Hi. Thanks to everybody for joining us. Our next speaker is Sanjeev Aggarwal. He is the President and CEO of Everspin. Sanjeev has been with Everspin, I think nearly 20 years, in a number of leadership roles, but the last 5 years as CEO. He is joined by Bill Cooper, Everspin's CFO. He joined Everspin, I think, a couple of years ago. He has a pretty long tech resume, most recently at AMD, before he joined Everspin. Welcome, gentlemen. It is great to have you. Sanjeev, I will turn the floor over to you, and maybe we can save Q&A for after the presentation. Thanks, and I will turn it to you.
Thank you, Rick, for the introduction, and thanks to Oppenheimer for this opportunity. I have a few prepared slides that we can go through and then open it up for Q&A. To begin with, I have just mentioned that any forward-looking statements that I make in this presentation, we are not held liable for that. Anyway, after that, what is Everspin? Everspin is actually a semiconductor manufacturing company that makes memories using magnetic materials. Because of that, it is actually radiation-hardened. It is actually very fast access, and it is actually very reliable in extreme environments, whether it is radiation or extreme temperatures. We have been in business since 2008. We actually spun out of Freescale in 2008, and we have been building these memories for mission-critical applications in data center, in industrial automation applications, in IoT, in automotive, and like I mentioned, in radiation-hardened application as well.
We have a fab over here, an 8-in factory that is housed in NXP. We recently made an announcement to increase our capacity. We actually signed a foundry services agreement with Microchip to duplicate our NXP fab that we have over here in Microchip, which is in Gresham, Oregon. We also have a joint development agreement with GlobalFoundries that was signed in 2014 timeframe. We use them as a foundry for our spin torque for our second-generation MRAM technology. At the end of this year, we are actually going to bring on TSMC as a foundry as well for spin-transfer torque MRAM. We will have a resiliency in our supply chain, both for Toggle and 8-in STT MRAM as well as 12-in STT MRAM with these two foundries across the world. Like Rick mentioned, we are about 20+ years in production.
I have been with the company that long. We have shipped about 200 million+ units since our inception. The good news over there is that we are very quality conscious, so we can actually count any returns, a handful of returns. It has been a very high quality technology and product that we have been shipping since 2008 timeframe. We have a diverse base of customers, 2,000+, and we are also very focused on our IP. We have about 700+ patents and applications worldwide that we have licensed to various customers, including GlobalFoundries, to allow them to ship embedded MRAM while they build a standalone MRAM for us. On this chart over here, it basically shows you the various locations that Everspin has satellite offices. Our headquarters are actually over here in Chandler, Arizona.
We do have a design center in Austin, Texas, and then we have worldwide sales and marketing people. We have some people here in Singapore, which is where we manufacture our STT MRAM with GlobalFoundries. Very top-level highlights. One thing I would like to highlight is we are the singular domestic provider of MRAM for mission-critical applications. There is no company in the U.S. other than us that can actually build Toggle MRAM or spin-transfer torque MRAM for the U.S. government for mission-critical applications. Like I mentioned in the previous slide, we have a diversified customer base. It goes all the way from IBM, Siemens, Schneider Electric, to mention a few, Mitsubishi Electric. Pretty diversified blue-chip customers across markets and applications. We have a large market opportunity, about $4.3 billion by 2029.
It's a large TAM that we can actually address, and we have the products and technology available to actually address that market, and we'll get into that in the next slide. We have a proven management team with extensive expertise from various segments of the industry. Bill Cooper, for example, he comes to us from AMD. We have people that come to us from Samsung and LSI, like Khaldoun Barakat. He's our VP of Operations. Then also our sales and marketing team that comes to us with experience from Intel, Marvell and such. We have a very strong financial position with zero debt. We have been cash flow positive from operations for the last few years, and we have about $43 million, $44 million in the bank, and Bill will get into that a little bit later in this presentation. This is what we bring to the table.
What is the value proposition of MRAM? For those of you that are technical, it's basically a CPU-attached memory, like an SRAM or DRAM that brings the nonvolatility of flash. It's a non-volatile memory, so it's persistent, so you don't need any batteries. You can turn the power off, come back after 10 years, and the information is still there. You can read and write to it unlimited number of times. It's basically very durable, strong endurance. Read/write cycles go all the way up to 1E12 to 1E16, so unlimited read/write cycles. In terms of performance, we are almost as fast as SRAM and DRAM. Some of the SRAMs go down to 5 ns, but we are in the order of 25 ns- 30 ns . Pretty comparable. Pretty fast read and write memory.
In terms of reliability, we have best-in-class robustness in extreme temperatures, in extreme environments, and we don't have a single failure because of any reliability issues. It's the only technology that actually delivers all these attributes, and therefore, I say it comes pretty close to universal memory. It's the only thing that does come close to that, although it is not a universal memory, but it's pretty close. With that, I'll hand it over to Bill to talk about our market opportunity.
Thank you, Sanjeev. What this chart represents on the left, you can see the classic market that Everspin has played in with its PERSYST products. The Toggle product is the originating product, then we moved into the STT product as well. You can see on the right, primarily those markets tend to be in industrial enterprise, casino gaming, transportation, those types of verticals. We are actually coming out with a new product, a UNISYST product, and that product will tape out later this year, and that is really geared toward going after the NOR flash market. As you can see in the chart on the right, that NOR flash market is significantly larger on the order of $3 billion-$4 billion. Everspin's product, the MRAM, is actually significantly faster. It's orders of magnitude faster than NOR flash.
It also is being much faster, but it also has better data retention properties as well. So we will bring out that new product. We expect to take that out later this year, and we'll bring it out next year. We'll start producing it in mid-next year, and we should start to see some revenues manifest themselves later in 2027. So pretty exciting for us in terms of bringing that product out to market. That product will actually be produced in conjunction with TSMC. So very pleased to see that. If we go to the next slide. One of the things that we've gone public about is Everspin has been in the $50 million-$55 million range of revenues for the last two years. But we expect to see that grow to $100 million or more by fiscal year 2029.
As you can see here, the mainstay products, the originating products, the PERSYST products, Toggle MRAM, STT MRAM, and that PERSYST family of products will form the base of that growth, and we have seen very good growth. We had the highest revenue quarter ever in Q2 of 2026 at $18.7 million, then we actually guided higher to $19.5 million-$20.5 million for Q3 of 2026. But as we look forward, we'd say we're on track to meet that $100 million target in 2029. Again, what you see is the PERSYST product, the UNISYST product that we're going to take out later this year and bring to market with production and revenue by later in 2027, that product will also contribute. One of the also very important areas is we do actually see licensing royalties as well as engineering service revenue.
That's under the NRE category. We just signed an agreement with a prime contractor to the U.S. government. We did that earlier this year in April of 2026. That's a $40 million agreement, and we'll recognize that over the next couple of years. But we expect that licensing and royalty and services revenue to be in the 10%-15% category. So even as we grow over time, we expect that to also grow with us. But at the end of the day, Everspin is a product company, and the PERSYST products as well as this UNISYST products form a really strong foundation. One thing that I just want to mention is the UNISYST family, the actual application of that particular product, is really geared toward the aerospace and the data center market.
We do have products in the PERSYST family that also play in those spaces, but the UNISYST part will open up new opportunities for us because it is going to be a higher density part, 256 Mb initially, and then we expect to scale that on up from 1 Gb to 2 Gb, and even higher. Finally, we want to show you some of the financial information here. As you can see, from fiscal year 2021 to 2023, we were in the pandemic area, saw a little bit of a drop in fiscal year 2024. On up to 2025, we saw growth there to $55 million in revenue plus $4 million in other income under a DoD contract. Starting to see that upturn, obviously, in 2026, we are on a path to do much better than that, much further north than the $60 million.
As you look at gross margins, again, very strong selling everything we could during the pandemic period 2021 to 2023. What we see here in 2024 and 2025 and what the company we have publicly committed to is we expect to see gross margins 50% or greater as we both today and as we go forward. On the right, the company has been very strong in spending off cash flow. We have seen some reduction in that. We have put in some additional capital into our Chandler facility, and then we have some capital requirements under the new deal that we signed, the foundry service agreement with Microchip. That capital, we feel like we are very well capitalized for all of the commitments that we have with $43.9 million, $44 million in the bank as of the end of Q2 2026, and no debt for the company.
Finally, one of the questions we get quite a bit is, "Hey, tell me a little bit about the breakdown of your revenues. How is it split by market?" On the left, we do see how it is primarily going to be, again, product revenues under the PERSYST family, both Toggle initially, but also the STT MRAM products, and then some of the licensing. Again, in that 10%-15% for licensing, but 85%-90% for products. Which is, again, even as we go forward, we expect to see that. For 2025, what you see is the mil-aero, low Earth orbit satellites market in the 10% range, enterprise in the 30% range. Industrial automation, that is sort of a classic arena for the company, classic market. Casino games, actually, also a very strong market for us.
Again, that has to do with the speed capability of the MRAM, as well as the persistence of MRAM. Really strong performance in that segment as well. As we go forward, we would expect to see a bit more enterprise and a bit more on the mil-aero and aerospace, as well as data center and enterprise. With that, I think we have covered all the slides. Rick, I do not know if we want to jump into Q&A.
Yeah, Bill, that sounds great. Thanks for that, and those are some good slides. It brought up a few questions for me, anyway. The first is, you mentioned the $100 million revenue target by 2029, and I think, Bill, you also mentioned you've kind of been range-bound the last few years, kind of between $50 million and $60 million, but clearly breaking out of that, right? Clearly accelerating with your guide. Right there, that's an $80 million run rate right there as we looked at calendar 3Q. I didn't know if you could spend just a couple more minutes talking about it, and you just mentioned some of those drivers. I don't know if it's mil-aero that stands out as a driver there, or if it's the enterprise stuff, or if you could just walk us through what some of those drivers are.
Then maybe the second part of that question, just trying to think about sustainability of that ramp, whether that's in terms of backlog, order patterns, revenue funnels. Basically, trying to get a sense of what your visibility is on the sustainability of this, what looks to be a real breakout or a real acceleration in the top line.
Yeah. Great point, happy to expand on that. We see very solid growth in the business. We actually saw 38% growth in our product revenue from Q2 of 2025 to Q2 of 2026. We saw that 38% growth. We saw 28% growth in Q1 2025 to Q1 2026 on the product side. We saw 9% growth from Q4 of 2025 into Q1 of 2026. Then from Q1 2026 to Q2 2026, we saw another 9%. So really strong, healthy growth just on the product side itself. Again, I can't particularly pick maybe one segment, but I would say, again, we see very good growth in the enterprise segment, in the industrial segment. Casino gaming continues to do well, but again, some of the growth drivers in the aerospace arena as well.
Then we did see a pretty significant jump in our non-product revenues from Q1 to Q2. That was driven by the $40 million subcontract agreement that we signed with a prime contractor for the MRAM process and engineering services. Again, really strong, healthy demand. What I would also say, and I've said this publicly, is from Q2 to Q3, we guided up to the midpoint of $20 million revenue in Q3. So you're right, times four, that gives you an $80 million revenue run rate. So well on our way toward the $100 million, for sure. But in our revenue at $18.7 million in Q2, so $18.7 million to call it roughly $20 million in Q3, most of that growth we do expect to be driven by product revenues. I just think that that's really a very important point to provide.
No, thanks for that. You highlighted UNISYST, and it seems to be one of the biggest opportunities that I saw on the slides out there, $3 billion TAM. I know you've talked about, or I believe you've talked about 5%-10% capture on that business. I don't know if you've given a timeframe on that. My impression was that UNISYST, you did put another chart up that showed UNISYST contributing to the $100 million by 2029, but it looked like a relatively small piece of that, and I could have misread the chart. I'm trying to get a sense of how excited should we get about UNISYST in the next couple of years? Is it more of a 2028, 2029, 2030 kind of a ramp? Or when do we see the 5%-10% penetration start to show up in your model?
Yeah. I'll start with this one, and then I'll let Sanjeev sort of supplement as well and add to it. What I would say is, yes, we are very excited about this particular product and this particular opportunity. Bringing out this 256-megabit part, we're going to take that out at the end of this year. That again expands into this much broader TAM for us with the NOR flash sort of competitive market. NOR flash tends to be at the lower densities, and MRAM can, when we take this part out, we'll be able to compete at the higher densities. Again, very sort of excellent applications in the aerospace and the data center Marketplace as well. We often get asked, "Hey, what are the milestones?" The milestones are, we're going to tape it out at the end of this year.
We'll start production mid-next year is when we expect, and then we'd look for it to start contributing some level of revenue maybe by the end of 2027. Really ramp to production volumes, and it should start to contribute meaningfully in 2028 and on into 2029. Again, I think we're well on our way to the $100 million, but I don't want to declare victory just yet. We feel we are very excited about that particular part and the contributions it'll make as we go forward.
Yeah. Rick, I think like you pointed out, UNISYST is actually going to be a large portion of our product revenue, and the question is when?
Right.
The challenge obviously is that in the industrial automation space, the typical design cycle is 12 to 18 months. So when we tape out our prototype at the end of this year, and we'll have some engineering samples available for our customers, let's say Q2 of next year, that's where you can start the clock of 12 months to 18 months. That's the reason why I think you'll start seeing meaningful revenue only in 2028. So I think we will hit this $100 million number with the majority of the contribution for PERSYST and the licensing revenue with a small contribution from UNISYST. But if you look at 2030 and 2031, I think we will start overtaking the PERSYST and the licensing portion of the revenue over here.
Got it. If I could pivot just a little bit, because you mentioned it a couple of times, but the $40 million, I believe, Sanjeev, it's like a two-and-a-half-year revenue capture on that government contract. So please correct.
That's right.
Oh, thank you. Yeah, please correct me if I'm wrong. I'm curious because I believe you're recognizing that. I think Bill might have mentioned that you're recognizing that as NRE. So I just wanted to kind of get a sense of what the margin implications are as that comes in, and then what does the slope of that capture look like as you ramp it? Are there certain milestones or hurdles that you have to meet along the way? Because it seemed like the implied guide for 3Q, and I could have misread that, was sort of flattish. Maybe it's not just a linear ramp for that two-and-a-half-year time period. I don't know what that looks like, that revenue ramp.
Yeah. I'll just make a comment and then turn it over to Bill for the specifics on the gross margins and the revenue ramp recognition. The bottom line is that this contract that we got awarded as a subcontract to one of the U.S. prime contractors is basically to provide engineering services to transfer the learning that we have on our Toggle MRAM to this Amentum subcontractor, under the prime contractor Amentum. By that, they will have a blue book, basically, that basically tells them how to build this technology in case they need it, and then they have all the know-how to be able to build this line, and they will be involved in our learnings that we have at Microchip that we will share with.
With that, I'll turn it over to Bill to actually talk about specifics of the NRE and the gross margins and.
Right.
Et cetera.
Right. Yeah. No, again, great question. What I would say is, won't necessarily be linear, but what I would say is it's probably a bit more like a bell curve that you would see. Yes, there certainly are milestones that are very objective that we have to track to along the way. But it's going to be based in engineering service efforts on the Toggle MRAM process. The final thing I would say is, yes, we do expect it to be or it is margin accretive to the overall corporate gross margins.
All right. Thanks. When you say bell curve, Bill, should we think the bulk of it is captured in calendar 2027? Would that be the right way to think of it?
Yeah. I think, again, it has to do with the efforts that we have to make and the milestones we have to hit. You'll see some decent. Obviously, it'll be more in 2027 than it was in 2026 because we didn't start.
Yep.
Essentially late April, May of 2026. You'll have a full year in 2027, so we'll see more revenue then.
Yeah.
Some of the more efforts, sort of again, as we go through time, and then on into 2028 as well. We'll recognize that it won't necessarily be a full year even in 2028. Yeah, 2027 will definitely be the biggest, and that's where you kind of get this bell curve as well, partly.
Got it. Thank you for that. Since we are talking about gross margin, just real quickly, I am curious. I know you have said you would get gross margin back above 50% and sustainably. I did not know if you had laid out longer term targets where you want gross margin to settle in your perfect world and maybe if you could include op margin target, to just kind of get a loose idea of where when this model is really humming, sort of where you see it.
Yeah. I would say, again, we would expect margins to be greater than 50%. I think as we ramp, we have different areas of the business that contribute in different ways. Again, the subcontract agreement is margin accretive against the 50%. I would say again, we would expect that gross margin. In Q2 of 2026, we had 53.9% gross margins on a GAAP basis.
Yep.
We are already kind of seeing some gross margins in that area. I think that is a reasonable gross margin for the company. Ultimately, we are going to be looking to push for ways to improve that, but there is a lot of moving parts, as you might imagine. Again, very comfortable with the 50%+ gross margins. We are kind of seeing it in that 54% range today. I think, again, anything in that arena, we would consider somewhat reasonable. As far as op profit margins, again, it depends on the mix of revenues and products and so forth. But yeah, we are definitely going to expect to be 10% or so as we move forward. Obviously, everything I say, I hope that we overachieve or we will work to overachieve.
But I think there are some areas where, again, we are making a lot of investments on the product development and roadmap, and sometimes you have to make those investments that your op profit might suffer or show less optimal at periods of time. But we are going to do what we need to do. Again, we have lots of exciting opportunities in the data center and aerospace markets. It is really quite fascinating.
No, absolutely. Thanks for that. I did not know if you could pivot to basically a couple of the fab projects. I am curious on the Microchip project, if there was anything you could. You mentioned it in your prepared remarks or in your presentation, but I just was curious if you would add any more color. Basically when, and if I missed it, I apologize, but when you expect to complete phase 1, move into phase 2 with the Microchip project and I know you briefly mentioned CapEx, just trying to get a handle on sort of how long, let us call it elevated CapEx, and maybe that is a mischaracterization, but how long we kind of run at these kind of CapEx levels.
Yeah. I think it is okay. I will just give you a quick answer, and Sanjeev can jump in. But we do see, under that particular arrangement, we would expect to see, we had something like $14.5 million of spend to sort of bring this up. This is, again, public information. It is filed as part of the 8-K. About a third of that is going to be NRE, and about 2/3 of that will be capital expenditures. If you think about $5 million a year for the next couple of years related to the Microchip ramp.
Okay.
It is probably not terribly far off, but again, some of it will be at different times, right? Because you got to get all the pieces and parts in place and spun up. Again, we feel the company is well-capitalized and certainly able to meet all of our commitments, including that particular commitment to bring up the interim process at Microchip.
Rick, on that one, phase 1 versus phase 2 and the time to bring first silicon from the Microchip fab. Basically, in phase 1, we are basically transferring or bringing up the Toggle MRAM technology, at the Microchip fab. For that, we are using the standard copper tools that Microchip has. Then there are some specialized tools that we use for our MRAM technology. We basically identified those tools. We are installing those tools at this time, and then we are basically starting to qualify the standard copper process at Microchip. I think all of that, the initial processing will probably be done in 9 to 12 months timeframe, and we will have a first qualified silicon over there from 18 to 24 months, one and a half to two years.
Somewhere in the middle of it, we will start bringing up the spin-transfer torque process as well. Sometime in Q2, Q3 of 2027 is when we can expect to start bringing up the 8-in spin torque process that we have over here in Chandler at the Microchip fab.
No. Got it. Thanks. Just quickly on Nokia, just with their purchase of the NXP fab there in Chandler, what if anything changes for you guys? Where does that relationship sit now with the new owners?
Yeah. We've traditionally had a very, very good relationship with Freescale as well as NXP, right? Even when NXP has been looking for buyers, they have been very upfront about talking to the buyers about our existence in the fab. Nokia was well aware of Everspin's operations in the NXP fab. Getting specific on the sale, the ownership doesn't transfer from NXP to Nokia till Q1 of 2029, so it's almost two years out. There is no disruption of our operations, at least through that time. We have talked to Nokia about continuing our operations over here. Our initial introductions have been, our conversations have gone very well. The hope is that we can continue our operations over here, of course, determined by the business opportunity that we have, right?
As long as our product revenue keeps going like it's going today, absolutely, we would like to keep our operations over here going. It looks pretty promising. I think Nokia is open to us being here, so.
That's great. Maybe I might just ask one more if that's okay, because I know you've got the litigation with Avalanche Technology. I know with litigation, it's always hard to comment too much, and we all understand that. But I didn't know how much you could share around that. I know in the first half, I think it was close to about $6 million in litigation costs, and I believe it's $4 million expected in 3Q. I think, and I don't want to put words in either one of your mouths, but I believe you kind of said that $4 million a quarter is sort of probably like a near term for the next few quarters run rate.
I was just trying to get a sense of sort of where that is and then kind of trying to get an expectation, maybe what litigation costs look like and as best you can, right, in calendar 2027.
I'll let Bill comment on the costs, but I'll comment on the litigation itself.
That would be great.
In January of this year, Avalanche brought a lawsuit on Everspin, claiming that we were infringing four of their patents for the magnetic materials that we are using in our products. Since then, we have gone through some review, and we believe that we have very strong defenses. Basically, we have prior art. Some of the claims actually do not apply to our products as well. We believe that we are in a very good position. Also some of the patents, they did not pay the appropriate fees for those patents in the patent office. Technically, they are not enforceable. That is something that we have highlighted at the ITC, and the patent office is actually working with Avalanche to sort that out. Overall, I think the defenses are going very well. I think we feel that we are in a good position.
The ITC case is very well-defined in terms of schedule. We will have a hearing at the end of November, early December, and that is when you start seeing the litigation costs going down. Then the decision is actually due in July or August of 2027. I think that is the silver lining to the cloud, that we will actually be done either way in the middle of 2027. Hopefully, we can get a conversation going between Avalanche and Everspin before then to minimize or avoid all these costs that are ongoing right now. Bill, I will turn it over to you for.
Yeah, I think just to add to that, I think certainly the litigation is expensive, right? It is challenging. It is complex, right? Because of the IP pieces involved. But we feel very strongly in our IP and in our position, and we are going to defend ourselves and we will continue to see it through.
Well, that is great. Well, once again, it was great seeing you guys and really appreciate you attending the show. It was, again, great to get the update for all of us. Thanks a lot.
Thank you for the opportunity, Rick .
Yeah. Appreciate it. Thank you.
Thank you.
Thanks.