BrainChip Holdings (ASX:BRN) June Quarter Update, August 2020

Company Presentations

BrainChip Holdings (ASX:BRN) President & CEO Lou DiNardo presents the June quarter update on the the company's product and business development timeline and financials.

Lou DiNardo: This is our June quarter update, our management presentation. I'm Lou DiNardo, the President and Chief Executive Officer, as I'm sure most of you know.

I'm going to slide number two, giving you a little bit of a sense of our agenda. We'll talk about our June quarter accomplishments, the product development of Akida, which I think you're all going to be very excited about, as well as the market's opportunities and challenges, as all companies face, a bit about the state of our business and what we consider our primary mission, which is increasing shareholder value.

So I'm moving on to slide three. We ended the June quarter with $5.3 million in cash. I'm sorry. I should always specify those are US dollars. We had a couple of favourable events that occurred during the quarter, where we got some money from the US government, we got some money from the French government. All money is good money. So we, again, closed with $5.3 million. We have continued to maintain very strict cost constraints. So I think it was about -- I'll use round numbers -- about $1.9 million in operating cash burn. So take capital expense out of that. But that was basically where we were at the quarter -- $1.8 something, and I'll round it up to $1.9 million.

It's a very exciting event. We completed wafer fabrication. So Akida, it's real. I mean, it's a piece of silicon. The engineering team is excited. After years of development on the part of the design team and Peter's vision for what Akida could and should be, wafer fabrication was complete. Assembly and test operations -- assembly is done. Test operations are really separated, and I'll talk about this in a few minutes, really into two phases. One is when our manufacturing partner, Socionext, gets the product back from assembly operations. They do really what's manufacturing tests to make sure you don't have shorts, you don't have other problems. I don't want to go into too much detail. Scan, memory BIST, just to make sure everything works on a logic level. Again, I'll talk about where we go from here in another couple of slides.

We have an early access program, which I'm sure everybody's aware of. We've got some major names. We've talked about Ford publicly. We've talked about Valeo publicly. We're being somewhat selective, frankly. We don't want a bunch of research or students being part of the early access program. So being very selective with respect to those companies that could be commercially successful, successful with us, and of course, successful with themselves in the marketplace. And maybe more importantly are these proof of concept agreements that we have, where people are stepping up for a little bit more money. So early access and proof of concept all include, you got to pay to play. We're giving it away for free. You need to bring a cheque. I'll talk about it and I’ll address certain questions that came in. If it's a light-touch deal, maybe it's less money. Maybe it's less access to evaluation boards or the number of devices that are delivered. Proof of concepts are significantly more with respect to cash that we get in the bank, and it will take more work.

It's a significant burden on our engineering team. Anil Mankar, our Chief Development Officer, shoulders much of that burden. You'll see that we've appointed two new... In addition to Adam taking over, dropping off the board, taking over as the chairman of our Scientific Advisory Board, Simon J Thorpe. Many of you know his name. He was very much instrumental in the founding of SpikeNET, which we acquired in September of 2016, and as important and I think a very big testament to where we're going with respect to smart health, is Barry J Marshall, a Nobel Laureate in science and physiology.

We’ve retired the convertible debt completely, which I think was a great achievement. It did allow us to get the Socionext deal in play. I think, in the end, it was a good deal all the way around. We did a small placement of, again, this is now Australian dollars, $3.15 million in ordinary shares, turned out to be about $1.8 million in round numbers in US dollars. So, overall the quarter was pretty darn exciting.

I'm going to move on to the next slide if I can figure out how to do this. For all of you that have been waiting, this is the money slide. This is Akida. These are actual live devices. They're sitting in a tray. Don't be bothered by the bottom left-hand corner. That's just the tray waiting for another chip that needs to be tested. But you can see BrainChip as a logo, the AKD 1000 dash ES, which is “engineering sample”. This has not been through all the production worthy, three lots, burn in, everything that's associated with it. But I have to tell you that Peter, Anil, the entire engineering team, those that are in Aliso Viejo, those that are in Toulouse, France, our software team that's in Hyderabad, are very excited. We now have silicon in hand, and I'll talk a little bit more in another slide or two about when that silicon and evaluation boards end up in the hands of customers. But I'm going to pause here just for a second, because this is pretty darn cool, guys. It's been years in the making, but BrainChip and the Akida device are now a live and real product.

So where are we with respect to AKD1000, which is the first generation of this device, of this processor? We've completed or we're completing, I should say, a functional test. I would say it's complete. I think there's some transit time between what goes on in Japan and what gets back to us here in the US. Internal validation will be underway very shortly. It's quite a bit different than the functional test. The functional test is doing scan, doing M-BIST, making sure that all the transistors work the way they're supposed to work. Internal validation is looking at all of the individual blocks and making sure that all of the interfaces, the CPU complex, the neural fabric, all work together. Evaluation boards are in fabrication. And as you have heard from us before, the early access program and proof of concept engagements will start in probably August of 2020. So, it's a real device and it's going to start to hit customers' hands.

I think it's important that we step up a little bit from talking about smart sensors and this sensor or that sensor, and really describe to investors, where are the focus markets that the company is addressing. And it really comes down to four or maybe five. Maybe one is missing here. Smart home, which is video analytics for speech, speaker recognition, gesture recognition. This is all the kinds of stuff that's going on in people's homes. Smart transportation, which is advanced driver-assisted systems or assistant systems, autonomous vehicles. Smart city for far edge surveillance. Smart cameras, enterprise edge surveillance, where you have, whether it's, call it unmanned aerial surveillance or enterprise servers that are doing the video analytics that are acquired from the data streams or video streams that are coming from smart cameras, and really a new and burgeoning opportunity for us is smart health.

It's called VOCs, volatile organic compounds. Whatever you exhale as a human being tells a lot about what's going on in your body. Whether that's cancers, early detection of cancers, ovarian cancer, breast cancer, other types of cancers, viruses. In this case, there's certainly a big focus on Covid-19, but the ability to detect by basically what is a breathalyser, viruses or cancers and a very large market is blood glucose. You have type one diabetes, T1D, or type two diabetes. So at a very high level, and I've tried to bring this up to a higher level than getting back into the weeds. Smart home, smart transportation, smart city, smart health and probably the next bullet on this slide should be smart security, where we're addressing what's going on in the cyber security arena.

I do think, you know, the learnings that we've had over the past year with the Akida development environment being in play, with all the customer contacts that we've had, building out and reorganising our sales force, there are two real different edge environments. One is far edge. You're really at the edge, you’re where the sensor is, smart sensor analytics, vision, acoustic, call it olfactory, which ties back to what I just mentioned, about the exhale or the ability to sense volatile organic compounds. Lidar, radar, ultrasound. A very large and emerging market for 3D point cloud. This has been historically looked at as Lidar, but it's becoming a far more robust marketplace for 3D point cloud for gesture recognition and a variety of sensors that are outside of Lidar. And physical inputs, which include pressure, temperature, vibration and flow. Those are really far edge AI opportunities.

Then you've got the enterprise edge, which is not in the data centre, it's not in the cloud. It's an on-premise analytic for... Object detection is one category, object classification for another. Could be a very, very wide range of applications. I think it's important to recognise that, at the far edge, it's more likely than not that these will be intellectual property licences to customers, potential customers that are building their own SOCs or system on chip, because of the space constraints and a variety of reasons that they can't afford to put another IC into a camera, into a radar detector, into a radar module, ultrasound or Lidar. Those are more likely to be than not, intellectual property sales or licences. When you're in the enterprise edge, you've got a lot more room, you've got the ability to dissipate heat. So those are really where the Akida device opportunities are probably going to be more prominent.

Call it challenges and opportunities, really should be opportunities and challenges. This is a big change. It's a big architectural change. There's no CPU required. There's no external memory required. And there's no dedicated matrix multiplication or dedicated DLA or deep learning accelerator. That's an opportunity, both for us and for potential customers, in that they can reduce the footprint, they can take advantage of everything that Akida brings to the table with respect to ultra low power, higher performance in most regards. But it is a learning exercise, and it's been a learning experience for us and it's been a learning experience for potential customers. But what we have found is that major multinationals, and we've talked about Ford, we've talked about Valeo. We've got others we've talked about, one of the world's largest storage manufacturers in the world, storage and server manufacturers in the world. They're all attaching themselves to the idea that they can offload the CPU. They don't have to use external memory. And they can use either Akida IP or Akida as an IC so that they can take advantage of the best of both worlds.

Again, challenges and opportunities. Customer communication. It would be silly to think that it's not a challenge right now. I typically travel a great deal. Anil travels a great deal. But travel restrictions certainly make this a challenging environment, but I think we're getting through it quite well. Folks are getting used to Zoom meetings. They're getting used to WebEx meetings. Internally, it's really not been an issue because we're somewhat of a distributed organisation to start with. Our facility in Southern California, our facility in Toulouse, France. There are two of our directors in Australia. Peter now having the Research Institute in Western Australia. So we're working quite well through the internal communication, and I think we're doing quite well, actually, with customer communication. But face to face interaction is important. It does put a lot of pressure, and I think it's good pressure, that we have high-quality collateral material, and we try to avoid, or at least clear up, market confusion about really what's going on. We're not a deep learning accelerator. We are an independent AI processor. And that is a message that we need to make, and we need to make it forcefully.

State of the business. Intellectual property, I think has been a really robust environment for us. I would hearken back to this slide, which I call the money slide, with the picture of the IC itself, because intellectual property sells when potential customers are looking at spending $30, $40, $50 million in order to build their own SOC. Simulation is great. The Akida development environment is great. But having silicon that they can actually run the network on, an evaluation board, I think that is really going to drive the possibility of intellectual property sales in a much more robust way. The Akida device, as you've seen, it's alive, it's well, it's real. I'm hoping actually to have one in my hand some time very shortly. So, intellectual property is moving quite nicely. The Akida device will now be one of our primary focuses.

Again, we're reconstituting our sales effort, and I expect some time in the next week or two, that we'll make an announcement about a sales executive that's dedicated to the selling process. Up till now, it's really been business development.

When we think about shareholder value, and there's a number of questions that I do want to touch on. I've tried to address many of them in the deck, but as we look at shareholder value, I think it's important to recognise or at least reiterate, that we are the only publicly listed pure-play AI company, at least that I know of. Again, I've said this many times, you could buy Google. You could go from Google to Tesla, to anybody in between, but you buy the whole bag. It's unique that we are publicly listed and shareholders can enjoy the increase in shareholder value that will come with our commercial success.

And it's also unique that we're listed on the Australian Stock Exchange, not NASDAQ, not in the New York Stock Exchange. So, that ability to participate is unique. And frankly, our current market valuation, market cap is at a discount to private companies. In some respects it's because shareholders come in, they come out. We have a good share registry. There's a lot of questions around... I'm going to move to a second screen of mine. Questions around what about capital, is the company adequately capitalised to fulfil the promise? We look at this every day, and we look at increase in shareholder value. What does it take to build a large company with an enduring brand that has a great multiple on sales, EBITDA, whatever multiple you want to apply to it? And we look at that constantly. But our commitment is that our focus is on increasing shareholder value. So over time, yes, we're going to look at different alternatives as to how we capitalise the company, how we continue to progress. But the major focus is to build a large company with an enduring brand.

Frankly, I got a lot of questions, so I tried to cover most of them in the presentation, but let me go through them just to make sure, because I think sometimes people feel like I've not answered specific questions, and if I don't, just send me an email. I'll just go down the list.

Do we have a working chip in action? Yes, we do. It's in manufacturing, I've shown you a picture, and it's not a Photoshop. That's a real IC. It's got our logo on it. It's got a date code on it. It's completely traceable.

How many companies would receive the evaluation devices? As I've said, we've been somewhat selective in the early access program. We don't want to just sprinkle things around to research institutes. We're really focused on what could be real commercial opportunities, significant commercial opportunities. So, a couple of dozen is probably where we'll land.

IP sales, when does the company expect some revenue? Frankly, we don't know. That is, I think that now having the IC so that people can validate what they see in simulation is going to be a big catalyst.

Are there any outstanding invoices? Of course there's outstanding invoices. I'm waiting on purchase orders from early access customers.

Is Akida considered complete? Yes, it's considered complete. And is it functional? Well, as we go through testing, we'll find out.

How many commercial partnerships does the company have? You know what? I hesitate to put a number on that. That's really very competitive information. But the early access program has been really successful. And these are household names, every one of which you would know and appreciate.

Large scale production? Certainly some time in 2021. Engineering samples will be available very shortly. The ADE provides customers what they need to proceed.

How will we be funded? Again, our goal is to support shareholder value and access capital commensurate with the company's needs, but to build an enduring brand and a big business so that shareholders can enjoy that increase in shareholder value.

Will we see dividends in the near future? No. Dividends are really related to much more mature companies.

Should we expect a consolidation of shares? We've had no discussion about consolidation of shares.

Does AI have a future with the economic depression? Actually, I don't know about the economic depression, but yes, AI has a huge future. It addresses so many... It's a horizontal market that's very wide, but it also addresses very deep vertical markets, and probably will be one of those drivers that gets us out of what might be considered an economic depression.

Of course the webinar will be available after the event. It will be posted on our website.

Akida 1.0 pricing, has it been established? Yes, but we're not going to talk about that publicly. That'll be private conversations with customers.

Can Akida interface with quantum computers? Yeah, there's no reason to think not. It can interface with virtually any computing architecture, so long as people want to take advantage of its capability.

Is the synaptic bus a limitation in defining the effective synapse and neurons? I would leave that for Peter to comment on. Frankly no, I don't believe so. You can make the buses as big and wide as you want.

Is BrainChip a takeover target? I would just say who really knows? But rest assured that the board is always cognisant of protecting and increasing shareholder value. So if somebody wants to come along, write a big cheque, that's okay.

Iterations of Akida and geometry. Are we thinking about five nanometre? Look, we're at 28 nanometre now. It's straight, flat out, digital logics. Anybody that's using our IT, where they're building their own SOC or system on chip, they can scale to any geometry they want. It's all very simple math. You go from 28 to 14, just divide by two, you get smaller die area, you get lower power. If you go from 14 to seven or five, you can do the same math.

Many of the questions are somewhat redundant. Again, there's a question here about developments in healthcare. I think it's a very ripe and target-rich environment for us. Again, volatile organic compounds tell you a lot about what's going on in your body, but those sensors don't do the analytics, and we can bring that to the table.

There was a question about a contract with NASA, which I think got a lot of attention. We have no contract with NASA. We certainly have lots going on in that arena. But wherever that came from, I'm not sure.

So, look, I've tried to cover all the questions. I'll move to the thank you slide. If there are any that I've missed, just send us an email. You can do it to the BrainChip update email address. You can send it to me directly. You could send it to our CFO, Anil or Peter and then we will get back to you.

So, thank you for your support. It's been an extremely exciting quarter. We've got chips in hand, and they'll soon be in the hands of customers. So thank you and we'll talk again in about 90 days.


Ends