Q1 2026 Earnings Call — May 5, 2026
Management: The ecosystem. Their test and development infrastructure helps us generate the kind of data that advanced node customers insist on seeing before engaging, and their endorsement will certainly help us engage a broader set of teams within each target account. Each of the large memory manufacturers are facing similar challenges to the gate all around customers as they develop their next generation transistors in DRAMs and high bandwidth memories. Our team is in discussions with them right now, and we are currently working on multiple solutions using MSE to assist in this area. Right now, memory manufacturers would do almost anything to get greater fab capacity, and they have the resources to evaluate different methods of doing so. We hope to take advantage of that opportunity with solutions enabled by MSD. The momentum we're seeing in the advanced node transistor space is the result of many years' work targeting current market trends.
The macro challenges that AI success has put front and center, capacity and performance of CPUs, GPUs, logic, and memory, the power demands of cloud providers, and the increased costs associated with these, are all areas that Adameric can help solve. For that reason, we believe that MST is a fundamental tool for the future of AI. Our customer pipeline remains very active across multiple domains. For example, our work with our large IDM customer continues to go well, and we expect additional results from wafer runs soon. Our efforts with STMicroelectronics are bearing fruit, and we are confident we will re-engage with them again in the near future, consistent with our view that MST can create value across multiple product lines, especially in a large, diversified IDM or foundry. In RFSOI, we are seeing strong results confirming our extensive TTAD simulations. The technical results we've been focused on, including for both PowerSwitch and LNA, have been confirmed through customer silicon runs.
The near-term question is less about performance and more about the most efficient path to commercialization, particularly in cases involving fabulous licensees where aligning the business structure with the manufacturing flow can be complex. In power devices, we are seeing excellent potential in new development work being done to target MST at both trench bed and HVT transistors, useful in high-frequency, high-speed, and high-voltage applications. At the same time, wafers continue moving forward with our second JDA partner, and we'll keep pushing those efforts toward production pathway. Turning to GAN, we made meaningful advancements this quarter, including a breakthrough that could give us a technical leadership in RF GAN on silicon to augment the advances previously outlined for power GAN on silicon. To explain the innovation, I need to give a little background. GAN on silicon is a much more economical growth method than alternatives built on exotic substrates like silicon carbide or sapphire.
But when GAN on silicon is manufactured, due to the GAN stack growth process, gallium and aluminum ions gather at the silicon substrate interface, forming an unwanted sheet charge layer called a parasitic channel which is well known to limit RF performance in GaN on silicon applications. In fact, its elimination has been the subject of materials and growth studies for more than 20 years. In the past few weeks, we received preliminary performance data suggesting MST can dramatically reduce the parasitic channel. It does this by using MST's fundamental interface engineering to block the gallium and aluminum ions from getting into the silicon substrate. An industry veteran told us that in his 20 years, this is the best measured sheet charge data he has ever seen. We're continuing to validate this very promising discovery with our test and measurement partners. RF GAN on silicon is of value in the wireless infrastructure, military, defense, and satellite markets. It's also being actively evaluated for high integrated RF front ends such as those for 6G cellular. So the market potential is large and growing fast.
We are actively engaging on both 200 millimeter and 300 millimeter wafer sizes in GaN depending on our customers' requests. That matters because the wafer size for GaN on silicon is one of its key advantages leading directly to a customer's path to high volume production, low cost structure, and a set of fabs that can support RAMP, including opening doors for new applications with conventional silicon fabrication methods and devices. We're seeing expanded interest in partnerships across the ecosystem, including engagements involving InSize, Synopsys, Texas State University, Sandia, and others. Those kinds of parallel paths, commercial customers plus research and ecosystem partners, can compress development cycles and accelerate the time from promising materials data to something customers can qualify and deploy. Work here is aimed at generating data that is both technically rigorous and directly translatable to customer device requirements. Finally, a quick note on our announcement last week about expanding our collaboration with Synopsys. We've worked with Synopsys for years to enable accurate modeling of MST inside the Centaurus keycat environment, through our MSD CAD tool set.
This expanded collaboration extends that relationship into GAN workflows for both high-value RF and power devices. Practically, this means we're working closely with Synopsys to provide feedback on their GAN models, and we'll be jointly developing marketing materials so customers and partners can evaluate the physical and electrical effects of MSD and GAN more quickly and with higher confidence. To summarize, we're making progress where it matters, expanding and deepening gate-all-around engagements, broadening GAN from power into RF with concrete technical innovations, and continuing to advance multiple customer programs across our pipeline. We remain focused on converting technical validation into commercial structures that can drive repeatable revenue and are confident in our ability to do so. This is indeed an exciting time for AtomEra. With that, I'll turn the call over to Frank, our CFO, to review our financials.
Frank (CFO): Thank you, Scott. At the close of the market today, we issued a press release announcing our results for the first quarter of 2026, and this slide shows our summary financials. Our gap net loss for the first quarter of 2026 was $6.1 million, or 17 cents per share, compared to a net loss of $5.2 million, which was also 17 cents a share in Q1 of 2025. On a non-GAAP basis, net loss last quarter was $4.9 million, or 14 cents a share. And our Q1 2025 net loss was $4.4 million, or 15 cents a share. GAAP operating expenses were $6.2 million in Q1 of 2026, which was an increase of $742,000 from $5.5 million of GAAP operating expense in Q1 2025. Stock compensation expense, which is excluded from non-GAAP results, increased by $397,000, primarily due to new hires and our adoption in Q1 of 2025 of performance stock units, or PSUs, for executives. PSUs vest over three years, whereas the time-based options and RSUs that we had previously granted to executives vested over four years. Although the vesting period is shorter, PSUs vest only if our stock performs well relative to the Russell 2000.
The first tranche of PSUs issued in Q1 2025 lapsed without vesting because we did not hit the required stock price performance threshold. With the exception of stock compensation expense, the drivers of GAAP and non-GAAP expenses are substantially the same. So I will drill down into other factors that impacted our expenses by focusing on non-GAAP numbers. Please refer to the slide presentation for reconciliation between GAAP and non-GAAP results. Non-GAAP operating expenses in the first quarter were $4.8 million, a year-over-year increase of $348,000 from $4.4 million in Q1 of 2025. Sales and marketing expense increased by $203,000, reflecting our two executive hires since October. R&D expenses increased by $127,000 from $2.8 million in Q1 of last year to $2.9 million in the first quarter of this year, primarily due to higher spending on outsourced engineering, to support the wafer runs for our gate-all-around engagements, our IDM customer, and our JDA customer, which drives spending on metrology. G&A expenses were basically flat from the first quarter of last year.
Turning to sequential quarterly results, first quarter 2026 non-GAAP net loss was $4.9 million, or 14 cents a share, compared to net loss of $3.3 million, or $0.10 a share in Q4 of 2025. Operating expenses were $4.8 million in Q1, which is a $1.6 million increase from $3.2 million in Q4. Let me offer some color on the magnitude of this sequential increase. As I explained on our last quarterly call, our compensation committee elected not to pay the full 2025 executive bonus, withholding approximately $669,000, which normally would have been paid out in January. The committee provided the executive team the opportunity to earn back the withheld amount in 2026 upon achievement of commercial objectives. This led to us reversing accrued bonus expense in the fourth quarter, which skews the comparison of expenses between Q1 and Q4. Our balance of cash, cash equivalents, and short-term investments on March 31, 2026 was $41.1 million, compared to $19.2 million on December 31, 2025. We used $4.6 million of cash in operating activities during Q1, compared to $3.2 million in Q4, and $4.8 million in Q1 of last year.
As is typical for us, cash use in the first quarter of every year is higher than other quarters due to payments for items that are expensed over the year. In February of this year, we closed on a $25 million registered direct stock offering, selling 5 million shares of common stock at $5 per share, netting us proceeds of $23.6 million after fees and expenses. Prior to this offering, we had also raised $3.2 million in Q1 by selling approximately 1.3 million shares under our ATM at an average price of $2.47. Currently, we have 38.7 million shares outstanding. With the proceeds of our equity offering, we feel that our current cash balance puts us in a strong position to execute on the opportunities ahead of us, but we will continue to be disciplined about controlling our costs. On our last call, I said that we expected our 2026 annual non-GAAP operating expense to be approximately $18.5 million, and we are holding to that number. To reiterate, the reason why the expense increase appears as large as it does over $15.9 million of OPEX in 2025 is the bonus deferral, which essentially shifted expenses out of Q4 and moved them into 2026.
Organic increases in spending mainly relate to the hiring of our VP of sales in Q4 last year and our VP of marketing in Q1. Revenue in Q1 was $11,000 and consisted of fees for wafer deliveries to the large IDM that Scott talked about, and we have $96,000 of deferred revenue on our balance sheet. Approximately $46,000 of revenue that we expected to recognize in Q1 pushed out to Q2 because wafer shipments that we anticipated making last quarter pushed out to early this quarter. Accordingly, we expect Q2 revenue to be in the range of $50,000 to $100,000. With that, I'll turn the call back over to Scott for a few summary remarks before we open the call up to questions.
Scott: Thanks, Frank. And before we take questions, I want to thank our employees, our customers, and our shareholders for their continued support. We're excited about the progress we're making, and we remain focused on translating our growing body of simulation and customer silicon evidence into commercial agreements that can drive long-term, repeatable revenue and a strong, sustainable business.
Richard (Analyst): Mike, thanks, Scott, and Frank. I just want to make sure the audio is okay. Scott, the gate-all-around stuff here, you made some very interesting comments. I want to touch on a few of these things here. So you mentioned that you now have measured silicon results here, and your customers have said that they're better than the other solutions that they have here. Just want to make sure that that's what you said, and then a couple of follow-ups on that topic.
Scott: Yeah, maybe – are you talking about GAN or gate all around?
Richard (Analyst): GATE all around.
Scott: On GATE all around, we do have measured silicon results, and we evaluated our results against another method that people in the industry are using to accomplish the same type of thing we're doing, and our results are a significant improvement. So, yes, we have definitely had that, and we're showing that to customers.
Richard (Analyst): To follow up on this, so I assume that the measured results are wafers run at one of these four targeted customers. Is that correct? Or is it independent?
Scott: The measured results are something that we did in conjunction with our strategic partner where they had gate-all-around structures, and we use those devices to grow MST on those gate-all-around structures in the wafer, and then we're able to conduct this testing.
Richard (Analyst): Okay, so now if you think about how we approach customers, we go out and we show customers our simulation data, which we can do without a strategic partner, but then having silicon-tested data is a massive improvement over that. So that's been able to really open the doors for us to get into the customers. The next step from there is the customer will typically say, okay, we can see you did that on your strategic partner's structure. Now we want you to do it on our structure because our structure is different. Everybody's is different. And when I mentioned that we have work underway with two of the target customers there, doing demonstrations, that's the step we're at. We're trying to implement our technology on their structures and show them that. We believe that the step after that, Richard, will be that they'll have to install MSD in their fabs to do any further testing because these structures are so small and hard to manufacture that it's difficult to do a lot more work by having us run demonstrations in our fabs.
Richard (Analyst): Okay, so to that point, do you have a commitment to attempt to do this on your customer structures or is just the discussions to get that agreed to?
Scott: We're working on it with two of them, actually. I don't know what you mean by commitment, but I guess they're sending us wafers and we're putting our stuff on it. So, yeah, that's pretty committed.
Richard (Analyst): Okay. That sounds pretty good. So what's the timeframe for this work to get done? And then I assume, given what I've heard for the many years that I've followed you guys, that the analysis of these can often take a while. These are more complex than most. So I would assume that analysis takes a while. So what's kind of the turnaround time between getting that done, analyzing, and getting to that next step? What do you foresee that taking?
Scott: It's going to take several months, just us doing the work. We have to really do a lot of development work to just figure out how to grow things effectively in these tiny devices that they're sending us. And so normally when someone sends us wafers within three weeks to a month, we can turn those around and send them back. In this case, my guess is it might take us longer than that, two to three months. And then when we send them back, they have to put them in their fab and run them for several months. So it could be in the order of six months before we start to see results coming out of this. Now, I mentioned a few times on the call, both structural analysis, which is where they are looking at what we did for deposition in those structures and making sure that what we did was appropriate. They can do that pretty quickly because you're taking TEM images, like electron microscope images, and looking at what we did. Those results will come quickly, but the electrical results will be the result of running the wafers through the whole line.
Richard (Analyst): Got it. Okay. And so you're expecting to run wafers with wafers from two different GAA customers then over the next few months then?
Scott: Yes.
Richard (Analyst): Okay. Going back to my first question here and understanding the results you measured with the runs you did with your equipment partner, they want to get a sense of whether the customers agree that the comparisons you've done with, I think, an industry standard approach to dope and diffusion, they actually agree with that as well, that that is much better than what they can get internally, or is this just what your equipment partner has concluded for you?
Scott: I think there's no doubt that the customers that we've been able to engage with and get down to lots of details on it, they agree. They have been impressed enough that they want to move forward with these further demonstrations. So, yeah, they definitely saw the benefit of using MST to block the dopant diffusion in the areas that we're talking about and how it works better than what.
Richard (Analyst): Okay, fair enough. Some really interesting stuff going on there. Thanks for all that detail, Scott, here. Maybe a couple other quick questions. So on the DRAM side, it sounds like we've made some progress here, but if I'm to compare that with the progress on the logic side to the memory side, it sounds like the logic is reasonably farther ahead than memory. Is that a fair comparison?
Scott: Yes, that's true. We are talking with the memory manufacturers, and one thing Memory is quite a different architecture than logic that we're using gate all around, but in memory they're having the same type of dope and diffusion problems with their newer architectures as the gate all around folks are, and our technology is directly applicable to that. So we have a lot of interest in, from the DRM guys about that, we're also talking to them about some other technologies solutions that may be able to help them in different ways. So it's lots of different vectors of how we're engaged with ERAM guys. I should say with the memory guys because it's also in high bandwidth memory, not just ERAM. But we're further ahead with the gate-all-around customers than we are with them.
Richard (Analyst): Okay. All right. Fair enough. Maybe a question on the GAN side here. So I think my recollection is you're talking more about applications of GAN into the power space, but more recently it's been in RF here. How would you characterize kind of the – which one is kind of the leader in terms of getting to the next step here and, you know, getting, you know, installation licenses? I know that's not the right term, but it's kind of what I think of it. Installation licenses or using the wafers with that already built in there, which one's kind of in the lead here if either one is notably better?
Scott: Okay, so it's kind of interesting where you're right saying that we initially targeted the power market for our GaN on silicon work. The power market is actually much larger than the GaN on RF market today, and that's one of the reasons why we targeted it first. And for the power market, our big value that we've been talking about is to improve crystal quality and, therefore, to allow people to manufacture on larger wafers because they'd be less bow and warp as they were growing the GAN and fewer defects and therefore would have a lot of inherent value. Now the only challenge with that is to validate all that work you actually have to build wafers and build electrical devices and do a lot of testing. So that takes some time. And everybody's GAN growth properties are different, so there's some tuning that has to happen. And so that takes time. The new things I just mentioned, GAN on RF, we got some test data, and we just spoke about it at a big compound semiconductor conference last week. And there is a huge amount of interest in the industry.
And just looking at this early data that we got, now it has to be validated and so forth, but just looking at that data could be enough for someone to adopt us because it's such a big breakthrough and such an area where the industry needs solutions. In RF, they don't actually have to do the full electrical testing before they can decide to move forward on something. So it could be that we're moving, although we're earlier into, again, on silicon for RF markets, that one could move faster.
Richard (Analyst): Okay. All right, fair enough. One last question for me. And maybe going back to STMicro here, and I'm not sure if this is who you're referring to, the IDM customer or not here, so maybe correct me if I'm misassuming that here. But maybe just kind of indicate where we're sitting here with those guys. And obviously, we'll put a pause on the power, stuff that you're hoping to move forward with that you talked about late last year, how about any other applications with them? Are they still moving as full force as you had expected and had been seeing since the cessation of the power work with them?
Scott: Yeah, just to clarify, when I talk about the IDM, it's not STMicro. STMicro is another IDM, and we think we have a lot of different areas where we can engage with STMicro, but that's a separate topic. So, yeah, we've been talking with multiple business units over there and been doing some work, some evaluation work, and we have recently got some results that lead us to believe that we're going to start re-engaging with them on developing a product. We aren't at the point where we can talk about that yet. ST hasn't specifically given us any okay to talk about it, but, yeah, we've been saying since we had to give that unfortunate news about the BCD program at ST that we were working with other groups and that our relationship with the company was great, and the thing is they really know and understand MSD technology and have seen it and they believe in it. So this is kind of an indication of those comments that we've been making, and I haven't been able to announce a new deal with them yet, but we hope to be able to do that in the future.
Richard (Analyst): Okay. Okay. Excellent. I will jump on the line, guys. Thank you very much.
Management: Thanks.
Management: There are a few questions that have been asked in the Q&A line, and I'll just bring them up one by one. So, the first kind of question is about GATE all around, and it's that given the evaluation periods that we've seen in other, are there specific milestones that need to be hit to convert these gate-all-around customers into JDA? And what's a realistic timeframe for such a conversion?
Scott: Yeah. At a high level, maybe I'll put a little bit more structure on what I showed. I talked about Richard before. It's typical of customers who want to see kind of four different levels. They want to see TCAD results that show that you have the potential to deliver performance, and they have to understand all the TCAD background and believe in it. Then they'll move ahead and say, we want to see that captured on silicon. So we've done those two steps and did all around. The next step, they say, okay, we want to see that captured in silicon, but on our silicon, on our structure. We're going to send you guys wafers. We want you to deposit it on our structure and send it back to us and we'll evaluate it. Now they know they're not going to get the most perfect performance out of that because, you know, this work we have to do together and tuning them up and getting everything to work fully integrated. But they're just trying to do a proof of concept on their platform, right? That's the stage we're at right now with two of the customers. Beyond that, the stage after that would be where they install and do the actual implementation on their device, tuning it all appropriately.
So, yeah, it's a fair question to say when should we expect to see a JDA. Sometime during in this period of us doing the evaluation on their devices and when we get to the point where we'll install there, because that would involve a license. Then we should be having a JDA in place. These companies do not move fast when you're talking about kind of legal agreements. But we're working hard to make those happen, and we hope to be able to announce them at some point in the near future.
Frank (CFO): Thank you. And, Frank, the question regarding the equity raise, an investor asked, he is curious about the background and reasons for the third-party private placement issue. And given the stock price rise, could we have had better timing?
Frank (CFO): Right. Yeah, thanks for that. You know, one of the comments I've made in talking about the capital that we raised in Q1 was some funding that we got via the ATMs. And if you look at that, the average price on that was $2.47, which is roughly about where we were trading about a week and a half or two before we did the equity raise. And so the $5 price that we executed on there, you know, given what we had seen so far, not only in Q1, but really looking back over the last couple of years, it made us look at this as a very good opportunity because sure, the stock had run up to $7 and now in the last couple of weeks, it's run up again. But, you know, given the past trading levels that we had, and again, a lot of geopolitical uncertainty in the middle of February, you know, which we've kind of seen play out since then. Of course, you can't know how the equity market's going to perform, but on balance, it seems like a very good opportunity for us to execute on that and then, frankly, be able to work, you know, toward commercial outcomes and not worry about the day-to-day movements in the stock price to have to use the ATM to keep our balance sheet strong. So we've now strengthened the balance sheet. It's always kind of easier with the benefit of hindsight to second-guess the price, but I think it was a very good decision to execute then.
Management: Okay. Thank you, Frank. Question on the tool partner. How has your relationship evolved with your tool partner, the strategic partner, and are they giving you more engineering personnel, and how has that relationship changed over time?
Scott: Yeah, that's a good question. We have been, you know, we try to be good partners with each of the big tool vendors. There's three main tool vendors that the industry uses for FE tools, and we typically want to be kind of an arm-stealer, work with whatever tool our customers want to work with, so we have good relationships with all of them. The tool vendor that we have the strategic partnership with, we've been working with for more than a decade and had a good relationship with, but now that we've entered into the strategic partnership, the level of co-development work that we're doing is at a whole new level. So we have weekly meetings with their engineering team where we are working on developing the test data that we need for marketing to customers. And as customers ask us questions and want to get more demos, then we dig in and do work on that together. So, yeah, on engineering, cooperation level, it's at a whole new level.
The second area is on the marketing and sales to customers, and that's something that we've never really done with them in the past, and that's where we would be, you know, developing the right materials for us to both go into target customers and talk about MST technology and what a good solution that is. Now, one thing I've calculated a number of times is that if we are successful licensing our technology to customers, in many cases, the tool vendor is going to make more money from us winning designs there than we will. So there's obvious advantages for them making us successful, and so they're not doing this out of the goodness of their heart. But the good news is I think they've recognized that in the last year since we started this, and we're really seeing the benefit as we're engaging with customers.
Management: Okay, and this is a follow-up kind of to the when would we get a gate-all-around engagement, but an investor commented that the last call sounded like 2026. We would see several deals being made. Is it safe to say that now that sounds unlikely, or is there still hope for inking an agreement this year?
Scott: We're only in the fifth month of the year, and I'm hopeful every month that we're going to be inking deals, so definitely would say there's definitely a very strong chance. And, you know, if you look at all the areas in which you are working, which of the segments do you think is closest to producing a royalty-bearing license?
Scott: So I spoke a call or two ago about wafer-based products, and I think that the development effort in a wafer-based product is relatively easier. So some of the areas where we're offering wafer-based solutions are in gallium nitride and in RFSOI. And we have wafer-based solutions that we're offering in the memory space. So I think one of those could be the fastest. But we also have been working on power and on RFSOI with customers for a very long time, so those could also be quick time to market. It's very hard to call with so many moving pieces.
Management: All right. And with that, Scott, I'll turn the call to you for closing comments here.
Scott: Okay. Well, I want to just thank you all for joining us to hear the progress being made within Atomera. I hope you're feeling the excitement that we are. Please continue to look for our news, articles, and blog posts, which are available along with investor alerts on our website, atomera.com. Should you have additional questions, please contact Mike Bishop. We'll be happy to follow up. Thanks again for your support, and we look forward to our next update call.
Management: Thank you. This concludes the call.