Interview with Brainchip (ASX:BRN) CEO Sean Hehir

May 25, 2022

Brainchip, BRN

BrainChip (ASX:BRN)

We spoke with BrainChip (ASX:BRN) CEO, Sean Hehir, about the company’s partnership with ARM, it’s commercialisation strategy and the many opportunities BRN is seeing in the market.

Check out our most recent article on BrainChip here and see a full transcription of this interview below.

And this is our latest Insights article on BrainChip from 17 June 2022.

 

Disclosure: Pitt Street Research/Stocks Down Under directors own shares in BRN. Also, BRN is a research client at Pitt Street Research.

 

 

No time to do stock research, but you still want to invest?
 
Stocks Down Under Concierge gives you timely BUY and SELL alerts on ASX-listed stocks!

 

GET A 3-MONTH FREE TRIAL TO CONCIERGE TODAY

 

No credit card needed and the trial expires automatically.

 

 

 

 

Transcription

 

Marc: Hello, and welcome to Stocks Down Under. Today we’re joined by Sean Hehir, the CEO of BrainChip. Welcome, Sean.

Sean: Thank you, Marc.

Marc: Great to have you here. You’ve been with your company for about six months now, more or less?

Sean: A little bit less, about five and a half.

Marc: Five and a half. What were the sort of the key priorities when you came into the company because your background is in go-to-market commercialization? What were the key priorities for you coming into BrainChip?

Sean: Well, it’s not just coming into BrainChip. I think any time you come into a new situation, you wanna listen. So, my first 30 or 45 days I really just wanted to listen, listen to the employees, listen to customers, listen to everybody that I could to really understand as much as I could about the company. And then I tried to play it back and clarify. So, you know, am I hearing it right? Make sure I’m making the right assumptions before I make any kind of action whatsoever.

Marc: Right.

Sean: But what I said yesterday at our AGM, I think is very, very true. We spent nine years building this amazing technology. It’s transformational in its capability and powerful, but we were not doing enough in our go-to-market motion. So, I started to build the go-to-market motion, and I laid about five or six steps. One was clarifying the business model. That was very, very important. Chips or IP? We clarified that it’s IP. A really important step. We do have chips available, boards and things like that, but the revenue streams of the company are gonna be IP.

Secondly, we wanted to tune the product fit to the market. So, we did a lot of market research saying, “Are we hitting it with all the attributes?” Thankfully, it’s 100% with all the research we did. The flexibility, the scalability, all those things about our IP that makes it unique in the market are very attractive. So, we validated that. Then we cleaned up our messaging. We just launched a brand-new website two weeks ago, and you’ll see all the business benefits about making companies more competitive and agile and things like that.

Marc: Right.

Sean: Before, it’s little more technology. So, it’s all about business value. Lastly, we put a lot more resources into our sales organization. We’ve cleaned up all the messaging there. We’re doing sales enablement, competitive analysis, talk tracks, outbounds. We introduced a whole new CRM system. So, this is all about taking nine years of research and development, and now taking it to market in a very loud and aggressive way.

Marc: Right. And you said messaging. What’s the key message at this moment for BrainChip?

Sean: Well, at the simplest level, we say cloud free essential AI. And what that means is obviously the attributes of our technology is you can do a lot on the local chip or the local sensor, whether we’re deployed. Less dependencies on the cloud. And of course, that makes less latency, less security and much more efficient, right. So, cloud free. Our product is aimed at the Edge. In fact, I made a comment yesterday. I thought it was fascinating. I was preparing for this trip. Michael Dell, Sony’s been very successful in technology for a while, said this year he assumed or guessed that it was roughly 10% of all computation and data was…10% of that was on the Edge. By 2025, he projects at least 75%. That’s a big transformation.

Marc: That’s a major step, yeah.

Sean: Yeah. Now whether he’s right or wrong, let’s call him directionally correct, right. Big, big market. So, cloud free for that. And the essential AI, what we’re saying there is the neuromorphic nature of our technology allows us to do things very simply, quickly with much less data and less, less computation. So essential. So, you know, the ability to do things very quickly and efficiently is the essential part.

Marc: Right. Okay. And then looking at that commercial deal…so MegaChips obviously was a big announcement in its own right, but I can imagine it’s done wonders for the company in terms of opening the eyes of other players in the industry. So, what’s happened since then for you guys?

Sean: With MegaChips or in the industry generally?

Marc: Well, following…no, specifically for BrainChip, what happened after sort of the MegaChips announcement? What it did for your commercial interaction?

Sean: Yeah, you know, they weren’t the first. You know, we have another deal in place as well and we’ve got tremendous engagement, but having companies like them…these are world-class companies. Renesas and MegaChips. I mean, these are really… Excuse me, Renaissance and MegaChips. It’s very, very, you know, important for that. You probably saw the announcement from Mercedes. Did you?

Marc: Yep.

Sean: Yeah. So, companies like that, having MegaChips helps validate a lot of it because MegaChips services tier one companies in Japan and actually around the world, but predominantly the Japanese market. So, when they see…when customers see that, they, we feel, have a sense of comfort for that.

Marc: All right. And then there was not an announcement but it was news around a partnership with Arm. Can you talk us through that, what it entails, what it means for you guys going forward?

Sean: Sure, sure. So, you know, before we started, you and I were chatting a little bit about my background. I said, you know, one of the things I’ve done in my life is I’m an ecosystem or partner person, and at one point at Hewlett-Packard I ran all the ecosystem for the entire company, all the silicon companies, all the software companies, all the system integrators, all the consulting company around the world. So, I believe in the power of partnerships. And why is that? With enterprise grade technology, it never stands alone. It never does. It’s always wrapped with other elements. And so, what we do is critical to the value of any enterprise customer but there’s always other pieces of it. So, these relationships like Arm, and I’ll get through it a little more specifically…

I embarked on a strategy, what we call ecosystem partners. We have technology partners like Arm and SiFive that we’re gonna work with closely and develop prepackaged solutions and have…so our customers can have confidence that it works well with these leading processor companies because they’re gonna implement it with these processor companies. So, the more we do that work together, which we are doing right now with Arm and SiFive, packaging, and characterization, and testing, and co-selling, these are very powerful things that can happen to make our customers feel more comfortable.

In addition, we’ve done things, what we call enablement partnerships. We announced things with NVISO which is a Swiss company. These are companies like Edge Impulse, and all these companies create models and datasets for the completion of the value. So, we make IP. You’re gonna work with the processor on one end, and you’re gonna work with models on the other. This is a start, not the end. We’re gonna add more partnerships and depth to these relationships.

Marc: Right. So just on Arm, integrating, you know, different types of semiconductor IP into one system on a chip probably is pretty hard, right?

Sean: Yeah.

Marc: So, what’s the timeline for that partnership to start sort of evolving into something that can generate revenues for BrainChip?

Sean: Well, we think…first of all, the awareness has been big already. We get calls already. The minute that went out, you know. And officially, it just went out a couple of days ago. Right now, we put it up on our website officially…as soon as it went out, we started to get calls and questions about it. We’ve already done some characterization work so people say, “Okay. Can we do this?” And the answer is yes. So, it depends on the level of integration you’re talking about. But we’ve got working demos right now, and in fact, most of the systems you buy from us right now have an [inaudible 00:06:51] processor on it. So, if you buy one of our boards or you buy one of our systems, it’s already working like that today. It’s just further type of characterization.

Marc: Right, okay. And in terms of deals and partnerships…so you must be talking to a whole range of potential partners, customers in different spaces. You know, it could be people like Arm. It could be potentially chip manufacturers. Could be system integrators. Can you talk a little bit about sort of…I know this is hard, but about the pipeline? Who are the type of people that you talk to, type of companies that you’re talking to? What sort of…you know, the things you expect to come out of that pipeline [crosstalk 00:07:27]

Sean: Sure. Sure. You know, this is one where a visual would be helpful but I’ll try to make it simple. I provided a slide yesterday at our AGM, and one of the things I did when you asked about the priorities was clarify who are our target markets, right, or who are our target type of customers. And we’ve really simplified it to four basic prospecting type of customers. One is what we call sophisticated companies. We talked about what Mercedes announced. Again, we didn’t announce it but Mercedes did because we respect, you know, our relationship with them. We’re not supposed to talk about that.

Marc: That was pretty significant, by the way, that they would announce something like that.

Sean: Yes, yes. We’re very thankful for that.

Marc: Yeah.

Sean: But so those are sophisticated companies. Companies like that that are, you know, very advanced in their thinking, very advanced in their capabilities usually with lots of AI expertise inhouse and actually SOC expertise inhouse. So, we look for those kinda customers to take and work with the ideal targets for IP. Buy a license, and of course, then design a chip either inhouse or work with a chip designer. Then we’ve got what we call simple customers. Not very, very fancy names, but simple customers. But many customers, this is kinda the biggest part of the market, they wanna go AI and they typically have a very small evaluation team. You know, three or four people evaluating AI technologies. Then they’ll work with a chip designer or a chip maker.

So, we’re targeting sophisticated customers and simple customers. The chip designers and the chip makers, of course, complete that. Some of them will proactively take a license like MegaChips to work with their customers in a sense or a channel for us. So, we’re addressing in dialogue with those kinda companies, and chip makers which will obviously wanna hold the license and make products using our technology for their own product distribution. So those are the four types of customers.

Marc: Right.

Sean: And it’s also a push-pull model meaning we talk, say, to a simple customer. They’re interested and they’re like, “Yeah, we wanna do this but we don’t know who to work with.” We put them together with one of the chip designers that we’re talking to. So, we call that a push-pull model, those four organizations.

Marc: Yeah, all right. And I looked at your AGM deck from yesterday. There’s a lot of slides on commercialization. Just sort of at a top level, can you talk me through the strategy, the commercialization strategy? Just, you know, boil it down for a simple guy like me. How you go to market basically with a complicated product like what you have at BrainChip.

Sean: Yeah. You know, it’s sophisticated and elegant. It’s really not that complicated when you simplify things down, right, because it’s tremendous business value. So how we go to market? First of all, it’s just what I said, simplify it. And we’re very proud of the work we did on our messaging and our website. And then value jumps off the page. And it’s interesting, most people know BrainChip, because it’s true, of our energy efficiency because of its neuromorphic nature, but part of our engagement since I’ve been here, we have also realized that if you’re not power sensitive, we can go toe to toe with anybody else on raw performance, okay.

Marc: Yeah.

Sean: So, we have the ability…because our IP is so flexile, we can scale the number of nodes, the clock speeds and things like that. So, we can go very power efficient for those things that don’t have a power supply in, like, an IoT device but if somebody’s gonna have a power supply and just want lots of raw performance, we can go toe to toe with anybody. So, we kinda look at all those four groups of people I talked about a moment ago, chip designer, chip makers, sophisticated and simple. We first said, “What are they looking for? Power sensitivity or raw performance?” We then tailor our message to which one of those attributes we’re looking at.

So that’s probably the biggest way we do it. The biggest thing in the IP business is you’ve gotta line up with design cycles too. So, we’re constantly talking to prospects and paying attention to when they’re gonna go into a design cycle because then that’s when they’re gonna buy. So, you have to do a lot of conditioning and business development activity and when they’re ready to go into a product design, that’s when you put the proposals on the table.

Marc: Right. Okay. Last question and it’s probably the simplest one for you. Looking at deals with, you know, likes of MegaChips, just for our viewers, can you just briefly explain what revenue looks like and at what stage that comes in?

Sean: Sure, sure. At its simplest level, our revenue model is very much a licensing or an IP model. So, when you sell to a MegaChips or others, they pay, you know, a significant amount of money for that license for a period of time. And they either have a single design or multi-design depending on what they decide to buy. But, you know, it’s a significant purchase for those companies to buy. Then you go through a period…when they produce their products, you get royalties. But I like to describe how that royalty stream looks like. Obviously, there’s a delay. Somebody buys a license. Then they’ve gotta develop their products so maybe a year, two, maybe even three sometimes before the products come.

But when their products start hitting the market, we collect a royalty on each individual product. And so, our modeling, the way we do it is it’s a factor, not a percentage. So, if somebody paid, you know…let’s just say they paid $1 for a license. Obviously, it’s a lot more. We’d expect $3 or $4 or $5 in royalties, not 50 cents. So that’s a way to look at it if you’re trying to model our revenue. As licensees come on more and more, there’ll be a delay but then, quite frankly, the royalties with no cogs come through, and that’s the beauty of this model.

Marc: Yeah. That’s where the scalability comes in [crosstalk 00:12:25]

Sean: That’s right.

Marc: Yeah.

Sean: Right.

Marc: Excellent. Well, Sean, exciting times. Thank you very much for coming by.

Sean: Thank you, Marc.