Kopin Corp (KOPN) CEO Michael Murray: AI Infrastructure Catalysts and Growing Defense Opportunities

We have a tremendous opportunity for growth at this company. Our expectation is that at the end of this year, the company will be profitable. We're growing into our capex structure. The United States government is investing in the company. We have a tremendous order book as we sit here today with some of the best customers in the world, market leading and market making customers in the world. Our European growth is being fueled by Thales International. We have great partners in Andros and the unusual machines and we're winning in the drone dominance competition because of our capability that we provide these flyers and pilots to see their terrain and score well in the drone dominance competition. And we're the only company in the United States that has LCD, OLED, micro LED and LCoS manufacturing all under the same roof, roughly speaking. So, we have a great platform for growth. I think the AI infrastructure market for Kopin could be 25 to 50% of our revenue next year and the year after it could be as much as our defense revenue in 2028 based on some of the opportunity levels that we're seeing.

Well, hey Martin, thank you very much for having us back. We had a great first impression when we talked earlier this year and looking forward to updating the audience on some of the great progress Kopin has made recently and talk about some of the things that we're expecting for the second half of this year.

We'll do Martin. So, firstly and foremostly, thank you for having us and to those folks that don't know me, I've been the CEO, chairman and president of Kopin for over 3 years now approaching my fourth year in its entirety. I succeed Dr. John Fan who started this company over 40 years ago out of MIT Lincoln Labs. The company has had significant turnaround over the course of the last 3 years mainly due to focus. When I took over the company, we have tremendous technology, a wealth of ambition here, but I think what the company really lacked was a definitive focus. We were very much focused on what was called the meta metaverse back when I joined the company and I looked across the markets that we were serving and saw that there were tremendous opportunities in certain markets that were closer to us that we could make more money on more quickly and there were just more focused growth opportunities for the company. So, with that we changed the strategy over 3 years ago. We're now in our second phase of that strategy and I'm happy to report that things seem to be moving along quite well as the markets that we hoped were going to move towards Kopin are moving towards Kopin and we're helping to shape those markets with the market makers in each of those markets. I'm going to talk about that here today. But before I go on, right now the company is just finishing up our quarter. So, I won't be talking about too many financial metrics other than to say the company has no debt, no warrants on the balance sheet as of the last posting on 427. We had over $60 million in cash and about 181 employees, which is currently growing due to some significant orders that we've received specific to our AI GPU strategy. So, we're going to talk a little bit more about that today. We'll also talk about drones. Everyone wants to hear about drones and what we're doing in that space. So, I'm going to double click on drones and I'm going to double click on our AI GPU infrastructure play with our friends over at Fabric AI. Okay. So, essentially the company is based around our microdisplay portfolio. Who we are is a semiconductor development and optics company. These tiny little TVs are semiconductors. We build them in a semiconductor flow. What I mean by that is we have a clean room here in Massachusetts as well as in Dalgety Bay, Scotland. And we recently purchased an entire OLED deposition capability that's going here in the United States because of the demand that we're seeing in first-person viewer drones as well as the demand in thermal weapon sights going up for OLED as well as other applications where OLED happens to be the better display for certain applications. Now, not all displays will solve all problems. As an example, LCDs, which we still build here in the United States up here in Westborough, is still a great solution for thermal weapon sights as an example because it's very robust. It can withstand tremendous shock like a weapon firing as a perfect example. OLEDs are a little bit more difficult, but we're getting there. We've learned how to design those into thermal weapon sights, and we're now building those here, or we'll be building them here in Massachusetts. Also, we've been contracted by the United States government to build micro LEDs here in the United States. And the reason why is not obvious to many investors. Micro OLED and LED specifically are two very specific technologies that allow you to see farther at lower power consumption and with better color and contrast. So, LCD is going to be phased out over time. It's going to take a while. OLED's going to come up, and then micro LED is the vanguard of technology. Now, I'm going to pause on micro LED for a moment. Micro LEDs act a lot like lasers. They can transmit data as well. And this is something that you're seeing in the industry, whether it be Nvidia as an example, or our friends over at Marvell, who have been doing some work in this area with one of our competitors as an example. Micro LEDs happen to be a great technology to transmit data over short periods of distance. So, I'm going to talk about that a little bit more in coming slides. Okay. So, with that in mind, Kopin has expanded our customer base with the tier one customers that are market makers. As an example, we partnered with Theon International, which is when you think about NATO. When you think about NATO thermal weapon sights and night vision goggles, you should be thinking Theon. That's who their biggest customer is, and they own about 60% of NATO's market share of thermal weapon sights and night vision goggles. Both of those technologies require displays. And we'll be building those displays with them, for them, in our European location, as well as here in the United States. We also have a tremendous relationship with folks like Wilcox, Lockheed Martin, Leonardo DRS, who's our largest customer, and a number of other customers that we're going to add to this list by the end of this quarter. We'll talk a little bit more about that in a second. So, I wanted to show you today our five pillars of growth. Many people have been asking me how did I think about reshaping the company and what our focus is going to be and why the market is now coming to Kopin and why we're being successful with it. So, when you think about the five pillars here, it's all based on being able to build this technology right here in the United States and why that's fundamentally to the sovereign nature of our defense organizations. So, when you think about mobile vision systems, these are fixed-wing aircraft. We have a huge franchise in one of the aircraft specifically that's classified here in the United States, we have three generations of displays going into that aircraft. I can't talk about which one it is, but it's getting tremendous orders through the European and Southeast Asian markets. And that's great for Kopin because in the actual helmet itself, there are two displays. But, what's not obvious to a lot of investors is in those other regions, not in the United States, where in the United States you have one pilot and one plane, in these other areas you will have one plane, potentially five pilots or four pilots or three pilots, but also have two helmets. So, that's why our international sales is going up, not down. Also, we've been able to secure several new orders in fixed-wing and rotary-wing in Europe for the first time. We've been very much focused on growing our European business as Europe and NATO starts to spend more sovereign dollars. We had to pivot and change our Daggett Bay location to focus more on defense, and I'm very, very pleased to say that our progress in Europe has been exponential this year. We've already announced several awards that will be going through our European location starting next year, which is fantastic news. One of the other growth pillars that I was very focused on when I joined Kopin is the soldier born systems pillar of growth. So many of you have heard about soldier born mission command, which is the old IVAS program. We've been working on several different solutions and several different partners in that program with our micro LED capability, which the United States Army has invested in Kopin to build that technology here in the United States. And that is a color micro LED that will be available by the end of 2027 for insertion into the program 28 when it goes into production. And I'm pleased to say that's well on track. Secondly, we've been working with the US government to put in a color micro LED production line here in the United States. And that too is on track. So we'll be one of the only companies, potentially the only company with our own micro LED production line here in the United States. And that same production line can not only build our micro LED solutions, whether they be monochrome or color, but it can also build our neural IO technology, which is our AI interface technology. And it can come off the same production line. That is a huge advantage over all of our competitors. The fact that we'll be able to go to our own production line and basically tweak the recipes that we need to tweak to make this technology work. And we know how to do that better than any company in the world in my view. We'll talk a little bit about first person viewer drones. This morning we announced that we've received several new orders from folks that are participating in the drone dominance competition. Right now, we've had several awards from the top 10 in the drone dominance competition. We're also partnered with folks like Andor and UMac, who are great partners of Kopin to bring first person viewer technologies and displays to them here in the United States. The FTC as well as the US Army has required every program to have non-Chinese displays and optics and electronics in it. And that's creating a huge opportunity for Kopin to provide US-based microdisplays, optics, and drive electronics to first-person viewer drones. And our Sentinel program, the Sentinel FPV program, is very unique in that it's bidirectional and situational aware. What I mean by that is you can still see out and outside of the eye box. So, if you're an operator in the field of war, you still want to be able to see your feet so you can run, shoot, and operate a drone at the same time. And this swivels off the helmet or what we call off the visor. So, we're getting tremendous opportunities for that technology, and it just makes a lot of sense for the US Army to go with a solution like that. While we're still working with the predominant manufacturers in this industry like ONDAS and UMEM on solutions that they'll bring to market as well. So, huge volume capability and opportunity for Kopin. So much so that it made better sense for us to buy our own OLED deposition line, and we'll install that in the first half of 2027. So, significant amount of volume that we're expecting for first-person viewer drones. So, the next market that I'm going to spend some time on is our AI infrastructure and co-packaged optic capability. This is an area that we think we have a significant market leadership position in because we've already been producing microLEDs in production that fly in aircrafts as we speak. So, we're going to double click on that a little bit later. So, here's some areas that I'm very proud to talk about. When you look at this slide, what's changed is the orders. And whenever I see new orders on a slide, it always brings a smile to my face, as an investor, but we've announced new pilot helmet awards recently, new thermal weapon sight awards recently, and several US aircraft rotary wing aircraft awards, as well as European rotary aircraft awards in Q1. DarkWave is a technology that we partnered with Teon International, and that is an area where we're also receiving awards right now. We expect to go into production in Q4. And then, of course, we've already received and announced several new orders in our Sentinel FPV headset. So, what this slide should tell you as an investor, I hope, is that when Kopin spends money in advanced research and development for things like DarkWave or Sentinel, etc., that we're actually getting that investment back in spades, and our new investment in new technologies is starting to pay us back. So, with that, just a brief update with our Soldier Board Mission Command, our color micro LED. We are progressing quite well. I think the program management team has done a wonderful job here. We're on target with our color micro LED, and this is something that the Army has awarded Kopin. And for folks that don't or aren't aware, this is a $350 to $500 million contract to Kopin if and when we get an insertion into the program with our color micro LED. And this will add tremendous power consumption reductions in the system, true color, excellent contrast, and certainly the best image for our warfighters to go into the battle with. So, we're very proud of our progress so far, but we have a lot more to do. Okay. I'm going to touch on a few things here, but when we think about the overall market for color micro LEDs, just here in the United States, it's a $1 billion serviceable available market. And remember, you have to build this tech in the United States to be able to support these programs. And right now, Kopin is the only company with a micro LED production line coming online, to my knowledge. And that gives us a tremendous advantage in the marketplace. We also expect further IBAS funding either this year or next year to help us complete and add more capacity to our production line and potentially other devices as well. So, right now we are working with the army on other micro LEDs for other applications, whether it be a smaller resolution or higher resolution device, and those discussions are ongoing. And furthermore, Teon has funded us to build a NATO version of that micro LED once and if we're ready to do so, and they have some requirements for us. So, that's ongoing as we speak. So, let's talk a little bit about why we think we're going to be successful in the optical interconnect and packaged optic market. Right now, we have a 2K by 2K, which is 16 million pixels or lanes or channels on 1-in device. That is very difficult to build. That's why many micro LED companies that are trying to get into co-packaged optics, as an example, are failing. They're going through the same lessons learned that we went through over the course of the last several years. We've also spent over $50 million in developing this technology for the Department of War, certain aircraft programs, etc. So, we've learned a lot in building micro LEDs, and we've patented much of this technology. And we have some interesting ways of doing what we do to make the yields and dispersion effects be improved. So, with that, we've received a $15 million research and development order from our friends over at Fabric AI. We're developing a parallel chipset which will achieve 1.6 terabits per second transmit and receive. Those prototypes will be available next year, but we're going to demonstrate this capability this year if not by the first part of 2027. So, I'm going to talk about why. Why this is important. And when you think about transmitting data at terabit levels, many people would want to use lasers, which is an opportunity if you're transmitting data over a distance that's quite long. However, from a chip-to-chip perspective or a board-to-board perspective, lasers just aren't necessarily the best solution because they're too power hungry, they're too physically large, but they can solve the bandwidth problem. Now, currently, the technology that is used is copper serdes. Unfortunately, copper, if you push it that hard to 1.6 terabits or 3.2 terabits, it heats up too much to the point where it is costing the average consumer and customer too much. So, that gives micro LEDs a really interesting in-between opportunity. We're not quite a laser, we're definitely faster and cheaper than copper interconnects, and we're not as hot, but we can't go as far as lasers as an example. So, one of the other things that I think investors don't necessarily value with our technology is that it's field programmable. Now, if you're the United States government that already has a tremendous amount of AI infrastructure in place, you don't want to start ripping and replacing all of this gear because A, it's already working, B, you've already paid for it, and C, you may not necessarily need every component that's coming out. But, field programmability and field programmable IO is a tremendous opportunity for these types of customers that don't want to rip and replace their AI GPUs. Lastly, one of the interesting things that we're focused on that may not be completely obvious to folks is that we're very much focused on the United States government and Department of Defense in their AI infrastructure centers. One, because we build this technology in the United States. Two, that rip and replace greenfield versus brownfield opportunity that I just talked about. But also, getting AI closer to the battlefield is uniquely important. And what we're seeing in Iran right now is the closer you can get your AI, the quicker the decisions you can make, and the more accessible AI will be to the battlefield. So, to do that, we have to reduce power consumption. And because of the issues around copper, pulling all that copper out, pulling all the cost to cool it out, and adding micro LEDs for GPU to GPU technology, GPU to memory, as well as board to board, we think we can remove 17 to 20 kilowatts per rack per month, which is tens of thousands of dollars of savings in electricity. So, very huge opportunity for us. The United States government is very interested in working with Coherent. We're actively working with them to create some projects for 2026 and 2027. And we expect there'll be more information on that soon. So, just to let you know, Coherent does not need more CAPEX to grow into this market. We're actually growing into the CAPEX that we already have. We inherited or I inherited three very large manufacturing plants that are underutilized. And now we're trying to utilize those fabrication plants specifically in Dalgety Bay for a lot of this back-end processing. So, as we fill up these fabs, our utilization rate goes up, our absorption rates go up, and our profitability also goes up. And that's why we're projecting to be profitable by the end of this year, which is an advance of what we were saying earlier just simply because of the amount of orders that we received in the first half of this year. So, with that, this is the team. We are adding to it. We just announced a small design center in Dallas, Texas for our optical team. It just so happens that Dallas is a great place for photonics and optical engineers, and we have to go where the talent is. So, a little bit of capex and opex that we're going to spend there, not much, but it's because we have orders from customers to fulfill those orders that we need those engineers in that spot. So, great team, we're focused, we have excellent customers, the markets are coming to us, and the strategy is starting to avail itself now. So, with that, I'll open up to some questions, Martin.

Yes and no. I hate to do that to you. It's a horrible answer. Yes, in some areas. We learned, as an example, on this page alone, several different things. As an example, we could have had a massive opportunity with the Abrams tank program called the SEP V4 until the Abrams tank went to Ukraine and sank because it was too heavy. So, now the industry is starting to learn that size, weight, and power consumption really matters as well as gas consumption. So, yes, we're starting to speed up and learn more quickly. We're also learning that next generation interceptor missiles need to be different because we'll be shooting different things. And as well as, having a first-person viewer drone capability that allows you to still see your feet and your peripheral vision is really important. So, yes, I think the United States Army is learning quicker than ever before. I'm very encouraged with the talent that I see in the United States Army and a lot of the divisions that are developing this technology. And certainly Martin, what an IBAS award means for us, that's the industrial base analysis and sustainment. What it means is the US government is looking at technologies for the industrial base, meaning how do we produce here in the United States for the United States? That's critically important and thankfully they're learning more quickly and investing that money now. I think there's areas where Europe quite frankly is struggling. We see Europe still trying to do things the old ways and companies like Theon are just growing exponentially because of their pace Martin and that's one of the reasons why we like to partner in with them is they're moving very quickly in Europe and Southeast Asia and Europe isn't moving quickly enough. Ukraine certainly is moving quickly, however, the rest of Europe isn't and I think the United States now is definitely catching stride and investing in the right ways in the right places.

Yeah, great question actually. We took the approach that customers that we have, let me scroll up here. Like the big primes as an example, they don't want to get smart on micro displays, optics and drive electronics. They're big metal benders. Their core competencies are in other areas. So we took the approach of more application specific solutions where we're adding more than just the lens as an example for your glasses. You don't go into LensCrafters and just buy the lens, you buy the lens, the eyepiece, the holders, and your prescription. And that strategy's worked very well for us specifically, as an example, with our Sentinel FPV program. Sentinel is not a product, it's a platform. You can tweak it, you can change its color, you can change its connector for whatever solution you want to provide, but the eyepieces themselves, the display optics and drive electronics, that's very much within our gift to commoditize and build here. So, that's where we've been focused. Our automation and capacity is in building those types of architectures so that we can rinse and repeat them many, many times. Now, what caught us a little bit off guard, and I'll be the first one to put my hand up here, is the demand that we're seeing for displays now here in the United States. We were thinking that OLED would have somewhat more of a traditional ramp, but for us with things like our first-person viewer and some of the thermal weapon sites that we're seeing, and that demand curve became a demand step. So, it wasn't a nice curve, it was a demand digital step. So, we had to respond accordingly.

Add more capacity because our current European suppliers wouldn't be able to keep up, and secondly, it was financially more viable for us to put OLED deposition here in the United States and utilize that and capitalize it here in our fabrication plant here. So from an OLED perspective, we're in very good shape for demand. MicroLED, we're already considering adding capacity as soon as the MicroLED production line gets operational, we'll likely add more modules and nodes to it to produce more because we're just seeing a significant demand curve, and that demand curve is very near field.

Yeah, it's been fits and starts with consumer. I think even the consumer companies would tell you that Meta has a 100 for 100 strategy. And I was on a panel with Meta and a few other folks recently at SID week. And their view, which is I think very interesting, is there's going to be a low cost part of the market. And you're seeing that with Meta's glasses coming out, Google's, etc. So this basis of lower cost AI headsets and glasses are seeding the market. And I think what you're going to see then is the Apples of the world, the more integrated type of glasses come later. If you recall what I said at the beginning of the call was we were focused on the metaverse first and Apple Vision Pro, etc. And I think the market's realized is we don't want a $3,000 headset right now. The apps aren't there to spend $3,000, $4,000 for a glasses as an example. So let's go low cost first, build up the applications, and then the big $2,000, $3,000 type of eyewear will come out from those companies. And what Kopin's focused is in truly differentiated technology like our neural display technology. We demonstrated this over two years ago. We invented it. This is a bidirectional microdisplay. It actually has a camera that's looking back at your eye. I use camera as a layman's term, but it's actually you view using the pixels of the display to do eye tracking as an example while you're still seeing an image. That reduces cameras that are doing eye tracking, the power consumption and processing that you need. And what we learned from the market is that they don't want that technology in an OLED display. They want that technology in a microLED display because OLEDs right now are not bright enough for daytime use. Where you use smart glasses most often is in daytime use. Therefore, LCoS and microLEDs are going to win the day in things like the Google Glass, Meta Glasses, Apple Glasses, etc. So microLEDs are required for those applications just strictly for the brightness. Therefore, we need to build a neural display in a microLED and or work with a partner in Korea or Japan that can license that technology from us for one of those big three on the West Coast and build it in their billion-dollar microLED fab as an example.

Yeah, it's a great question. We don't want to compete with our customers. And we want to be somewhat agnostic since we have this great opportunity to support multiple different customers in multiple different ways. Where I think we want to automate and focus first, at least in the near term Martin, is the display manufacturing, have to have it here in the United States and we have to be great at it. And we are and we're getting better and we're expanding that footprint with OLED and microLED which is fantastic. But secondly, optics. Marrying that microdisplay with the right optic is tremendously difficult and our customers don't want to do it. They want us to do it. And then wrapping that optic around with some sort of plastic or housing, then they get a module that they can plug in and play and go off and add value in the way that they want to add value. So we see that in the first-person viewer drone market as an example, with all these customers competing against each other, whether it be on the motors that they use, the distance that they have, the airframes that they have, they don't want to necessarily compete on the headset. They want someone else to do that for them. So I'm not going to produce a product in that space. I'm going to produce a solution that's 80% of the way there, and you, Martin, can tweak on that, paint it a different color, make it a different connector for your helmet system as an example, or make it a full goggle, it's up to you. But the fundamental LCD, OLED, microLED, or LCOS, plus the optic, drive electronic, and some sort of housing, that should come out of the Kopin manufacturing plant fully and completely warranted and ready to go.

Yes. So part of why and how we build microLEDs, there's some reasons we have programmability in those systems. I'm not going to go into it because it's a significant advantage for us. But having that programmability will allow us to do different things in the market. As an example, let's say you have an Nvidia H200 chip sitting on one rack, and in your IO card, you want to be able to talk to a legacy Intel chip, as an example. One is new, the other one is old. Our ability to program our chip to an H200 NVLink, and then the receive side, an Intel custom interface as an example, allows us to talk to both of those systems and transmit data over optics. But we won't know what we're going to have to talk to in a brownfield already installed environment. So what this allows you to do is take older architecture and bolt on new architecture as long as the IO interface cards have our chip in it. And we've heard that from the United States government that that is their largest fear of having to rip and replace every time Nvidia comes out with a new GPU or Intel comes out with a new CPU that they have to rip and replace their IO cards. Well, if the IO card is already programmable and your chip is programmable, then you have the ability to program that interface and use that old technology where it is and how it is, but still link those boards using optics as an example. So that's one of the reasons why we do it that way. There's a significant amount of capability that we actually do through software to make those LEDs behave and also things that we can do for redundancy as an example of those LEDs because to place 16 million LEDs, 16 million, on the size of my thumb perfectly is really hard to do. Go ask some of the customers that are out there that are getting investments from folks like Marvell or Nvidia or whomever, putting 16 million LEDs on the size of your thumb perfectly every time is really hard to do. And that's why they're finding out that a lot of these technologies that they promised for micro LEDs have not come to fruition yet because they've not gone through the growing pains that we have building this type of technology. And we've built it for several years and it's being flown in an aircraft today. So we know how to do this better than I think anyone in the United States, period, full stop.

You have it exactly right and we have a very exciting road map. I want to make sure that people understand we're not a one and done with Neural IO. We have a full co-packaged optic road map that I think is two years farther advanced than the leaders in the industry here in the United States. We're not ready to talk about it yet. But again, when I start talking about the ability to put 16 million pixels on a 1-inch chip with microlenses and controlled dispersion, I think we're about two to three years advanced than everybody else in the industry. Folks just haven't really realized it yet because we haven't really been talking about it that much. When Marvell came out with their announcement, it kind of pushed my hand to announce what we were doing with Fabric AI. And I think this new technology called Neural IO, which is that transceiver chip, it's a receive and transmit, Martin. I think our road map, we're now contemplating maybe just focusing on the transmit side because we do it so well and we know that the other players in the industry are really struggling with transmit. The receive side of that data is actually fairly well known. The science is well known on this and there's many companies on the receiver side of optics, but not too many on the transmit side that can do what we do, which is 16 million pixels on a 1-inch substrate. We're thinking that we can get well over 3.2 terabits per second of data near term. We're not talking years, we're talking near term. And I think 10 terabits per second is definitely in our view. So we think we can do it and we have a significant amount of capability on the transmit side. But we're learning on the receiver side and I think Neural IO is going to prove a lot this year that we can do a tremendous amount of capability in this space with technology that we have here already.

I think slow and wide. Meaning, if you have a laser as an example, you have to run that laser because it's one line, as fast as you can. Many, many lines, you can transmit data in parses, meaning in pieces, or you can transmit tremendous amounts of data at the same time across multiple different lines, so your bandwidth is exceptionally large. So think slow and wide. Now, we're able to do that. We invented the bidirectional microdisplay 2 years ago, if you remember what I said. We knew that we could transmit data back and forth at the same time. Our challenge is how much receive can we do? On the transmit side, we know that we can outperform pretty much any company in the United States and the world today on the transmit side. It's the receive side that we're doing a lot of learning and development on. But yes, essentially each of those lanes would drive data to and from the micro LED. Now, our ability to program those LEDs specifically by LED, we literally can control one LED with one pixel or one register in our programmable backplane. So as an example, let's say you have one LED out of the 16 million that isn't working. Then what do you do? That's a problem. For us, it's not a problem. We just program it out, and we use another LED. That's one application that we have figured out that nobody else has, and everyone's struggling with the yields because of that. It's a very simple solution for us. So programmability is critical for us. The ability to transmit at 1.8 million foot-lamberts as a degree of brightness as an example, 1.8 million foot-lamberts will blind you. Literally blinds you. So where again I come back to the market, many people that I talked to recently are talking about we're going to kill copper. We're not going to kill copper with this technology. It's an and technology. There'll be copper serdes. Folks like Credo do that very, very well. Highly competent company. They're focused on that. Copper will still exist for a long, long time. But it was there when I started engineering, Martin. So it's been there a while. And then there's lasers. Lasers are great for longer haul and higher frequencies. But in the middle, chip-to-chip, board-to-board, there's a better solution than those two, and that happens to be a micro LED. It's less cost. It's more effective, high bandwidth, programmable, low cost to build, and certainly much, much less power consumption than either of those two solutions. And the meantime between failure is actually better with a micro LED than it is a laser. So for all these reasons, micro LEDs are very well suited for chip-to-chip, GPU-to-memory as an example, and then board-to-board. Rack-to-rack, getting a little more difficult for a micro LED. You might want to add a cable plus the micro LED, which is what Microsoft is doing as an example. That's publicly known. But for us, we're very focused on the short-range, chip-to-chip, board-to-board, because that's where I think micro LEDs are going to have their day.

The FABC market? I think it's... So Fabric AI is called FABC, that's their ticker on the Nasdaq. Great company. What Fabric, as I understand it, is going to be, this isn't going to be their only product and only company that they partner with. They're very focused on solving the four big pillars of issues around AI infrastructure. And James and Josh are focused on bringing other companies like Coherent into the Fabric fold. But if you just think of the market spaces out there for transceiver like as an example Credo, I think is 12 billion or 20 billion dollars, and the Fabric AI market share or market size is, what, 30 million? The arbitrage between the two is so massive that I think Fabric's ability to grow its share price is exceptional. It's massive. Especially if we get this chip right and I get one customer, that stock price is going to run. So I think their market potential at Fabric AI, if they do the right things, is exponential. Just think of it as a play between the market size that it is today and what you see with Credo or Marvell. That's the opportunity that I see in Fabric AI. And Coherent for that matter. It's an arbitrage. Even valued at a billion dollars, Coherent, if you go back to that slide that I had, this one, and look at the market opportunity in these markets in the United States with a fabrication plant that nobody else has and capability that nobody else has and we're already in the market, these market opportunities and the numbers that were put up here by 2030 are very real for us. And I think our growth trajectory over the course of the next 3 years is exponential. So I plan on addressing that with the market over the course of this year as we see our order books start to fill up and I see that Fabric AI's order book starts to fill up, we start partnering with more AI infrastructure companies together, and I see Fabric bringing in other companies like Kopin that has bespoke technology, I think both market values of both companies right now are underserved by a lot. So we're very excited that we're positioning ourselves for great growth.

It's a good question. I actually see the defense industry really wanting this technology first.

Again, because they have the biggest need and they have this addressable market, Martin, that I think a lot of investors don't realize. The United States government is the second largest buyer of AI infrastructure behind the hyperscalers. They're the second largest. And more importantly, the chips that go into the US government cannot be built in China. So all these other companies that would necessarily compete with Kopin, like the Marvells and Credos of the world as example, they're going off and building their chips in China or Taiwan. We're building our chips here in the United States for Fabric AI as an example, for the Department of War, for NIST, for DISA, for three-letter agencies and four-letter agencies. So I see them really wanting this technology now because we also solve that big problem, Martin. Brownfield and greenfield coming together so that you can use the next H200 without ripping and replacing everything that's out there or the A300, whatever it might be. That co-packaged optic side of that is a different equation. So let's talk about the co-packaged optic size for a second. The ability to support a co-packaged optic from Kopin is actually very unique for us. Because we do the transmit side so well and our pixel sizes are far smaller than anybody else's, and I understand this to be true, our next competitor's pixel size is like 25 microns to 50 microns, we're below 10 today. That means we can put more LEDs in a smaller space, run them faster at a cooler level, lower power consumption. So our ability to put together a 3.2 terabit co-packaged optic for transmit is far advanced than everybody else's. I think instead of 3.2 we should jump straight to 10. And we're debating that right now. I think there's no reason to go to 10 terabits right now because the GPUs can't run that fast. So that's where we are in the industry and I don't think people have really figured that out yet.

We've been focused on execution right now, Martin. We don't even have a salesperson selling the stuff yet.

So when we're ready and we're coming soon, we have a full road map, we have a demonstrable chipset that we'll have by the end of this year that customers can use and test and what have you. But until you really have something that customers can test, it's difficult to engage. But we do have several customers, they're semiconductor companies that have signed NDAs with both Kopin and Fabric AI to be early adopters and testers of Neural IO. On the co-packaged optic side, we really haven't been addressing the market yet. We're waiting to see how our test chips come back. But from a co-packaged optic perspective, if you're building LED anyway, how many can you build on 1 mm is the question. And our answer is we can build a lot more than people think we can, and they're faster than they think they are. So when we're ready, we'll go out to the market and hopefully sign two, maybe three partners that one will hopefully be a semiconductor company, and then the other will be an infrastructure company that we can work with. And then the third obviously is going to be the United States government. No matter what we do, we'll have a United States government customer for this technology because we'll be able to take more risk with them. They'll be willing to fund that risk. And I think that's great for Kopin investors is that the technology risk on Neural IO as example is purely borne by the United States government and Fabric. That's it. So we'll have a chipset funded by those two entities as opposed to Kopin's own internal research and development. And that's a huge thing. That's a big difference that I bring in from my background versus where Kopin was is that we're working with our customers and our partners to fund this technology development with them and for them so that they get what they need as opposed to us giving them what they think they need. And that's been a huge change in the culture of Kopin.

I can say it's difficult to talk right now because we're just finishing our quarter, but we're doing well this year. I think next year our growth potential is more exciting than I think people realize. Potentially I was a little shocked actually. Happily so. But I think where we want to be by 2028 is going to be very significant revenue growth and we'll definitely surpass the $100 million is my goal. So we'll see how Q2 shapes up and get that announced. We announce in August and at that point in time I'll take a judgement of whether or not we're going to raise our guidance for this year and next or not. But right now we are very focused on just delivering on our order book, increasing our order book and closing the quarter. That's about all I can say today.

I think so. I think we're going to have a good year this year. We guided 52 to 60 million this year which is a good step up from last year. Last year was a little depressed Martin because of the government shutdown in Q4. That hurt us. No one expected it to be that bad. But I think this year we'll definitely hit our guidance in my view and if things keep percolating here we'll see where we end up at the end of Q2 and whether or not we want to adjust our guidance for this year and next year. But I do foresee next year being a significant growth year for us. In fact, when I started we started talking about 75 million for next year. Actually it's right here. So I can talk about that. 75 to $100 million next year is still well within our scope and I'm happy to report we're getting the orders and that's what really fuels me is if we're getting the orders and the order book is growing the way that it is, I have a high degree of confidence that we'll be in that $75 to $100 mark next year.

Yeah, update is it's in federal appeal as we speak. Documents have been entered on both sides and we'll wait and see the outcome of the federal appeal. I don't have any timeline for when that's going to be resolved.

Yeah, I think firstly and foremostly, we have a tremendous opportunity for growth at this company. Our expectation is that at the end of this year, the company will be profitable. We're growing into our CapEx structure. The United States government is investing in the company. We have a tremendous order book as we sit here today with some of the best customers in the world, market leading and market making customers in the world. Our European growth is being fueled by Theron International. We have great partners in Anduril and Unusual Machines and we're winning in the drone dominance competition because of our capability that we provide these flyers and pilots to see their terrain and score well in the drone dominance competition. And we're the only company in the United States that has LCD, OLED, microLED, and LCoS manufacturing all under the same roof, roughly speaking. So we have a great platform for growth. I think the AI infrastructure market for Kopin could be 25 to 50% of our revenue next year and the year after it could be as much as our defense revenue in 2028 based on some of the opportunity levels that we're seeing. So exciting opportunities for growth for the company and yeah, look for more during our Q2 announcements.

Thanks very much, Martin.