 180 of our competitors in the same garage.

And these are state of the art clean rooms space. These are bunny suited test floors. And I always have to remember that when someone says, oh, you can ship, you know, five XPs a month, you know, if they're 18 blade machines, they're 100 testers a month of capacity. and like in silicon carbide, testing all 3,000 die in parallel in one insertion per wafer. These are very high volume, very capable machines. It's, you know, why we get several million dollars a piece for them, upwards of six, including wafer packs.

Analyst: Oh, we lost you, Max.

Analyst (Chlebina Capital): Hi, Gayn. I just want to quantify the capacity in Fremont. You mentioned one shift, two shifts. So on the Sonoma side, what I call Sonoma Max, I guess those are fully automated systems. What is the throughput capacity expected in Fremont? Is it still 20 Sonoma Maxes per month?

Executive (Title): So if you were to come and look at our facility, as many shareholders have, what we did in our big modification upgrade a couple years ago was add enormous amount of additional cooling and power capacity. That gives us drop locations for upwards of potentially 16 locations, physical locations for the equipment to be plugged in, another four mechanically, plus the test labs. What does that mean? That means that theoretically you could have 16 of these machines plugged in and powered up and going through different levels of test. Now, depending on the tool, they could take from several days to several weeks to be on that floor. What we've been doing in parallel is bringing up the CMs to be doing most of that work so that when it comes into the facility, maybe they only sit there for one week. So if you do that math, you're like, well, wait a minute, you could do 16 a week and four per week. Wow, that's a lot. What I've shared with people is, you know, we can reasonably see the ability to do, say, 20. and a combination of package or wafer level systems a month out of this facility within the supply chain infrastructure that we have.

But we have not initiated all of that. But in certain cases we have, like chambers, we've been talking to the CMs about auto aligners for both the package and also the wafer level. And so, you know, if those were, if those were, say, million dollar Sonoma Maxes, as you call them, we actually call them Ultras, it's $20 million a month of Sonomas. If you are 20 wafer-level burning systems a month, they have an ASP closer, with the wafer packs, of $5 million. They're $100 million a month. So people are like, oh my gosh, you have $1 billion worth of capacity. Conceptually, yes. For my employees that are listening, we need to hire more people and there's a lot more work that needs to be done. But with a running head start and a customer forecast, we could do enormous amounts relative to our current size just out of this facility. So actually the number would be 20 Sonoma Ultras per month or 20 XPs per month, not both, not 20 each. Yeah, okay, so you're right. You caught me on something. Except for we just initiated the 20 Sonomas outside of our facility. No, I know about that.

What it means, Larry, is we could run the 20 Ultras shipping directly from Southeast Asia and then still do 20 XPs here a month.

Analyst: And by the way, I need to put some disclaimers for all the attorneys listening. Some of these challenges would be particular material or getting access to. Some of the challenges that we've seen on the power supplies, for example, on Sonomas, is some of the power supply manufacturers that we use in Sonoma are supplying to NVIDIA and others. And they've come back and they've raised their prices 40%. if you want to get them. And so we've been buying components with long lead time items and things like that. So, you know, if we make it sound easy that we're overstating, it's actually quite difficult to do what we're doing right now, but so far we're able to keep ahead. And then I think almost everything in my backlog, we could ship before the customer has requested it. I'm thinking like all the Sonomas, Chris, you know, all my Silicon Photonics customers, So, so far, we've been on top of that. And our goal is to stay ahead of that. So, customers, if they want it, we could ship it as soon as they'd like it.

Analyst: Okay. Then on the paid evaluator, did you say that in your guide for the full year, you don't really have any revenues in there from this particular potential customer? Did you say that?

Executive (Title): Let's call it little to none, you know, at this point. You know, it's in the noise level. So that's a potential big upside, assuming that gets kicked off.

Analyst: And then lastly, in your written release, you mentioned on a memory side that you're working with multiple memory suppliers. Does that mean more than one HBM? I mean, how far are you engaged with the HBM potential customers?

Executive (Title): We've had specific conversations with suppliers... Okay, let me make sure I do this right. Two to three on flash and two on DRAM. Okay. And the primary flash... For those that are listening that aren't aware, there's literally like four memory manufacturers in the whole world that matter, so I apologize. And if you're not... For those that are in the fifth and sixth, they might be offended by that, but of the top ones, I'm listening from those.

Analyst: Okay, and then the primary flash, you engaged them over two years ago and yet they have any revenue or orders from them. Is that nothing going to happen there unless HBF gets kicked off or as far as you can see?

Executive (Title): I don't know. That was a huge change amongst some other organizational things that happened to them pretty dramatically about a year, a few months ago. So at this point, feedback recently has been, you know, they really need something for the HBF, but they really want something for their standard flash. HBF architecturally is way more power, more power supplies. It's technically a more expensive tester, but they really like the price point of the current test, you know, capability. And so we're kind of working through that. We got to find the math for that. And, you know, we need their help to sort of get the ball going because we obviously have a lot of opportunities ahead in all kinds of markets, and we just need them to sort of help us help them. But on a memory side, that would still be the first potential sale revenue for any of the four memory products.

Executive (Title): I think so, Larry. I would put some bet on a wild card. Because this new HBM has a logic interface for its testability, it's breaking the memory testers. And technically, when I describe a new memory tester, I am talking about, well, most people don't know or care on this call what a memory tester is, but a memory tester is built with algorithm pattern generations. It's sort of a different architecture and a different beast than a logic machine. So if you have an APG-based memory tester, it would be really good for flash or DRAM. But if you had a logic BIST engine in the device, then it may not even be very good at doing BIST. and so you would need something that looks like a logic tester, which candidly is more like what we build today with flash. The Fox XP was originally designed for the best engine of a couple of the big flash memory manufacturers.

Analyst: You mentioned that they were coming to you, these HBM potential customers. Would that be a function of potentially this paid evaluator telling them, hey, get over the air test and get your reliability better before you take out our accelerators? Is there any angle on that?

Executive (Title): That would be a little too bold for me right now. I don't think that's the case. I do think there is a general tone amongst and growing amongst the Data Center buyers, all the hyperscalers, you know, from Google and Meta, all the papers that you see, you know, Apple, certainly Tesla, of an increasing expectation of testing reliability to screen out defects so that doesn't show up on my data center floor. So that's more of a go get your act together in general than necessarily go to error. I will say that by contrast to the automotive guys, we specifically have automotive EV suppliers telling their suppliers go get air. I mean, that's true. So that's a different, we love that. But imagine one of the biggest, one of two biggest automotive suppliers in the world, buys our wafer packs, tells their suppliers, we'd like to qualify you for your silicon carbide reliability, and we're going to use the air chest box system with our wafer packs to validate your devices, and good luck to you. Pretty impactful, right? So I don't have that with the AI customers today. That would be nice, but we don't have that.

Executive (Title): Good job. Hey, I'll let you go. Thank you.

Analyst: Thank you.

Analyst: Thank you. I'm not seeing any other questions in queue

at this time.

I'll now turn the call back to management for closing remarks.

Executive (Title): All right. Well, thank you, everybody, for listening in and joining us. We're really excited to work hard for you guys this year. This is going to be a great year for Aira and hopefully for our customers and our shareholders as well. And we'll look forward to getting you guys an update as we go along. Take care.

Analyst: Thank you. This does conclude today's conference. You may disconnect your lines

at this time.

Thank you for your participation.

Q3 2026 Earnings Call — April 7, 2026

Mark Shooter (William Blair): Hey, Gain. Hey, Chris. You have Mark Shooter on here for Jed Door Timer. Congrats on all the progress, especially with the hyperscaler. I'm curious how you guys are looking at this internally, and what percentage of GPUs or ASICs or XPUs do you think are burnt in today? And how do you guys size the vector space?

Gain (Executive): That's a really good question, and I think we're still getting our arms around it a little bit here. I would say that we've been a little bit surprised at how many devices are not yet doing production burn-in. One of the things that we mentioned strategically when we purchased NCAL, what, 18 months ago or so, Intel does a type of burn-in, and they were known for it, called qualification reliability burn-in, which all processors go through, in fact, all semiconductors. It's what determines their lifetime reliability specs and that they will last long enough, et cetera. It's sort of a one-time deal you do with a large number of devices to do the statistics on it. Then certain devices go through a screening in production to weed out infant fatalities because the failure rate is higher than the market will bear.

InCal was doing this with a large number of AI customers, but actually prior to that wasn't doing any production burn-in. When we acquired them, we've now, because of the capacity we have in terms of people and infrastructure, we've been able to capture this large hyperscaler and are engaged with multiple others. But one of the things that I've been surprised at is that how many of the, I guess, particularly the ASIC suppliers don't do production burn-in yet or are talking about doing it. And that goes for a lot of different devices that are out there from edge, robotic, ASIC network processors, and even, you know, I've got to always be careful with GPU because everybody just associates GPU only with NVIDIA. But, you know, not all devices are burnt in still today.

And so there are certain ones that are, there are certain ones that aren't. And even within a company, they may have some of their products are burnt in and others aren't. However, the common theme is they're all moving to burn in. The data is out now that there's solutions out there like Sonoma or the wafer-level burn-in of our Fox system that can cost-effectively do it. And so now there's a very viable alternative to doing it at the system level or the rack level. We've said in the past that many of these guys would actually build it all the way to the rack. And then at the system integrator, they would burn it in for a week or two and weed out the infant mortality to ship it. Or in some cases with the ASIC suppliers, they just shipped it into their data centers and dealt with the fallout.

So it's growing. I'm trying to think if I'd try and put a percent, I think on ASICs, it might be, you know, by unit, like skew. I mean, I don't know if it's 20%. Maybe it's 5% of the number. So most ASICs are not printed. I would say on the AI accelerators that are out there across the wide variety, you know, maybe half. But then what's happening is the processors are getting higher power from generation to generation and breaking all the tools that are out there. So even the tools that were out there, and I'm not giving any inside information whatsoever, but just what's classically understood, like NVIDIA's processors of three couple generations ago compared to their current ones, their power is substantially more which would require new tools.

And the ones that they're working on and others in a year and out, and again, just what's publicly available, break the current tools. And so there's a continuous roadmap. And so even within our Sonoma platform, we're continuing to add capabilities. One of the key features we have is the ability to adapt it and add higher and higher current and power as you go forward. So there's, you know, how many times you hear a CEO say you're at the early innings, but this is still at the kind of the beginning phases of this. And over time, people will be buying a lot more burn-in systems as a percentage, meaning to cover the percentage of total, and then just ensure quantity.

Mark Shooter (William Blair): I appreciate all the color gain. That's very helpful. To zero in a bit on your hyperscaler customer, can you bring us a little into the room a bit here? And what was the decision process to go with package level, right? Not package part anymore. It's package level versus wafer level. And do you see a transition potentially with this customer to move to wafer level? And if you get a new customer... Do you think that they'll make the same decision, or is there a track towards wafer-level? Try to help us out with that.

Gain (Executive): Okay, so to be fair, two, three years ago, if you would have asked me, I've said this before, can you do wafer-level burn of AI processors? I think we would have said absolutely not. We didn't have the power in the system, and the belief was that there weren't the test modes that we now understand there are to be able to do it. And now as we've gone from customer to customer across a wide variety, there's commonalities about it that allow us to be able to confidently tell them we can do wafer-level burn-in.

So prior to that, it was whether you did package-level burn-in or not, or did it at, say, the rack level. So people first step is, do I do burn-in? Then they're going to default to thinking, I'm going to do it at the package level. But then what we're seeing, and I mentioned this before, we have customers, I don't want to get too carried away here, but the last two customers that we're in in the last two weeks, Alberto is our package level Vernon VP, and Vernon really runs kind of the wafer level side of things. The customer will come in and say, I want to talk about package level. And about halfway through the tour, they're like, what is that? We talk about wafer level. They're like, whoa, whoa, whoa, whoa. How do I do that?

And so we kind of joke about it around here. It's like, ah. But the reality is we don't care which side you go to. We have both. Specifically on the hyperscaler, and I've said this out loud before, the first device they ran with us, it's not their first device, but it's the first one they went to production on, is on Sonoma. Their second device, they just awarded us with production for that one and are planning the ramp of that with us right now. They're already on the roadmap talking about the third device, and they've asked us about the DFT to specifically put into the third device because they would like to consider that for wafer level on our Fox systems.

So I think that's sort of a progression that we will see, and I would actually imagine large customers that have multiple different product lines some they would do wafer level on and some they might do package level on. It becomes particularly valuable when you have a package that has multiple processors in it and all the HBM memory. So in those particular ones, I mean, the co-op substrate is more expensive than the silicon itself of the processor, which sounds crazy. So they would be very interested in doing the wafer level to screen out the dye before they have to throw away everything else. So I think there's a progression over time where people will move towards wafer level on the things they can, default to package level where they can't.

Christian Schwab (Craig Hallam): Hey, good afternoon. Thanks for a tremendous amount of detail regarding the different target markets and your success in each one of them. The most common question I receive is, is there a way to gauge over a multi-year time frame? Obviously, you gave guidance for this year in support of substantial growth the following year with bookings in hand and others to come. But have you had enough time to give some thought to the range of potential outcomes over a multi-year timeframe that you guys could do in combination of your target markets and potential entry into the market, memory market down the road?

Gain (Executive): The short answer is we have. The long answer is we're just really cautious about trying to get too carried away with our projections. But the numbers are very significant. If you just, because, you know, particularly now that there's some HS that hung down with memory kind of angle on this thing too, if you look at the dollar spend that people are going to do on, whether you call it compute or AI or, you know, if you look at the compute capability, right, that are going into training and inference in data centers, inference in edge, automotive, robotics, you know, the number of different applications and the way people are using it and deploying it.

The amount of silicon wafers is staggering. And, you know, why people talk about these enormous dollars. Those devices, a processor has always been burnt in. It feels like I'm contradicting what I said earlier. You know, it's widely known that Intel and AMD, the primary processor suppliers of the world, burnt in every one of their processors and always have. When the first GPUs were coming out, those were using graphics, they were not burnt in. And the initial people that are all related to AI are our foundries, and they're out looking for burning capability. There were no burning systems in the foundry OSAP models. And so people weren't spending on it. They spent enormous amounts of money on tests, and it's growing, and they're going to be spending a significant portion of their test budget on burn-in going forward.

I hear things, I mean, I hear it constantly from the customers rotating through. So, you know, the TAMs are, you know, multi-hundreds of millions of dollars for, you know, package-level burn-in, wafer-level burn-ins. If you say it displaces, package level is even higher. The average actual price per unit time of wafer level is actually more expensive than package level. But the yield pays for all of it. And so it's cheaper to the customer to spend more money. And so the TAMs are larger there. If you look at the memory side of things, if you look at the memory spend of the number of fabs that are coming out in the next, you know, five years, what percentage of budget is for their test budget, these are big numbers.

And so, you know, the spend is in Vernon is probably total spend measured in multiple billions of dollars per year in the next couple of years, you know, on an annual basis. And, you know, the question is, well, then wait a minute, how come you guys aren't, you know, $500 million? And the answer is, we think that we have a very good opportunity to significantly grow our package-level and wafer-level business across the biggest segments that are driving burden. And one of the reasons we're leading with putting infrastructure and capacity in place to be able to have the conversations we're having with these customers, they're throwing out some really big numbers.

And somebody in legal... going to warn me, you're getting carried away here, but it's an awesome place to be. And it's not only silicon carbide for EVs. Lots of people are wondering if the EVs are ever going to make it. As you guys know the history, it's like people got ahead of themselves. And I was even saying it. It's like, come on, you guys. We're not all going to be driving EVs. But the TAMs in these segments are significantly larger than anything we ever talked about on the power semiconductor side.

Max Michaelis (Lake Street Capital Markets): Hey, guys. Thanks for taking my questions. First, I want to start out here. We look at the demand environment from the package level and wafer level. So demand seems strong on both sides of the business here. But, I mean, to me it looks like wafer level is outpacing on the demand side and maybe the order side. Can you let me know if I'm wrong there? But anything else you can add as well?

Gain (Executive): The challenge with our business and for all of our shareholders is we know how to be lumpy. And, you know, by having more markets, and more customers, it can make it less lumpy. But the ASP of a production order, you know, a set, you know, in wafer level burn-in, you know, can be $10 to $20 million in an order, let's say, okay? Package level can be that big or bigger, too, okay? So when they come in, it looks like, oh, right now, we see demand on both significant.

Now, the engagement level and the work to get a wafer level burn-in is definitely harder than package level. And the obvious reason is, in many cases, we're already testing the part for the call on our tool. So now they have to just say, oh, I need to buy a whole bunch of them and add automation and go to production. Does that make sense? On wafer level, what we found is that there's a learning process by both sides a little bit, but to understand how they can use our tool to be able to test their part.

And in some cases, they're like, okay, I know if I just did this, it would make it a lot easier, but it's too late. I already taped out this part. That would be an example of this benchmark I'm in right now. It's like they're having to use some little fancier wafer pack to do it. And if they just did some specific DFT, they could use a very simple wafer pack, the same wafer pack we're using for like silicon photonics or silicon carbide and some of these others. Their vocabulary with us is, oh, I'll be able to do that for the next gen. But can you just work around it with the current one? Well, it's kind of harder.

The other one, as I mentioned, I want to get a little too carried away. I mean, I get pretty techie on these things. But, you know, we had a miscommunication on the clocks, which is something really simple, candidly. But if you do them wrong, it doesn't work. And so we've had to jury rig some stuff to actually get it to work, and we're going to spin it to make it work. Nobody's freaking out about it because this isn't rocket science, but it would be something we would never mess up again with that customer because we now both have the same vocabulary.

Second one is always easier. And so there's a little bit more startup thing with the wafer-level burn-in, but if you're technically astute and engaged and you look at it, you're not going, oh, this isn't going to work. You just go, okay, gosh, that's too bad. Okay, now let's keep going. And so there's a learning process. We're getting faster at it. And I think over time, wafer-level burn-in, like the silicon carbide or the silicon photonics customer we won this last quarter, it was just, yes. I mean, there was no on-wafer benchmark. It went from can you do it to how fast can you deliver. I think that is a natural progression. You'll see it in our package level, and you'll see it in our wafer level over time where customers will engage. They'll know we can do it, and they'll just think, let's go.

Larry Shlebina (Shlebina Capital): Hi, Dane. Your contract manufacturer that you're starting up, when does that start and when will it be fully capable of doing your 20 Sonomas a month?

Gain (Executive): They're in the process of building the first batch, I would say, is the best way of looking at it. It's a little more complicated than the way I described, but there's actually two contract manufacturers together, and then one feeds into the other one. The one that feeds into the other one did their prototypes. They sent to us. We were going through kind of an acceptance process to validate it to work out any kinks. Then those go to the other contract manufacturer for final system integration and shipping.

The other one is when we were out there, we visited them last September, I think. We did kind of an audit of facility power infrastructure and cleanliness, and they did a kind of a remodel similar to ours if people have seen it. You know, it's all white and fancy, clean floors, more clean room space so that we can actually build these things in a clean room area. They had facilities. They were doing some stuff for solar, as it turns out. And so we were able to leverage from that. And that is in place now. And we think first products would be ready to ship to customers this quarter, you know, through May.

And what we want to make sure is they're ready to go by late summer, when we see this Sonoma ramp hitting.

Larry Shlebina (Shlebina Capital): Right, that was the root of my question. Are you keeping any capacity? Are you planning on producing those systems in Fremont as well?

Gain (Executive): Oh, yes, for sure. This is in addition to, we've kind of talked about like about a 20 system per month capacity here from an infrastructure and footprint perspective. We'd actually still need to hire some more people, maybe take on a shift. But we'll use – we use that facility for, like, large volume orders of the same SKU, if you will, make it simple. And then we'll use – we'll continue to make Sonoma systems here, and all of the XPs will be built out of here, all the Fox products.

Larry Shlebina (Shlebina Capital): And then did I hear you say that your first expected XP sales to an HBM customer will be – this calendar year or this or next fiscal year, 27?

Gain (Executive): Yeah, I didn't quite say anything. I was a little more elusive than that on purpose. What I will tell you is that we have identified some interesting opportunities with HBM, probably the new 4E, that it has some interesting challenges that people would really like to do this wafer level burn-in on. And between our Fox system as it stands and the roadmap that we've been working on, as people know, with a team of people here for a memory extension to the Fox system to add what we would call channel modules into the Fox that make it memory focused, we think that there's some real overlap there.

That just, as you know, Larry, you follow this a lot, that's an uptick. I thought HBM had a past flash, and it is in parallel with flash now. I would say that would be an uptick. Yep, I agree. A little bit of an uptick. In order to be a good uptick, I'll agree with you. But right now, I'm excited about the discussions.

Larry Shlebina (Shlebina Capital): Yep. So the flash engagement, is that? Do you think that will bear fruit on the enterprise side here shortly before HPF gets underway, the effort that you're going to have on that front?

Gain (Executive): That's a good question. You know, I think it really is up to the customer. Kind of the timing of what we would build would be something that would be a superset that could do both. So, yes, if HPF were delayed a little bit, maybe we would intercept their standard products. They've asked us to build it. The definition discussion has been to do both. In some ways, HBF is easier. Because if you start saying it's all flash, a lot of times what happens is people say, well, I want to be able to test everything I've ever had before.

And then as the interfaces evolve, they tend to converge in voltages and speed or whatever. And if you say, well, I want legacy, it's like, well... okay, I've got to support this old voltage or something on a device you don't make anymore. So part of the challenge for us is to try and kind of converge on what do you really need going forward? You know, where are you going to spend the money? They'll probably never buy a system for legacy products from us in general. So I think that's one of the challenges we get to work through.

Larry Shlebina (Shlebina Capital): Well, that's all I have. Boy, you got a lot of irons on fire.

Gain (Executive): It is. It is so much fun, you guys. I'm telling you. Yeah, Vernon and Alberto and the R&D teams and the, you know, poor Nick and our wafer pack team is very busy right now. And we're doing some things to offload that, adding additional resources. We're hiring anybody looking for a great job with a company that's growing, you know, let us know. We've got a lot of recs out there and we're looking for great people.

Larry Shlebina (Shlebina Capital): It sounds like it is a lot of fun, and congratulations. I know you've been working at it for a good while to get to this point.

Gain (Executive): Thanks, Larry. Take care.

Management: Once again, if you have a question or a comment, please press star 1.

Management: All right, Operator, if there's no other questions, we'll end on that really happy note. And as always, if you guys have any questions, please feel free to reach out to us. If you happen to be in the Bay Area and want to try and stop by, we're always happy to, you know, give a short tour to key investors and things like that. And we look forward to a great quarter talking to you next quarter. I guess with our new fiscal year, now, our quarterly earnings will be the same time, the next time, and then there'll be, I guess, a one-month push out or something like that. But it should work out. This will be a good thing for our customers, which, honestly, that's the key to all of this.

Management: All right. Thank you very much, folks. Bye-bye.

Management: This concludes today's conference, and you may disconnect your lines

at this time.

Thank you for your participation.