BrainChip Holdings Ltd (BRN) Earnings Call Transcript & Summary

It is my pleasure to welcome you to BrainChip's Quarterly investor update for the second quarter of 2026. My name is Troy Dahms, BrainChip's Head of Investor Relations and Financial Communications. And I'm joined today by Sean Hehir, BrainChip's CEO. Good afternoon to you, Sean.

Hi, Troy. Great to be here with you today and more importantly, with all of our shareholders.

Thanks, Sean. The format for today includes a short presentation from Sean outlining BrainChips strategy and how we continue to execute on that strategy, along with a summary of recent activities. At the conclusion of the presentation, similar to our podcast series that some of our investors may be familiar with, I will ask Sean to address a number of questions that we've been asked by shareholders during the period. Over to you, Sean.

Thanks, Troy. I think it's always important to start with the market that you serve before you talk about the company because it sets the appropriate context. And I'm going to start on the right-hand side of this page, if you don't mind, and talk about some of the markets, and we'll talk a little bit later about use cases and companies and things like that. But on the right-hand side, you're seeing is a whole large market, what they call Edge AI, which is markets sort of AI outside the data center. I believe all of our listeners are using AI in their daily lives around the world and are very familiar, but most of those implementations are in a data center, things like Claude or ChatGPT. But the use cases and the type of flexibility associated with the edge market is a market that started later, but it's maturing very rapidly at this moment in time. And that's what you're seeing on the right-hand side is these kind of markets. All these markets is where BrainChip serves and others as well in Europe. Now what constitutes the market, it's basically the ability to do inference outside of the data center without moving things back to the data center. And there are some really important things that drive that kind of why that matters. Some use cases, you just can't do it in the data center or it doesn't make sense because it takes too long with latency. Sometimes people are concerned about privacy because when you move data around, it could be a concern because somebody could take your data or perhaps even if you're not concerned about latency, but you're concerned if you lost a network connection, if you're making a decision in space. You're making a decision in a wearable device on somebody's health. You're making a decision in defense situation. You cannot afford to lose a network. You need to do your inference locally. And if you're doing that, you're also -- there's an imperative to do that on a battery typically. And so if you want to do it in a battery, you got to go ultra-low power. And you're going to hear me talk a little bit in the next slide about BrainChip. That's where we excel is ultra-low power. So the brief introduction to the market is the following: very large, growing markets coming to bear right about now, later than the data center market to do inference outside the data center with some real strong drivers that are getting clearer every single day. Next slide, please. So if a company is going to do inference or AI outside the data center, they have some choices to make. And one of the critical question is, why would you choose BrainChip? First and foremost is that top left-hand box that says, first, to commercialize neuromorphic. Neuromorphic is an approach that's different than the rest of the industry, which is traditional kind of brute force multiplication or the von Neumann architecture, something that's been around for many, many decades. This company was founded on a vision from 2 founders was about mimicking the human brain, which is the most efficient computational engine on the planet. And basically, your brain ignores things that don't matter. If nothing is happening, your brain sits idle, but it only fires when something is happening. That's what neuromorphic is or in the parlance of computer world, 0s and 1s. If something is a 0, we just don't compute. So that's how we get these attributes of very low power. We're the first to bring this to bear in a commercial form in all kinds of form factors that I'll describe a little bit later. As an investor, you should be very concerned and rightly so, and we're very proud to talk about our patent portfolio. We have a very strong portfolio of critical elements of this technology and those patents have continued to grow every single quarter. You want to make it easy. One of the very interesting dynamics about technology that I've been in technology a long time. All companies are striving for greater breakthrough, defining in this case, lower power, lower latency, high-performance and compute. But you got to make it incredibly easy to adopt. What I drive this company to every single day is the ability to get that breakthrough performance but easy for our customers to deploy, whether that is putting their models onto our hardware, deploying our IP into their custom chips. It's all about ease of adoption. And quite frankly, it's a customer centricity I drive through the workforce as well, the ability to make things easy for it. I think from a customer point of view as we design that for customers and also all of our customer interactions, customer support. And then of course, there's a very interesting thing that's happening in the edge market right now about hybrid compute. And we're going to have an opportunity to talk a lot more about this later in the presentation, Troy. But the edge market really was established on discrete individual kind of tasks with hybrid compute, we're seeing it come together the ability to aggregate chips and to pool and accomplish greater things is those kind of attributes that set BrainChip apart from anybody else in the edge market. Next slide, please. So I talked about just why BrainChip, and I want to illustrate some really important elements on this slide. Instead of walking through the individual pieces of our offering, I just want to kind of put it in the context of the strategy of the company. When I joined the company, we had a much smaller set of offerings for our customers. It was a single offering IP. As of this broadcast, customers can now buy A few months after that, they'll be able to buy 4 levels of IP to address different types of models, different requirements that are getting clearer every day in the edge market, listening to our customer feedback. We also now have a plethora of hardware offerings, which you can see right here as well. And these hardware offerings allow us to address other customers who cannot afford IP. So the point of all this without going through all of the individual parts of this is to show that we got into the market early, listened to the requirements of the market, expanded the offering based off of the customer engagements and the feedback we're having to address a broader set of use cases technology-wise and form factor-wise. This is part of the multiyear strategy of the company has been on and we're executing exceptionally well. Next slide, please. I'm often asked -- and we gain traction. How does it look on the commercial front? And what I'm trying to share today is a little bit of insight into that. Our funnel is incredibly strong. It continues to grow quarter-over-quarter. These are 4 stages of typical funnel management when you're in a sales organization. Not surprisingly, those that are really early in the conversations as more as you get down to negotiations, you can get down to less and less the real serious buyers to go through it. But a really important message in here is it illustrates some of the sectors that are most important and growing well with us. I'm often asked, where do you get the most traction? I mean you can see by looking at this that the defense vertical and the wearables are something that are getting a lot of traction when I say wearables, medical health care. Now there's some reasons for this. These markets have unique and strong requirements. With the state of the world today, the defense world has changed dramatically from more kind of fixed position kind of things to something that requires mobility. If you're doing mobility, you have to do things that you couldn't do before like AI inference in a mobile platform. But if you're going to do that and you need to do it an edge platform. So the whole defense world has changed dramatically about how they're going about conducting their business and the ability to do a AI activity on the edge is growing exponentially. No surprises a lot of activity in that sector from us. Wearables by definition, if you're going to do something and you got to have on your body, on your face, you need long battery life. You need long battery life. You need a company like BrainChip where neuromorphic technology that allow us to have incredibly long battery life. Next slide, please. I'm often asked about how we position ourselves against competitors. And the right way is the simplest way to look at this is this little graphic right here. The edge market can be defined as something or anything outside the data center. And you can see we address or this market could be something very close outside the data center, which is very large, high number of TOPS, which is a term talk about kind of compute power. But in those kind of examples you get incredibly high power signature. That is not where BrainChip is competing. We go to incredibly low power situation. You can see sub-watt slightly over a watt and a more modest kind of compute -- that is where the growth is in this market in the edge market. And this is where we can differentiate ourselves considerably. There is virtually nobody in this market that can do the things that we can do. It's a segment we occupied uniquely. Next slide, please. Another question I get is let's talk about some real deployments. And I really could spend a lot of time in here and talk about this, but I'd highlight just a few more to just highlight the attributes of the product and why it's so important for the edge. The top left Onsor I talked about medical devices. This is a company that's creating seizure prediction glasses for people who have epilepsy. In the past, people of epilepsy often could not even leave their home. They were afraid to drive because they were concerned about having a seizure. This company, using our chips, has created a solution on a pair of glasses on some of the space that can predict within 1 hour with a 98% accuracy if they're about to have an epileptic seizure. This creates a whole potential freedom of life for people and very, very powerful. I'd love to share that example about the kind of the innovation and creativity going on at the edge. I can talk about some of the things that are happening perhaps around radar and drones with our partners like Raytheon. And I mentioned earlier about the traction in defense. When you think about defense and radar traditionally, radar would say we sense something is coming, but it'll lack the ability to do classification and it certainly lack the capability to do this mobility. Our type of solution developed with Raytheon and AFRL, the Airforce Research Lab, allows us to say, yes, something that's coming and classify it, which is very powerful in a radar situation. And then I just want to pick on 1 or 2 other really important things that are important for us because if I highlight some of the other capability. I mentioned some of the IP levels that we talked about. We have a second-generation IP, which we have very high focus on its capability. We have a license, our first major licensee, Edge AI, which is on track to tape out their chip later this year. So we're very excited about that. Next slide, please. I'd like to highlight a little bit about Q2 and some really important things that happened. We talk about hybrid computing earlier. I'm going to the middle part, and I want to talk about the IBM Symphony community bundle. Here is a wonderful example, not the only example, but a wonderful example of the work that's going on to work in a kind of a hybrid computing model. This was some breakthrough type of work. The work continues on. We work closely with the people that are doing this. and there are other companies that are out there doing more hybrid computing right now. Neuromorphyx, over there on the top right, very, very important kind of customer for our chips as introducing some very cool modules. So it's another kind of downstream potential distribution for us and creates more modules for customer choice point. And then some of the announcements on the bottom -- on the bottom right, around Edge AI, micro IP and ASICLAND. ASICLAND is a really important relationship for us. Think about it as a channel relationship. In that case, they have bought some, what we call MPW licenses from the company, which are a smaller company like trial licenses. And as they're out talking to their customers, ASICLAND is a design shop, they'll deploy those licenses and that creates potential other opportunities for other downstream licensees. The Akida 2500 tape-out is going incredibly well. And then I think to maybe highlight a little bit on the Akida 1500 chips because I'm often asked about that. It's a very important revenue stream for us. You saw some of the modules we had earlier on the previous slide in there. and we'll talk a little bit further on another slide about the traction we're seeing in there. Next slide, please. I highlighted just a moment about the Akida 1500 chips. This is critically important for the company. We've had great uptick since we placed the original order with 25% of the chips already sold out. We're also spending part of that inventory to build out the modules we had -- we talked about earlier. Those modules are critically important for future sales because it allows people to buy volume and plug and play. We have tremendous interest in some of the reference designs particularly around Akida design. And of course, the pipeline is growing very, very rapidly now that we have chips in-house. So a couple of customers that I'd highlight is one I mentioned earlier was Onsor and Neuromorphyx, but perhaps I'll highlight one more, which is Parsons, a very large defense contractor who's doing some really interesting things with more than one chip on an individual board. And so it advances not only the chip orders but the hybrid computing model. they took an initial order of several thousand chips, and it's the first of a multiyear partnership, and we look to see that grow considerably in the years to come. Next slide, please. I've already touched about hybrid computing a couple of times, but I think this is such a really interesting thing that's going on in the future is really bright for this. It's the ability to do this true work between Akida, CPU and GPUs, other NPUs or whatever that is because it makes all the sense in the world. Without getting overly technical, specialized inference is really what's driving the industry. the ability to put the right platform, the right tool. And that most of these things, when you look at this graphic or you read some of the posts on there. Usually, it's right next to the sensor where the Akida goes, make some quick decisions, some of the larger decisions go up to a centralized CPU for bigger decision, but it's really the right tool for the right job. We continue to learn every single day and the use cases look limitless for this. We have very high expectations for this kind of use case going forward. And with that, I look forward to taking a lot of questions. I know we covered a lot of material in a short period of time, Troy, but let's take some questions.

Yes, sure. Great, Sean. Thanks so much for the update. As said, we'll now turn to the question. So if you're fine, I'll just fire away at you. In terms of commercial success, what verticals or sectors are likely to drive the most demand for BrainChip products in the near term?

Well, I think we touched upon it in the presentation, which is certainly the defense organizations and the wearables because the demand is so strong there right now. The requirements are just lined up perfectly with our product. But I think probably what's more importantly than picking on the ones that we're seeing the traction right now is where are the ones that are building. We have very strong interest in all type of verticals, one of which would be, let's call it, IoT, Internet of Things. The world is waking up for the power of inference on the edge whether it's preventive maintenance or monitoring and things like that. But the ones right now are radar and drones, but there are many more that are coming right behind us.

Yes. Great. The second question we got, and I've sort of amalgamate a few questions from shareholders. But on the Akida 1,500 silicon chips, it appears that the yield is lower than expected. What is being done to address this? And when do you expect to have answers? This question came from Rajesh, Tom and Matthew.

Sure. I think the important thing to note from the beginning, Troy, is this is not an uncommon issue in the silicon world. There are many variables that come to a manufacturing process that can impact the yield process. BrainChip and our design partners and manufacturing partners are analyzing every element of this to find exactly what it is. We have very high expectations that this will be resolved to a satisfactory level. Our MPW had very high yield. So we're just isolating it, walking through the process. It takes time, but this is not an unusual problem. We have very high expectations as we resolved to a satisfactory level.

And I guess, probably good to let our investors know that we'll have more on that in our third quarter update when that comes out. The third question, there is a strong interest in defense and humanoid robotics right now. Can BrainChip benefit from these segments of the market?

Of course. I mean it's similar to the first question. But think about it, robots is very similar to drones and things like that. A robot is -- and obviously, defense is drones require mobility. It's just by the nature or you don't typically have fixed robots or I'm not sure it would be a robot. A robot is mobility and so when you have mobility, you need the ability to do things typically on a battery or at least even if you don't, even if you're on a power source, you want very quick decision-making and quick actions going on. And so these are ideal use cases for our kind of technology.

Great. And what is the revenue model for Akida, given you are now producing silicon chips? Are you moving away from Akida IP?

Absolutely not. This is one where we could talk a lot about but I know our time is limited, but I'll try to do this succinctly. I touched upon it in the presentation, some of the largest companies in the world will want to build custom computer chips or inference chips or NPUs. In order to do that, you've got to have a pretty large capital budget. You've got to have large technical resources to do the planning and designing so. For that, we will offer IP. And we think that's incredibly important for a couple of reasons from the business model. One, it allows you to get out there with that, you get a perpetuity, the royalty as the chips come off the line. So we -- it's foundational that we offer IP. That allows us to address the segment of the market in a really critical fashion so -- to do that. I also think it's important to know the trends of what's going on in the industry. big companies are trying to specialize silicon more and more. If you think about the competitive world that they operate and the ability to get more performance is usually typically coming down to tuning models and tuning silicon. So we see the trend to more custom whether it's ASICs or SoC to continue. So IP will be a critical cornerstone for them. Having said that, offering silicon is equally important because you've got a whole segment in the market who doesn't have the capabilities. So doing the way we're doing offering both are additive. It allows you to address the really large companies and maybe the midsized companies and even the smaller companies when you have silicon, and not to make this overly complicated, too, sometimes even the larger companies that want to build a chip will want to buy some computer chips from us first to say, let's try and buy because we're not going to build something exactly like the Akida 1500, but we can make some modules send it out, get customer feedback. So it's a very supportive virtuous circle where they can buy some chips, and we're seeing that on some customer engagements right now, building some prototypes putting it out, getting customer feedback as they refine their final designs on an IP selection. So IP and chips together and modules is critical, fundamental to our well-planned strategy, something we have envisioned for a period of years. We just have to get to that point where we are now.

Great. And as you say, it's all about customer choice. The next question is a hot topic, obviously. With $20 million in cash at June 2026 and the current burn rate, what is the company's funding strategy? And associated concern is with the number of shares on issue and potential for a capital raising. This comes from Devin, Ryan, Pawan, Gary, Dennis and Jacob. Did you want to make some comments on this, Sean?

Of course. I think my opening comment would be around, let's call it, capital in general. As you know, I report to our Board of Directors. We do a strategic plan every year, at the end of the year. And the Board has set requirements around not only the revenue trajectory or the booking trajectory, but some goals around cash flow or cash burn neutrality over a certain period of time. And the Board has expectations that I will take the company to that. As and if we need capital raises in between, Ken, our CFO, and myself, will examine the best options of the best flexibility of the best rates as possible and make that presentation to the Board. But my opening comment is really important. We have line of sight and a plan at some point to make the company cash flow neutral. And in the meantime, we will examine the right options at the right time and bring it to the Board.

Right. Question, next one comes from what is BrainChip view on Kevin Johnson and why aren't we doing more with him? This comes from Paul Roger, Glenn and Mario. And I'm sure you've heard a bit about this, Sean.