Everspin Technologies at Q3 Investor Summit: Expanding MRAM Horizons

On Tuesday, 16 September 2025, Everspin Technologies (NASDAQ:MRAM) presented at the Q3 Investor Summit Group Virtual Conference 2025, showcasing its strategic vision and financial health. The company highlighted its leadership in MRAM technology, targeting substantial market opportunities while maintaining a strong cash position. Despite ambitious goals, Everspin faces the challenge of doubling revenue in a competitive landscape.

Key Takeaways

• Everspin is targeting a $4.3 billion market opportunity by 2029, driven by its diverse MRAM product lines.
• The company aims to double its revenue from $50 million to $100 million in the next three years.
• A strong cash position of $45 million and zero debt underpin Everspin’s financial stability.
• Everspin is launching a 256-megabit X5 product to capture the $3.5 billion NOR flash replacement market.
• The company’s technology is integral to high-profile aerospace projects, including NASA missions.

Financial Results

• Revenue for the first half of 2025 was $26 million, with expectations to maintain over $50 million annually.
• Everspin plans to double revenue to $100 million within three years.
• The company maintains a cash reserve of $45 million as of June 30, 2025, with no debt obligations.
• Gross margins fluctuate between the upper 40% and upper 50%, influenced by product mix.

Operational Updates

• Everspin’s MRAM technology supports mission-critical applications in sectors such as data centers, IoT, and aerospace.
• The product lineup includes PERSYST, Genesis, and AgILYST, with varying production timelines.
• Manufacturing is supported by an 8-inch line in Chandler and partnerships with GlobalFoundries and TSMC.
• Over 115 million units have been shipped to more than 2,000 customers globally, with revenue distributed across Asia, Europe, and North America.

Future Outlook

• Everspin is eyeing a $4.3 billion total addressable market by 2029.
• The upcoming 256-megabit X5 product aims to replace NOR flash, offering faster writes and lower power consumption.
• Significant growth is anticipated in the satellite market, with collaborations involving AstroDigital and Blue Origin.
• Continued development of the PERSYST, Genesis, and AgILYST lines is a priority.

Q&A Highlights

• The new 256-megabit X5 product will enhance write speeds and read-write cycles, reducing power consumption.
• Product sales account for 85% to 90% of Everspin’s revenue, reflecting a diverse geographical market presence.
• Key markets include industrial, gaming, medical, avionics, aerospace, and data centers.
• Everspin has progressively scaled its memory densities, reaching up to 1 gigabit.

Everspin Technologies remains committed to revolutionizing the memory market with its MRAM innovations. For more detailed insights, refer to the full conference call transcript below.

Full transcript - Q3 Investor Summit Group Virtual Conference 2025:

Operator: Welcome to Q3 Investor Summit Virtual. We appreciate your participation in today’s virtual event. Up next, we are pleased to introduce Everspin Technologies. If you would like to ask a question during the webcast, you may drop them in the chat box button on the left side of your screen. Please type your question into the box and click "Send" to submit it. At this time, it is my pleasure to hand over the session to Sanjeev Aggarwal, President and CEO, and Bill Cooper, CFO at Everspin Technologies, who will lead the presentation. Gentlemen, the floor is yours. Bill, Sanjeev, you can go ahead with your presentations.

Sanjeev Aggarwal, President and CEO, Everspin Technologies: Hello, everyone. This is Sanjeev Aggarwal and Bill Cooper from Everspin Technologies. Before I get into the details of Everspin, I wanted to share this that any forwarding statements that I make during this presentation, you know, we assume no obligation for those statements. To give you an overview of Everspin, we are the leading provider of embedded technology and products for mission-critical applications. I just wanted to explain what we mean by being a provider. By that, we mean basically we are an end-to-end supplier. We do the design, we do the R&D, we do the manufacturing, and then we also do the testing of the device. We do outsource to our OSATs for packaging and the final test. Other than that, we do the entire design to manufacturing, either in our own fab in Chandler or with our foundry partner at GlobalFoundries.

These parts are used in mission-critical applications, in the data center, in industrial automation, in IoT, on the edge, automotive, as well as radiation hardware applications. Our memory is actually radiation immune. Because of that, we are designed into various U.S. government devices. I’ll get into that in a couple of slides. We have been in production for 15-plus years. We actually spun out of FreeScale in 2008, and we were already in production in 2006. We have shipped over 115 million-plus units to over 2,000-plus customers worldwide. To give you an idea, our revenue is coming from worldwide, which is shown on the graph on the right over here. We have offices in Asia, Europe, Middle East, as well as North America. Our revenue is equally distributed one-third each in all of those geographical areas.

We have about 650-plus patents and applications worldwide that we have licensed to various customers, including GlobalFoundries. I will come to that detail in a few slides. Why Everspin? To give you an idea, Everspin is the singular domestic provider of MRAM. We are in mission-critical applications. What we do is we accept radiation hardened CMOS wafers from customers like Honeywell or QuickLogic. Then we put our MRAM on it. Because MRAM is radiation immune, these final devices can then be deployed for deep space. A good example is we are actually designed into the Perseverance that is actually on Mars today through Honeywell. We are actually accurately collecting data from these MRAM devices. We are on our way to Jupiter on the Lucy mission with NASA. Other than that, we are also in various aerospace and defense applications that I will cover in a few slides.

I talked about how we are a company that is serving 2,000-plus diversified chip customers. I’ll get to that in a few minutes. We have some marquee customers like IBM, Siemens, Mitsubishi, and as such. We have a large market opportunity that exceeds $4.3 billion by 2029. I’ll explain how we get to that in a couple of slides as well. We have a proven management team with people from Advanced Micro Devices, our CFO, Bill Cooper, from an operations point of view, Kazum, that brings experience from Samsung, Intel, and similarly, our sales team, as well as an R&D team with experience from Motorola, for example. We have a strong financial position, a strong balance sheet with zero debt. We continue to expand our operating margins with positive free cash flow. To give you a little bit of an idea of what MRAM is, right?

The first thing and the foremost there is persistence. What that means is you can turn the power off. There is no battery required in your device. This information is basically stored for the amount of period that is designed in the spec. For example, for the automotive customers, it’s storing the data between minus 40 and 125 degrees C for 10 to 20 years. On the bottom left, performance. We basically are as fast as a DRAM or an SRAM. It basically means 25 to 35 nanoseconds read and write. Very, very fast speeds. That is the reason why they are used for data logging applications. I will cover that in a couple of slides. Endurance. We basically have unlimited read and write cycles through our memory. Last but not the least is the reliability. It’s best in class in harsh environments.

What that means is automotive, as well as industrial applications, as well as deep space. Based on that, we’ve actually created three different categories of products. The first one is Persist on the extreme left. What that means is if you are looking for an application where data logging or fast write speeds are of importance, or fast read and write speeds, and extreme temperatures, and unlimited read-write cycles, that is the Persist memory over there. What you use it for, for example, industrial automation, casino gaming. I’ll give you an example of casino gaming, for example. The idea over there is in the casino boards, on those slot machines, they are actually using our MRAM memory. Before they were using our memory, they were using a flash memory.

If every action has to be recorded two or three times in two or three locations, and the write speed is basically on the order of microseconds, you can only have two or three actions per minute. With our MRAM, unfortunately for the gamblers, you can actually have 10 to 15 actions per minute. That’s how the fast data logging actually shows up in the application. Another one is industrial automation. Basically, we control the robots on the manufacturing floor. If there was a power outage, with our memory, you have instant on and instant off. With a traditional memory like NOR flash or SRAM, you basically lose all the work in progress. That’s where we bring value to our customers. Organizations like Siemens and Mitsubishi are taking advantage of that application. The second one is Genesis, where we basically combine code and data memory.

This is where NOR flash, when it stops our configuration memory on NOR flash that stops scaling at 49 millimeter CMOS, that’s where we bring in a roadmap to extend the NOR flash using our MRAM technology. Identically, NOR flash is not available above 256 megabit or 512 megabit. That’s where we come in. Our first part at 256 megabit will be in production 2026, Q2 of 2026. You will have the tape out, and by the end of the year, we will have some samples. We can go into volume production in 2027. That’s the $3.5 billion market, total addressable market that I was talking about. We are the only players in that market today. Even 5% to 10% of that total time would be a huge inflection point in revenue for Everspin. The last one over here is AgILYST.

The idea over there is you’re looking for even faster SRAM-like performance, right? 5 to 10 nanoseconds instead of the 25 nanoseconds that you’re talking about for AI applications on the edge. Today, people are using SRAM, but it has very leaky transistors. Then you’re also using NOR flash for configuration. You can replace both those memories with an MRAM on the edge. You can basically do all the compute like SRAM on the edge, and then you can have an inference right on the edge with the part of the MRAM memory being dedicated for FPGA computation or for, yeah, basically for configuration memory. That’s where we are targeting for AgILYST. PERSYST is in production today. Genesis, we go into production next year. AgILYST is targeted for production probably a couple of years or maybe three years down the road. Where are these memories used?

That’s what you can see over here. Plotted on the y-axis is the number of read-write cycles or endurance. The x-axis is the write latency. You can see PERSYST, you are at 35 to 45 or 50 nanoseconds with SRAM and DRAM-like performance. We have STT-MRAM as well as Toggle MRAM in production today. We also have PERSYST Data Center STT-MRAM in production today. That’s where we are providing these 1-gigabit parts to IBM for the FlashCore arrays. Genesis, like I mentioned, is actually going into production next year. That will basically be providing performance that is much better than NOR flash, which is shown at about 1E-5 cycles on the bottom right over here. With our MRAM that we compete with, we have much better latency as well as much better endurance performance. That’s the one that’s going to address the $3.5 billion market in a couple of years.

Genesis is the equivalent for that product that we’re using. Let’s look at the total addressable market or the TAM for these technologies for these opportunities. You’re looking at, in the dark gray over here, a $320 million market for PERSYST Toggle as of last year. Then you have the PERSYST X5 that we actually brought to production over the last two years. That’s another $300 million market. The data center market is about $400 million. A billion-dollar market today is actually going to grow to $4.5 billion by 2029. What you’re seeing over there is the PERSYST Toggle market growing up decently by $370 million, $2.37 billion. The X5 market is growing from $330 million to $440 million. The PERSYST DDR4 is going from $400 million to $600 million. You can see that we add the Genesis enhanced NOR or the NOR flash replacement of $2.5 billion over here.

The Genesis MPDDR4 is on the order of $600 million. AgILYST is on the order of $50 million or so by 2029. All of these add up to the total addressable market being about $4.3 billion. What does that mean for Everspin? For Everspin, I think with the $1.1 billion that we have over here, we are at about $55 to $60 million. We expect this to grow by 2029 to a $100 million revenue for Everspin, capturing all these markets that we are showing in this circle over here. Next, I’m talking about a few examples. One is Avionics and Transport. This is taking advantage of the fast read and write that I was talking about. Think about recording the health of the battery in EVs as you charge, as the batteries charge and discharge as you’re accelerating or decelerating the vehicle.

Because of the 35 nanoseconds read and write, you can actually learn how the battery is charging and discharging and then improve the battery performance and the performance of the vehicle as you’re driving it. Similarly, for medical applications, think about dialysis machines or heart monitors or the pacemakers. You want to understand how the heart is performing at as frequent intervals as possible. That’s the 25 to 35 nanosecond data logging that comes in. Similarly, the real-time monitoring with the sensor logging. Any data that you’re collecting with the automotive vehicles, you want to record it as quickly as possible the next you get in place. That’s what you have with the automotive applications. We have about 5% to 10% revenue coming in from our automotive customers as well.

Casino gaming, this is what I was talking about, you know, on the slot machines on the casino floors, where you can basically improve the performance of these machines and improve the experience for the customers as well. Aerospace and data and defense, I’ll talk about this in the next couple of slides. The idea is because we’re radiation immune, we can actually be deployed in space applications. As far as industrial controls, this is what I was talking about with the programmable logic controllers that actually control the robots on the manufacturing floor. The robots will keep talking to the PLCs on a regular basis so that in case there is a power loss, the robots can tell the PLC where they were. With our memory, you can do instant on and instant off. You don’t have any scrap on the manufacturing floor.

Similarly, for electrical and power grids and the battery charging units that I was talking about a couple of slides ago. This talks about how we are relevant in mission-critical applications. Taking off a Everspin MRAM is a little bit clear on the words for aviation applications or flight control systems. On the right over here, we are designing into the engine cockpit in the flight control systems, also with the personal flight systems with EVTOL, as well as the flight control systems over there. I talked about the NASA Jupiter mission, as well as the Perseverance on Mars. The automotive, the engine control units.

If you think about the racing bikes, if you understand how our racing cars, if you understand how the engine is performing based on the fast data logging that we can do with MRAM, you can then provide more battery or more power at the appropriate time to the engine. That’s how we’re designing with the BMW racing kit for the bikes. A good opportunity over here is the NEO satellites. You might be aware that there is talk about 70,000 satellites over the next five years. We have a couple of design rooms over there with AstroDigital with the flight control systems and also with Blue Origin that we can talk about. The idea over there is they’re taking advantage of the fast data logging, the reliability of the memory, and the radiation immunity. We’re designing about under four to eight units per satellite or flight control system.

As we grow over the next five years, we expect to capture a large portion of this market, the 70,000 satellites over the five years, and expect a huge growth in the industry over there. In the FPGA programming, again, looking for faster over-the-air updates if you’re using NOR flash. That’s at three orders of magnitude slower than MRAM, two orders of magnitude more power than MRAM. For example, in 2024, when all the Teslas were parked on the side of the road, waiting for the NOR flash to be updated over-the-air updates. If they were using MRAM, we could have it done in seconds or minutes versus 30 minutes that you were parked on the side of the road. Those are the design rooms that we’re working on. We have some design in with Lattice FPGAs. We have some designs with the battery management systems.

We expect to grow this as well over the next few years. This is what I was talking about in the last couple of slides. You can see that we are actually designed in with the Mars rover, in the Perseverance on the way to NASA’s Lucy mission. Also in the EVs, we are actually designed in with the Lucid car as well as SUV. We also designed in with Bugatti that we talked about and also the BMW race motorcycles that we were talking about a couple of slides ago. Any mission-critical application where you care about the data in flight or where you care about understanding the health of the battery, I think that’s where you’re looking for using MRAM or standard memory like DRAM, SRAM, or NOR flash. This is a snapshot of the various customers that we have in the different segments in enterprise.

Like I talked about, we designed in with IBM, with Broadcom, with Dell. In industrial automation, Siemens, Schneider that I mentioned, Mitsubishi, Omron. For medical, I mentioned Canon and GE Healthcare. This is the dialysis systems or the heart pacemakers. Network and infrastructure, you know, Supermicro and HP, Cisco. The idea over there is any general memory that you have, that’s where you would be using Everspin’s MRAM because it’s faster to write and storing data in flight. Casino gaming we talked about. We also talked about the aerospace and defense. Typical examples include Honeywell, Airbus Frontgrade, and the BAE Systems. For the defense applications, Northrop Grumman and also Bombardier. A wide range of applications. If you’re looking at industrial automation, high volume, casino gaming, high volume, aerospace and defense is low volume, high margin. That’s the way you think about these various segments that we talked about over here.

A little bit about our capabilities. Like I was saying earlier on, we do the entire end-to-end from design to manufacturing. We have done various interfaces from XPI to DDR to random access and also some custom designs for our customers like GlobalFoundries Embedded MRAM that they provide today or Honeywell or Frontgrade’s custom STT-MRAM as well. We do have, like I mentioned, an 8-inch manufacturing line here in Chandler. That’s been in production since 2006. The idea is we can actually do all our R&D and contain all the innovation that we have over here in Chandler. We develop this technology and we transfer it to our foundry partner, GlobalFoundries in this case, where we’ve been in production since 2018. We are bringing on another foundry now, TSMC, to bring in our next 256-megabit product on TSMC’s 16-nanometer filter as well.

I talked about the IP development that we have. This is a non-exhaustive list of some of the licensing that we’ve done. For our—sorry, there’s a typo over here. This needs to be sensors. We have licensed our technology to AlphaVault for our TMR magnetic sensor technology. Also, with Frontgrade Technologies and Honeywell for deep space applications and aerospace and defense. GlobalFoundries for their embedded MRAM application. In enterprise, for Seagate and TDK for their hard disk industry. I apologize for the typo here. I’ll get it fixed later. This talks about some of the forward-looking technology. I’m going to skip this slide. Just know that we have applications in FPGA and also MRAM for AI inferencing on the edge, as well as some projects ongoing for neuromorphic computing with Purdue University that we talked about in our earnings call recently. This talks about our roadmap.

Again, distributed by PERSYST, Genesis, and AgILYST. We are in production with PERSYST, going all the way from 1.8 kilobit to 1 gigabit. These are the two X5 products that we brought to production recently. We have a high-density product that is planned later in 2026, 2027. The next product that we bring to market is this 256-megabit to 2-gigabit X5 product for replacing NOR flash. This is going to be on TSMC’s 16-nanometer and 10-nanometer. The AgILYST product that we have, we’re working with QuickLogic to demonstrate with our 8-inch. We will bring that product to market with an advanced CMOS node, either with TSMC or GlobalFoundries for FDI applications in 2026, 2027 coming up. I want to introduce the Everspin executive team. Bill Cooper, who’s on the call with me today, CFO, brings about 20 years of experience with Advanced Micro Devices.

David Schrank comes to us from Intel, as well as Sean Dougherty, VP of Sales. Recently, we brought Sean on to head VP of Sales. David Schrank is now focusing on VP of Business Development only. Kenny Nagel, who is leading our R&D group, comes to us from Motorola. Almost 20, 25 years of experience with magnetics. Jan Kemp comes to us from Cypress, who’s our VP of Product Development. Our operations is led by Amit Shah and Kalvus Bergkjær, with experience with Intel and Samsung, whereas Amit’s experience is with Marvell and Semtech. A very diverse background and a very deep background for each of our executives, bringing in their expertise to drive Everspin’s future growth. I’m going to hand this over to Bill.

Bill Cooper, CFO, Everspin Technologies: Thanks, Sanjeev. As we look forward and look a little bit backwards, you can see that Everspin had a fairly sizable inflection point back several years ago in 2021. Like many folks in this area, we saw some pandemic tailwinds. Some of that saw a little bit of drop-off in 2024. In 2025, we’ve achieved revenue of $26 million in the first half of the year and raised our guidance into Q3. We’re kind of still in that strong $50+ million range. We have a stated goal externally as well that we expect to double our revenue from roughly that $50 million range to $100 million in the next three years. The company actually has a strong cash position, $45 million at June 30, no debt on the balance sheet. We’ve been consistently free cash flow positive.

We’re cash flow positive for the last several years, as well as the first half of 2025. You can see that our margins here have been kind of hovering in that upper 40% to upper 50% range. That always depends upon the mix of product as well as non-product revenue, which is high margin type activities, as well as the introduction of some new products as well. Certainly expect to continue that in terms of the gross margin achievement as well. With that, I will stop.

Operator: Thank you, Sanjeev and Bill. Let’s move forward to the Q&A section. What are the key products in development that position Everspin to meet emerging industry needs?

Sanjeev Aggarwal, President and CEO, Everspin Technologies: Yeah, good question, Steven. The next product that we are planning to bring to the market is a 256-megabit X5 product to actually replace the NOR flash, obviously, to enter the NOR flash replacement market. We call it an enhanced NOR product. The reason is, like I said on the call, it’s two to three orders of magnitude faster writes, two to three orders of magnitude lower power, and two to three orders of magnitude more read-write cycle capability. Since NOR flash has stopped scaling in density, this is a good opportunity for Everspin to actually enter that market, which, as we talked about, is a $3.5 billion market by 2029. It gives us a good opportunity to enter and increase that revenue. That’s where we see our inflection point over the next three years or so.

Operator: Can you describe Everspin’s business model, including how revenue is generated through product sales, licensing, or partnership?

Bill Cooper, CFO, Everspin Technologies: I’m sorry, can you repeat the question? I didn’t catch it until the end.

Operator: Yeah. Can you describe Everspin’s business model, including how revenue is generated through product sales, licensing, or partnerships?

Bill Cooper, CFO, Everspin Technologies: Yeah. Most of our revenue, 85% to 90% of our revenue, is generated through product sales. We go through a distribution model, so we have distributors. About our revenue is dispersed geographically throughout the world. It’s about a third in Asia, a third in Europe, and a third in North America.

Operator: OK. Which markets are the primary focus for Everspin? What drives your focus on these sectors?

Bill Cooper, CFO, Everspin Technologies: Yeah, certainly, we, as Sanjeev mentioned, we have strong presence in industrial, gaming, some medical, avionics, also aerospace, and satellites are one area in particular that’s growing as well. Of course, we’re also focused on the data center market as well. As Sanjeev mentioned, this newer product that we’re going to be coming out with is going to be focused on moving into the NOR flash market as well, which is sizable.

Operator: Got it. What are the most significant milestones in Everspin’s journey that have shaped its position as a leader in MRAM technology?

Sanjeev Aggarwal, President and CEO, Everspin Technologies: Yeah. The good thing about Everspin’s approach to this thing is we developed a technology that can be tuned to look like a data logging memory. It can be looked like a data center memory or even a configuration memory. We have all three types of memories in production today. That’s the uniqueness that we bring to the market. We’ve been able to scale our densities from low density, like 128 kilobit, all the way up to 1 gigabit, which is a high density, highest density product that we’re selling today. I think given this growth, the milestones that we have is we first brought the data logging memory to market in 2006. In 2017/2018, we actually brought the data center memory to market. That’s the 1 gigabit that we are shipping to IBM today.

Going back to the data logging market from 128 kilobit to 16 megabit that we’ve been shipping since 2006. Recently, in 2022/2023, we brought our X5 Spin-Transfer Torque MRAM memory to the market. With all these variants, we are actually addressing all types of memories in the market. We do believe that we’re going to be able to revolutionize the memory market with Everspin MRAM. That is the goal.

Operator: OK. Are there any other questions from the audience? Feel free to drop it in the chat. I think there are no other questions. We will now close this session. Thank you, everyone. Thank you, Sanjeev and Bill. Thank you for your time. Have a wonderful day.

Sanjeev Aggarwal, President and CEO, Everspin Technologies: Thank you, Steven.

Bill Cooper, CFO, Everspin Technologies: Thank you.

Operator: That concludes Everspin Technologies’ presentation. You may now disconnect. For details and upcoming presentations, please refer to the conference agenda. Thank you for your participation. We look forward to welcoming you to the next session.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.