Good morning. Next up is Jonathan Klamkin with Aeluma.
Thank you very much for the introduction. Thanks everyone for being here today. Look forward to introducing you to Aeluma if you're not familiar or providing an update if you have been following the company. I'm Jonathan Klamkin, Founder and CEO of the company, joined here today by Aeluma CFO, Chris Stewart. Here's a slide with forward-looking statements. Let me start by sharing Aeluma's vision, t hat is to build the world's highest performance semiconductor chips with scalable manufacturing. If you're familiar with semiconductors, high performance doesn't always go hand in hand with scale. High performance is typically specialized, low volume, whereas scale comes with sort of mainstream microelectronics. What Aeluma has done, it sort of brought high performance materials into mainstream microelectronics manufacturing. Here's a background story.
This is not what we do at Aeluma, but it's a very nice story to sort of set the stage for what we've done, which has made materials breakthroughs to hopefully enable an industry. This background story is about Shuji Nakamura, who's a Nobel laureate. He's a friend and colleague. He invested in Aeluma and was an early advisor. If you're familiar with this technology and invention, gallium nitride is a compound semiconductor material that nowadays is used for many things. It received a lot of intention to try to build blue LEDs for lighting applications and displays.
Dr. Nakamura made some breakthroughs in the manufacturing of that material so that you could build things like blue and then white LEDs that are now used in lighting and other applications. He, of course, was subject to a number of naysayers that said, "This will never work. This will never replace traditional incumbent kitchen lights." But as we know, the LED really revolutionized lighting and displays. We can all go to Home Depot and other stores and buy LED light bulbs. They're in cars. They're in specialty lighting. What happened here was the innovations made in that material also enabled other industries. This was not just about lighting. This material is now used in power electronics, in electric vehicles, in data center servers.
It's also used widely in what's known as gallium nitride RF for 5G, 6G wireless applications. The Aeluma story is analogous. We have made breakthroughs in the manufacturing of a different set of compound semiconductor materials that are based on gallium arsenide and indium phosphide, if you're familiar. One of the compounds is InGaAs or indium gallium arsenide that's used for shortwave infrared sensors and a number of other optical or electronic components. And we've also heard, you know, a number of things like InGaAs on silicon won't work, p eople have tried this for years, y ou can't grow quantum dots by MOCVD. But we've made significant breakthroughs in the last few years. I'll show this picture again later.
This is the world's first indium phosphide on silicon wafer produced, and that's a 300 mm wafer. These materials are typically manufactured on very small semiconductor substrates, so not so scalable. By now, over the years, we've continued to mature the technology, and we build components, photodetectors, lasers, and other optical devices that could be used in communications and sensing applications. I'll come back to some of these later, but what I'll say is that in our minds, this could have even more broad reach than that gallium nitride materials breakthrough that you saw a moment ago. It can be used in consumer sensing applications, in communications inside data centers, in 3D sensing, in cars, in robotics, in space applications, and so on, and we'll touch on just a few.
Let me provide a brief background of our company. We have a very transformative technology. This is about taking the best-in-class compound semiconductor materials, but manufacturing them at scale the same way mainstream microelectronics are built. IP is foundational. It's not just patents, but we do have 36 issued and pending patents to date, as well as a number of sort of process recipes and trade secrets that we maintain in-house. Our manufacturing approach is capital light. We primarily outsource the manufacturing, and we do that to keep CapEx investments down and to enable rapid scaling. So, we work with materials and foundry and packaging partners, but t o date, we do still conduct some of the proprietary steps in-house, and I'll show you some images a little bit later inside our facility.
We do have a quality management system, ISO 9001, to assist with manufacturing. Significant business traction. Our technology's been validated by U.S. government agencies and defense and aerospace companies, as well as some commercial companies. We're in sampling and a redevelopment with customers and also started sales initially in small volumes. This number of customer engagements is upwards of 30. I'll tell you a little bit more about what markets those are coming from, in a later slide. We've been growing our team. June 30, 2025, our last fiscal year, we had 12 employees. We're now up to 27 people, so we're growing steadily. Those people are across engineering, operations or manufacturing, G&A, as well as business development.
We've got state-of-the-art facilities in Goleta, California, which is just a couple of hours north of here. A couple of buildings, a gain, I'll show some more images later. One is R&D and manufacturing site, and the other just some executive office and meeting space. As I mentioned, what we do to date is materials production in-house, so w e produce those wafers with the compound semiconductor materials. We have a nice test and validation laboratory where we test the chips that come back from our fab partners. Many of these steps that we initially developed in-house have also been transferred to foundry partners. Strong financial outlook. Recurring R&D revenue fiscal 2025, we reported $4.7 million.
That's not the revenue that we're looking to grow, but this is recurring R&D revenue that has come primarily from government agencies or government customers, n on-dilutive funding that has supported development to date, and also led to significant customer traction. We're a Nasdaq-listed company, no debt or overhang whatsoever. $37.8 million cash reported at our last quarter in March. Again, there's some pictures of our site in Goleta. Let me spend a moment on the innovations that we've made and the technology itself. This is all about enabling the next generation of photonics with large diameter wafer manufacturing.
Over on the left, these represent indium phosphide substrates, and you might know what that is, because it's showing up in the mainstream nowadays because of geopolitics and supply chain shortage that is hindering some of the build-out of AI data centers. Indium phosphide substrates are typically small, 2 in, 3 in, or 4 in in diameter. There's some 6-in production, but that's sort of slowly growing. And that's about where indium phosphide would cap out because it's an expensive material, i t's very fragile, a nd the fabs that run these wafers through to manufacture them into chips that we can use are low volume, and I'll show you some numbers later.
This is an example of what's known as a shortwave infrared sensor that could be used for defense applications, industrial applications, robotics, automotive applications, as well as consumer sensing. It doesn't scale because it's manufactured traditionally on indium phosphide. This is one aspect of the Aeluma innovation, is that we take those materials that you ordinarily layer up on these small, fragile, expensive indium phosphide substrates, and we put them on bigger substrates that can scale. Gallium arsenide and silicon, up to 300 mm silicon, and there's that image again. Why is this so important? It overcomes supply chain constraints. I'll mention again the indium phosphide substrate shortage a bit later. This lends itself to highly automated manufacturing. When you go to a silicon fab, people don't touch the semiconductor substrates. They move around with robots.
They're loaded in and out of the manufacturing tools with robot arms, so v ery high yield, high- volume processes. Lends itself to integrating directly with CMOS and also to what's known as 3D wafer-scale packaging, another thing that comes up in the mainstream nowadays. There's a lot of onshoring of semiconductor manufacturing, and that's not only the front end, but that's what happens in the back end in this wafer scale packaging that traditionally happens in Taiwan and other places. But now, the U.S. government is making investments to ensure that we've got some capacity here on shore. This ultimately leads to much lower manufacturing cost when you compare it to indium phosphide, c ould be as much as 20x in some cases, but we more conservatively say 5x-10x lower manufacturing cost.
Again, a picture of one of these large wafers and some of the chips that we've manufactured. So, why Aeluma? Why this technology? Why now? I'm gonna spend a little bit of time on two particular market verticals but give you an idea of how broadly applicable the technology is. You're probably familiar with consumer sensing if you have a smartphone, especially an iPhone, that has a technology such as the Face ID. How does that technology work today? It uses what's known as a near-infrared sensor, and it operates at this wavelength or color of light that I'm depicting here.
The reason Apple chose that wavelength is because there's a dip in the solar spectrum, so it works okay outdoors, and b ecause they can mass manufacture the illuminators, the emitters that send light and illuminate our face, and they can mass manufacture the receivers or detectors that detect the light that reflects off of our face in order to generate sort of a 3D point image and recognize our face so that it unlocks the phone.
The consumer electronics market would be very happy to adopt SWIR, and this is not a new concept. They've wanted SWIR in consumer electronics for as long as sensors have been deployed in consumer electronics. The reasons to go to SWIR, this spectrum over here, are multiple fold. One, the solar interference is much lower, so it works even better outdoors. It works at night. It enables night vision. This black curve here, that's eye safety.
When you go to these longer wavelengths in the SWIR, you have higher levels of eye safety. That means that you can illuminate with higher power. We don't see this light, i t's invisible in all cases. But if you can illuminate with more power, it means you can see better, y ou can see further away, y ou can do more sophisticated 3D imaging as well as a number of other things like health monitoring. That's what InGaAs enables. This InGaAs material that traditionally doesn't scale, it enables you to transition to SWIR. Better resolution imaging, longer range imaging, and actually it turns out you can put this technology behind the OLED screen, so you can reduce the size of that cutout in the phone.
This has long been sort of the roadmap of some of the mobile OEMs to reduce or eliminate that screen cutout that we see. Again, broadly applicable in consumer electronics, but to take that SWIR technology to a large volume market like this requires a radical approach to scale and reduce cost. I'll just briefly show you the numbers. This is what's required if you do manufacturing on these smaller substrates. If you use 3-in or 4-in indium phosphide and you need it to yield 20 million chips for, say, a consumer application, and that's a small order for the mobile market, a s you know, this is more than 1 billion phones a year produced, and Apple has about 15% of that market, y ou need hundreds of thousands of substrates.
If you look at the capacity of the indium phosphide fabs, it's usually around 1,000 wafers a month or at best, maybe 10,000 wafers per month. There's a clear manufacturing capacity bottleneck there. When you switch to the larger diameter substrates, not only do you need fewer wafers, the capacity of those fabs is much more significant, 10,000, 50,000, 100,000 wafers per month. It's very clear why going to sort of a more mainstream microelectronics manufacturing would enable transitioning to consumer markets. Let me spend a moment on the products we've developed and sort of the roadmap to address a multitude of markets. We're very focused on defense, on mobile consumer electronics and AI data comm, but this is broadly applicable.
I've spoke somewhat at length about the opportunity in mobile and consumer electronics, and very interesting things happening in that industry today to adopt SWIR, whereas in the past it didn't seem feasible. Data comm and AI infrastructure, if you're following the kind of investments that the hyperscalers are making, the kind of investments that NVIDIA has been making in optical, that's a space where we're very active. Th is is for AI infrastructure, data centers, high-performance computers, and also traditional telecom and wireless communications. Our technology is very applicable to defense and aerospace. This is for communications or sensing, for high-definition imaging, for LIDAR, comms, surveillance, and autonomous systems. There's an example of a very large format camera on a gimbal on an aircraft.
The technology is also applicable to quantum. Again, we work with these high-performance compound semiconductor materials. We know how to scale them, and that's very important for some of our quantum customers. So, a moment on AI data comm and the investments that have been made in this industry, more than $300 billion in 2025 only from the top four hyperscalers. So, that's Microsoft, AWS, Google, and Meta. This investment, from what I hear, is gonna approach $700 billion in 2026, might fall a little bit short because of some supply chain constraints, and growing to more than $1 trillion in 2029. These are not just sorta guesses or forecasts.
These are based on projects of record already in place with some of the hyperscalers and could grow significantly even beyond 2029 and 2030. Optical networking. So, ho w do you connect all those GPUs and CPUs and XPUs? How do you connect the compute nodes to memory? When you get to the data rates that AI workloads are demanding, that can't be electrical anymore, that can't be copper, and you can't carry the data over a certain distance and enable the low latency that's required to ensure that all of those servers sort of act in unison as one computer. Around 15% of that investment is for photonics, for the kind of components that Aeluma and Tier 1 and Tier 2 suppliers build.
This is a huge opportunity, one of the biggest we've ever seen for photonics. You've probably seen in the news, NVIDIA is not a hyperscaler, but very important in this ecosystem. They've made investments of $2 billion in Lumentum, $2 billion in Coherent, $2 billion in Marvell, also $3.2 billion in Corning. That's about components like lasers and other photonic components. That's about some of the packaging and the fiber that's required to connect some of these compute nodes. What's happening is, demand is completely outpacing supply for traditional pluggable optics modules. Those are the modules that plug into the faceplates on these server racks and connect each of them in the rack or across the data center with optical fiber.
Then, there's investments being made in next-generation optical interconnects, what are known as near package optics or co-packaged optics, taking the optics inside to the server rack to further increase efficiency and reduce latency. These are technologies you may also be hearing about in the mainstream now for architectures such as slow and wide or fast and narrow, micro-LEDs, micro VCSELs, silicon photonics, EMLs. These are externally modulated lasers that are built by some of these Tier 1 suppliers. Thin film lithium niobate. These are mostly on the laser and modulator side. On the receive side of the link, you need the photodiodes, and that's where Aeluma excels because in addition to building these photodetectors for sensing applications, we build the high-speed versions of these that can be used in optical transceivers.
As I mentioned a couple of times, what's happened is major indium phosphide substrate suppliers are out of substrates, in some cases sold out for five years. It takes time to increase capacity. The suppliers didn't believe the requests for quotation that came in over the last year, and so, t hey didn't increase capacity and couldn't do so quick enough. Now, there's a shortage of substrates. There's a shortage of indium phosphide fab capacity. It takes time to build this all out. There's even shortages of some of the packaging components, and that's why you're seeing investments being made by some of the major players in the ecosystem to secure some of that supply. Aeluma's non-indium phosphide photonics, that's the key. We don't use those indium phosphide substrates that all other suppliers use.
This manufacturing can provide performance and scale that's critical for components like InGaAs photodiodes, quantum dot lasers, and high-speed modulators, and these are all in very high demand for the data center applications. We only got a couple of minutes left, so let me just walk through briefly what the products look like that we're building. In some select cases, we do sell our base wafers, especially in the case where a customer wants to leverage that for a market we're not necessarily planning to become vertical in. These large area InGaAs photodiodes that are used for basic sensing applications and power monitors. The high-speed photodiodes and arrays that are needed for those slow and wide and fast and narrow applications in data center transceiver interconnects. InGaAs imaging photodiodes, that's a fairly large shortwave infrared camera more suited to a defense application.
Imagine a smaller version of that for mobile and consumer electronics application. Quantum dot lasers that are used both for sensing and communications applications, and also our III-V materials for quantum nonlinear photonics. I can't spend a lot of time on each of these, but just to give you a snapshot of what some of these look like, these different templates that we build for our own products and for some of our customers' products. These large area photodiodes that you can buy from distributors that source these from suppliers that use indium phosphide substrates. The difference being here we do this on silicon substrates. We can scale to much larger volumes, and we don't need those starting indium phosphide substrates that the traditional suppliers can't find. High-speed photodiode, same.
Built on non-indium phosphide substrates, and that can enable data center transceivers in these very high data rates that are expected to deploy over the next coming years. Little bit more detail on the imaging photodiode arrays. If you look carefully, there's some tens of thousands of pixels in that imaging array. That's essentially an infrared camera that leverages our technology. The quantum dot lasers that we build on large diameter substrates and, in some cases, directly integrated with silicon photonics, which is key for the data center market. Lastly, the quantum photonics application and the nonlinear elements that are critical, especially to quantum networking and sensing applications. I'll close by just sort of touching on our facilities and capabilities in this capital light manufacturing approach.
We have this MOCVD capability in-house where we produce the wafers all the way up to 300 mm or 12 in in size. We do some proprietary cleaning steps in-house. We can do quick turn fab rapid prototyping and small volume manufacturing to develop our materials and refine them before we send those substrates to our medium and even large volume fab partners. We haven't shared all our fab partners, but we made some announcement recently about a relationship we have with Tower Semi for front-end fab and with Sumitomo Chemical Advanced Technologies for scaling our substrate manufacturing. As I mentioned, we're north of here in Goleta, California, r eal ideal ecosystem for development and commercialization. Number of companies in sort of defense infrared, AI, quantum, in the same location as us.
Some more details of our facility and the equipment inside that includes the deposition tools as well as the test capability. We can test everything from chips, packaged components to chips on 6-in, 8-in, and 12-in substrates with semi-automated and automated wafer probers. This enables us to do sort of full-scale tests when the wafers come back from our fab partners. Thanks very much for your time. Happy to take some questions.
For your InGaAs or some of the things that you're making, is it a die-to-die replacement or does the end customer need to change their firmware or make other developments? How are they reaching this?
Good question. The question was, for our technologies, it's sort of a die-to-die replacement or does customer need to change how they do the integration and the firmware and so on? It's a little bit customer- dependent, but generally, this is a drop-in replacement. If a customer is already using, say, InGaAs, our technology's a drop-in replacement. It's just on a different base substrate, so we can build in larger volume and at lower cost. If a customer is using a different technology, for example, I mentioned that in the mobile application, silicon detectors are used today in the near-infrared, so InGaAs would be a different technology, so that would require a little bit more reconfiguration. But it's very enabling 'cause it's adding functionality to the system.
It's a little bit dependent on whether the customer already uses InGaAs or if they use another technology that has sort of subpar performance.
[audio distortion] where there's quite a lot of shortages as well with InP chips, those are [audio distortion]?
No. Anywhere they're, where they're already using indium phosphide substrates, it's essentially a drop-in replacement.
How quickly can you scale up?
We can scale up as quick as our fab partners can. You know, there might be some CapEx investment required for some of our partners to scale, but i n the near term, they have pretty significant capacity already. The idea is leverage what's there and don't build a fab and worry about filling a fab. We've got good partners for doing that. Yes, please.
Speaking of, do you give the recipe to your fab kind of contractors?
Today we do not. There's a few ways to do that. One, we have a fragmented supply chain, so w e don't have one partner that does everything.
Yeah, you bring the cost down?
Yes, of course. We're doing this with large volume, supply chain partners.
Yeah.
But we do different steps at different sites so that no one has a good view of the entire picture.
Yeah.
In some cases, we use consignment. The idea being we load the recipes and the partner doesn't actually see the detailed recipe. We do our best to hide a lot of the detailed information. I mean, patents are helpful. This technology compared to some mainstream silicon technology is very hard to reverse engineer. That adds a little bit of level of protection. You've got a bit of a head start. Yeah, we're very mindful of sort of protecting the technology.
[audio distortion] today we're doing everything in the United States. That might not always be true [audio distortion] . Can we come see you?
Yeah. I'm gonna step off 'cause I'm being told to leave, but y eah, p lease reach out to us i f you'd like to learn more. Thanks very much for your time.