Hi, everybody. Thanks for joining us. My name is Ruben Roy, I cover applied technology here at Stifel, and I'm really excited to have QuickLogic back with us again this year. I think it's a really interesting story, a company that's been around for a long time, which Brian Faith, CEO, standing to my right, is going to talk about, but I think which has also reinvented itself over the last several years, and I think it's getting even more exciting. We also have Elias Nader with us, the CFO of the company. So we're going to be drilling down into some of the model bits a little bit later.
But, Brian's got a few slides just to set the tone and, and get the, you know, kind of less familiar folks up to speed with QuickLogic, and then we'll get into Q&A. With that, Brian?
Thanks, Ruben. Morning, everybody, and thank you to Stifel for inviting us to the conference to present our story. So for those of you that were here for the previous presentation by Sammy of Efinix, you got a background of programmable logic. I'll echo some of the same sentiments. So programmable logic is going through sort of a renaissance right now. Although it's a 40-year-old industry, there's a lot of new use cases being deployed now for programmable logic. But fundamentally, people look at the technology as a way of extending their designs into the future through that on-chip programmability, enabling new use cases, in certain cases, offloading compute, enhancing power consumption or battery life. It's a very pervasive technology across many different segments, and it's predicted to be a very high-growth segment within semiconductors.
I saw a recent report that had FPGA consumption at about $12 billion in 2024, more than doubling by 2029, and that's across, again, a very variety of different use cases and end systems. So it's exciting to be in that segment today. It's exciting to be a company that's been in that market for a long time because, as Sammy of Efinix said, it's very challenging technology. One of my favorite trivia questions to ask people is to see if they know how many companies have tried to do FPGA technology during that whole 40 years of the industry. And most people are surprised to learn that 50 companies, 50, five zero, have actually tried to do that technology and essentially failed.
It's a challenging technology, but if you can get it right, the market is really, your oyster. One of the neat things about programmable logic now is that it's not just a discrete device, it is also an IP. As we know, there's a lot of companies now that are starting to vertically integrate, and as opposed to just buying standard products, they are doing their own custom silicon. It's very well documented that large tech companies are actually doing that now so that they can either control the supply chain access, or they can put some secret sauce IP in devices that nobody else can do. But they still like the notion that you could have programmability in that same chip, and if they're not going to be one of those 50 companies that tried and failed, they'd rather just license that technology.
And so that sort of gave birth to this whole notion of FPGA as an IP, and that's called Embedded FPGA or eFPGA IP. And we are a company that resides in both of those product categories. So as you can see from here, again, we sort of centralize on our Australis tool there in the middle of that cornerstone. That is what gives us this ability to generate our embedded FPGA IP in a very scalable way. We have services that we can customize the IP for customers for either unique use cases or different domains. We do have a device storefront on the right-hand side, which essentially means that somebody is relying on us to run the supply chain and ship them a known good tested device. And then on the far right, we have an AI/ML software subsidiary company.
It's called SensiML. That was a spinoff from Intel many years ago, and that is a software tool chain that allows people to take sensor data and turn that into edge inferencing models to run on microcontroller class devices. So in total, you can see the spectrum of technologies that we bring to bear. We feel our addressable market is about $1 billion, and that spans many different end segments. During the last four years, and actually a little bit before that, we started reinventing how we develop our programmable logic technology. We started to incorporate a lot of open-source components to that because that gave a lot of scale, and in some cases, with customers, it gave a lot of inspectability into what we were doing.
You can imagine certain markets really care about knowing exactly how we are designing our products and how the software is using them. That's been a journey getting us to last year, which is really when we started to see a huge uplift in the revenue and the profit from this foundational work over the last several years. As far as the vertical technology stack goes, as I said, it starts with having deep domain expertise in programmable logic. We were founded in the late eighties, and so we've really garnered a lot of that know-how, trade secrets, patent protections across those decades of being in the business. Now we're starting to see how those are coming out in different vectors of products, like the devices, like the IP, and then very soon in the future, chiplets.
Chiplets are a different way of doing semiconductor design, where you're stacking multiple chips in the same package as opposed to just doing something monolithically. There's a lot of reasons behind that for optimizations for certain workloads or cost optimizations that you'll see from us in the coming future. Business model should be pretty straightforward. For devices, when we sell a device, we get revenue and gross margin. If we're doing development work, we're getting funded development revenue from that as we're doing the development. For IP, it's very, I'll say, Arm-like in IP license and back-end royalties when customers ship products that use that IP. So pretty traditional in each of the product categories.
The nice thing about having all of these under one roof is that we get the nice benefits of the revenue, the gross margin, both in the upfront and in the long tail as we continue to ship products or our customers ship ASICs that have our IP in them. The biggest win to date for us as a company is a contract with the U.S. government. We are the prime contractor for a very specialized FPGA for the Department of Defense. This is a strategic radiation-hardened FPGA. We are being funded by the government to do something that doesn't exist today, and it's very purpose-built. And I'll get into how and why we're- we were able to win this.
But this is the type of thing you see when markets get large enough and technology gets pervasive enough, that then you can start to see very domain-specific applications popping up, and this is an example of that. Leveraging the fact that we've been serving the defense industrial base for almost our entire existence of 30 years. But it goes beyond defense. So there's lots of use cases in industrial. That's probably our second-largest market. AI/ML, obviously, a lot of people are talking about that. We do have some wins in AI/ML. And then you can think about security and consumer devices as well, in the future. So some recent announcements from us, broadening the scope of our technology. We have some 12-nanometer embedded FPGA contracts that we've announced.
We've also announced some ecosystem developments with companies that specialize, again, in sort of domain-specific areas, like radiation-tolerant. You could imagine that that would be very applicable for space, and satellite applications. We're just-- we're cranking on these new contracts, and the way that we can do that is... I'm gonna go to this slide first, is through this tool that we have called Australis. This is how we are able to automate the process of doing our IP design and porting it to different foundries and process technologies. When we started as our company, you know, 30+ years ago, everything was full custom, so we would have to allocate our entire engineering team for probably about a year and a half to move our technology from one node to the next.
By bringing in the automation capabilities and automating that whole process, we can now port from one process to another in about 3-4 months with about 2 people. So you can see there's a huge operating leverage difference between the old way and this automated way, and this automated way is how we're able to broaden the, the use and the ability to tackle these designs. In fact, in Q1, we already said that we were doing 6 different technology ports in the same quarter. That would be completely unheard of 20 years ago when we started as a company. So that's what's allowed us to execute on these contracts, expanding our customer base. You see the nice logos here. These are either customers or partners of ours.
Again, at the bottom, for a long time now, we've been working with all 5 of the top 5 and 8 of the top 10 U.S. defense prime contractors, and that's either devices or IP, but it really shows, I think, the trust of the customer base that really cares about quality and longevity of the technology that we offer. This actually we announced about a week ago. I'm very proud of our team for this, but we got BAE Systems Supplier of the Year. This is hard to do, but we got this for the FAST Labs Technology Innovation Partner of the Year category.
And I think, again, this just speaks to not only do we have good technology that meets all the needs of customers like BAE Systems, but we also have that longevity and the trust built in the industry, that we do what we say we're gonna do, and when they buy something from us, it does what we say it's gonna do. Lots of contracts in the last couple of years. So like I said earlier on, once we got Australis and that automation in place, we really saw a lot of uplift in our ability to execute on those new contracts over the last couple of years. Funnel pipeline's almost $180 million, and $32 million of which are contracts just in the last couple of years.
So really starting to see the operating leverage come into play now, through the automation. Good question: "How do we define pipeline?" Pipeline, in this case, are qualified opportunities where our technical teams and business teams are engaged, and we see that there's a technical fit, there's a business fit. It's not one business yet. Once it becomes one business, it flows into the contracts that are in the top, the $32 million. I'll also say that this is a 2-year horizon of revenue. So as an example, if we have, like, a 4-year contract, which that defense contract was a 4-year contract when we first signed it, the last 2 years were not in the sales funnel at that time. As we march through time and stuff gets into that 2-year horizon, it falls into the funnel. Good question.
Just briefly on the AI/ML tool, so this, again, is a company we acquired about 4 years ago. They were spun out from Intel. They focus on making it incredibly easy and efficient to take sensor data and create inferencing models that you can run on very resource-constrained devices like microcontrollers. As you probably read, it's actually very difficult to take data and create inferencing models, and you have a garbage in, garbage out situation where you have models that identify things that shouldn't be identified. This toolkit is very efficient in making sure that when you have data coming in, it's clean, it's labeled, and you can create very low-power models to run on edge devices. And you can imagine that being consumer devices. You can imagine it being predictive maintenance for industrial.
Today, SensiML targets x86 by virtue of being part of Intel originally, and many different Arm architectures all across there at the bottom, and looking at how to expand that in the future to things like RISC-V. And they also announced a couple of things in the last six months or so. One is a private label, so one of the large microcontroller companies has private-labeled SensiML as their toolkit. Secondly is there's an open-source initiative now that we launched about a month ago on our last earnings call, so that some of the capabilities of SensiML are open source, which is gonna encourage even more developers to come in and start to improve the toolkit and add new features and capabilities that don't exist today. Now getting to the brief financials. So again, last couple years, a lot of uplift in the revenue.
Almost all of the revenue growth has come from the embedded FPGA IP products and services associated with that, so strong growth over the last few years. Relatively flat OpEx over the same time period, so we reached profitability last year. We're forecasting profitability again this year, and another 30% revenue growth for fiscal 2024. And then here's the punchline. So we've just, again, been in the market for a long time, but I think we've found the magic way of automating what we do in engineering to go after more customers at the same time, lots of operating leverage, and that's right at the right time.
As I said in the very first slide, FPGA technology, programmable logic's going through this renaissance of being designed into so many different systems, including custom ASICs, and we're uniquely positioned to take advantage of that market opportunity with that scale that we have. And we see that in contracts, and we see that in our, our funnel growth. So with that, I'll turn it back over to you, Brian.
Thank you, Brian.
Thanks.
I wanted to start with the pipeline and Australis, because I think, you know, obviously, that's something that you've been talking quite a bit about. We were here last year talking about it, and we're seeing, you know, continued momentum, let's say. How much of the pipeline is built on Australis? And maybe you can walk through, like, kind of how that's built up as that tool was made available to your customer set.
Yeah. So I think, firstly, almost all the growth is driven from our embedded FPGA products, which all of that is based on Australis.
Yep.
We're not doing anything that's non-Australis at this point. If we double-click a little bit on that, I'd say close to 75% of that growth is in the aerospace and defense sector.
Mm-hmm.
That could either be just an IP, where we're licensing an IP, like in the case of a 12-nanometer technology. It could also be that we are doing IP development work for devices like the strategic Rad-Hard contract. That's how I break up the growth in the funnel.
Okay.
Yeah.
When you talk about these ports, I think you mentioned six different technology ports, you know, happening, and Australis is driving that. Can you maybe walk through? Is that mostly also aerospace and defense, or are you seeing some broadening of, you know, sort of exposure by end market?
So for those six, I would say about half of those are related to aerospace and defense in number. Some of those were industrial, and one of them in particular is for an edge AI use case, where the customer is designing their own ASIC focused on really low-power positioning, and they wanted to use the FPGA capability in there to do offload from a processor to do some algorithmic acceleration-
Mm.
With the intent that you could lower the energy, since batteries store energy, lower the energy consumption for this inner loop function. That's one of the great things about programmable logic, is that you can parallelize algorithms, and you can do some algorithmic exploration because the silicon is programmable. And so, we actually had a published research a few years ago for a very similar use case, FPGA tightly coupled with a RISC-V processor in 22-nm, and I think just being able to articulate the value proposition to them convinced them that this made sense also at 12-nm, that the value proposition would scale. So again, back to your question, though, of those six-
Right
... about half were defense.
So we're seeing some broadening, I guess, yeah, is kind of where I was going with that, so that's interesting. Another area I wanted to touch on with Australis specifically is, and we've talked about this, Brian, I don't know, for years now. It's really interesting to me, the automation part of it, right? And time to market's important and all the rest of it, and it just seems like everything's accelerating in the market itself now.
Yeah
... by a lot of different end markets. NVIDIA talks about this, accelerating the cadence of how quickly they're coming to market, and they're using some AI tools. They're eating their own cooking and developing these, you know, kind of NeMo specific, you know, toolkits and models and algorithms to help their engineers come up with designs faster. So, you know, as that is evolving, in obviously other parts of the semiconductor design ecosystem, is that helping you with your customers? Are they becoming more aware of, you know, the benefits of Australis and automating, you know, some of these designs for different markets?
I think they are. Firstly, we're starting to just see the snowball effect, where people-
Yeah
... can see, as we announce things that are based on Australis, you just get a lot more interest coming in. And I think the fact that we are incorporating open-source components is piquing people's interest, because they feel like if it's open-source, it's sort of safe in the sense that there is a community of people developing around that.
Mm-hmm.
One of the things I was mentioning earlier, too, is just the inspectability. So a lot of people don't like this black box concept, where you have to completely trust what you're getting from a supplier. If you have this open-source components inside, you can actually dig in and see how people are building things. I think that inspectability is sort of lowering the risk perception in people's mind if it's a new technology for them, like embedded FPGA.
Right. Are there any questions from the audience?
The defense side of your business, is there something specific that you guys are doing that gets you that business? Or I guess a better way of putting it is, who are you competing against?
Can we repeat the question?
The question was around what is, you know, unique to QuickLogic that's enabling them to win, you know, these big contracts with, you know, aerospace and defense and government contracts, and what's the competitive environment like in that end market specifically?
Yeah. So I think if we look at that particular market segment, I was just talking about risk, right? There's a huge resistance to taking on risk for obvious reasons. So the fact that we've actually been selling into that market or across different generations of products in the defense industrial base, I think helps take down that risk, and that's one of the reasons why we won this contract, because we're sort of a known supplier in that space. The second thing from a technology perspective, as I said on that main slide that I presented on the strategic Rad-Hard, that's something that doesn't exist today. But in order to design that, you really have to know FPGA technology.
Going back to my trivia question earlier, there's just a handful of companies that have been in the market long enough that know and have learned through the mistakes of how you can build something that meets the needs of that mission in a way that's going to be manufacturable and reliable and meet the cost targets, as well as have the software in place for users to actually target that device. I think those were reasons why we won that. The other thing I'll say is, I think within the defense market, FPGA technology is probably the highest revenue or highest spend category in all the defense companies, and so there's a lot of desire for investment to go into that.
Because I said earlier, as the market gets big enough, then you start to see fragmentation, where people are optimizing for different domains or use cases, and I think this is an example of programmable logic has been used a long time. It's a high-spend category. I'm not getting exactly what I would like from the main FPGA companies. How about I go and invest in a company to build what I want across those use cases? And the last thing I'll say is that a market like that is large. It's certainly large for a company like QuickLogic. It's probably not large enough that the really large FPGA companies would really go after that kind of a space, 'cause it's probably too small to matter for them.
It's obviously big enough to matter for us, and I think that mix is really why we were really positioned well for this and why we won that contract.
I think the Rad-Hard part of that, I would agree with you, you know, completely, having followed Xilinx and Altera pretty closely. I don't think they really focused on that at all. Then, you had other companies that are no longer with us. They're, you know, kind of consolidated into larger companies that were looking at secure FPGA as one-t ime programmable. Well, maybe, like you said, Brian, there's a lot of different-
Yeah
... use cases and feature sets that just got too broad for the big guys to really worry about.
And then, I'll throw one other thing on there. There's a very large push to have onshore manufacturing-
Right
... for assurance of the supply chain, and so my belief is that there's no commercial FPGA manufactured in the U.S. today. Like, I haven't double-checked that, but I believe that to be true. In this case, this will be, the one that we're doing. And so if you think about, like, technical capabilities, company that's backing that whole thing up and then where it's manufactured, I think that really placed in, into how we won that.
Yeah, that was the same question I was going to ask, which was: Who did you compete against, and what's your win rate?
So on that particular one-
The question, just so everybody can hear.
Yeah, so the question was, who did you compete with, and what is your typical win rate when you go after these big opps in this market?
So for that particular contract, there was a general RFP that went out to the community. I actually don't know exactly who proposed solutions, because we're not privy to those things once they go into the government. But I think, again, the reason why we won that is for the reasons I already said. As far as the hit rate moving forward with other opportunities, I think anything in the aerospace and defense side, we have a very good hit rate, probably more than 50%. On the other markets, like industrial and the upcoming AI ones, it's not 50% yet, and it's probably still too early, since we've only been at this new flow for about two years, to really have the statistics be meaningful.
But it's something that we are tracking, and we've said publicly we'll probably start disclosing more about hit rates maybe next year.
Elias?
Yes.
You're up here with us. Let's... We've got to get it, got to get you involved. As the business, you know, starts to, you know, kind of unfold in a lot of these areas, and you're gaining traction and momentum, like we just talked about, Brian, FPGA companies historically have been really profitable companies. They've had great margin structures, both gross margins and operating margins, and as Brian said, the outlook looks pretty good for, you know, longer term CAGR over the next 10 years. How are you thinking about margin structure for the business as you ramp, you know, into some of these, you know, areas of revenue?
We've had some fantastic margins the last few quarters. 78% was our high, but I'm not going to tell you we'll hit 78% again. That was, that was a one-off. Certainly some, some product mix and, capitalizing on certain, R&D tools because they have multi-use purpose. But our goal internally is high 60s, so I, I'm very comfortable with the high 60s until we get to that storefront that Brian has talked about in the, in the past, that will then shoot up to, to probably a higher number.
Okay. Thank you. Any other questions? I don't know. There, we have one over there.
I understand, yeah, but when you talk about bringing something to the industry, what's the end that you're addressing, but you're smaller a few years ago. Do you expect to be much larger?
It was materially smaller a few years ago, prior to us really, I think, figuring out how to do the IP automation, because at that point in time, from an IP perspective, if it takes us a year to do a port, it really narrows down, and we have to really be wise about how we choose which ones we go after with our engineering team, just because of the opportunity cost. And then, from a device perspective, from a low-power consumer point of view, we had some wins there. Like, we still ship into Kyocera in Japan quite a bit for smartphones, and we had our existing products, but I would say a substantially smaller TAM.
Once we got that automation in place, now that opens up so many more possibilities just simply because of the scale, because the opportunity cost is so much smaller now on that. And one thing I will say, I'm glad you asked that question, a lot of the revenue growth right now is in the IP and the development side. If you think forward a couple of years from now, especially when that strategic Rad-Hard FPGA is done, and the ability to sell that device into the defense industrial base, that unlocks a huge market that's not even reflected in the funnel at all today. And if you think about how much, again, to my earlier point, like, how much revenue or how much spend the defense industrial base has allocated for FPGAs, that's a multi-hundred million dollar a year market.
And so I think the TAM actually is going to jump up quite a bit, step function increase, in that sort of 2.5-year horizon when that device is actually able to be sold. And that's one of the most exciting things to me, is, yeah, 30% revenue growth is great, IP is great. It's getting to that storefront ability, which is what Elias was talking about also, to me, that's the really, really exciting part of it. 'Cause as you know, once those devices are designed in, you're talking about years, if not decades. We have precedent for stuff that's designed in decades ago, still shipping, right? It's, there's some critical systems where the cost of change is simply too much. They'd rather just stick with that architecture, and that's what we want to get to in the next couple of years.
It's an interesting question. We had Sammy from Efinix up here just before, Brian and company, and Sammy had talked about sort of, you know, looking at the market and, you know, potentially getting into some areas that FPGAs, you know, hadn't competed so much with from a socket-to-socket basis, like microcontrollers, for instance. And we've seen, you know, companies like Lattice do pretty well, and now they're moving up into the mid-market. We've seen Intel say, "Hey, you know, maybe this isn't core to what we want to do, and we're going to spin Altera back out." What do you think is happening, either by end market or, you know, just technology-wise, that's maybe bringing back, you know, sort of the concept of FPGAs, which have been around since the mid-1980s, right? So what's going on? Any higher-level thoughts there, Brian?
Well, I think what happened during COVID really paused-
Mm-hmm.
What people were thinking about their supply chain and made them reflect, "How do I take more control of this so that I can't be prevented from shipping a, you know, a million-dollar system because of a ten-dollar component?" And the fact that semiconductor design now is starting to become more affordable for people to do themself, I think that's really opened up possibilities where they're looking at, "If I do my own chip design, take control of my supply chain, how do I, how do I use programmability for that to amortize that investment over many years?" People like that notion of, of being able to program things in the future. And so we're starting to see that across sectors, including ones that never really did their own custom design before. Like, we've been talking to industrial customers-
Mm.
very long tail, but, you know, those were the ones that were hurt the most, I think, during COVID, looking at how they use programmable logic for that, that fewer-
It makes sense.
-flexibility.
Right. Another, you know, sort of question that I've been asking most of the companies that I've been speaking with here at the conference is just around near-term-
Mm
... dynamics. Obviously, the macro has been tough in some of these end markets that you guys are looking at. You have a different, you know, sort of, you know, kind of path forward coming with, you know, new products and from a little bit of a smaller footprint than, you know, some of these other folks. But, you know, in terms of the market versus design activity versus, you know, customer engagement, how is that playing out, you know, over the last, let's call it, you know, six months, and how do you see that playing out over the next six months?
I would say we've actually seen an acceleration in people from a design perspective, I think, coming through the lens of the IP side of our business. Because I think ASIC design is such a long-
Right
... long development cycle, people aren't slowing down on that because of near-term economics. And so we're actually seeing, I'd say, good growth in our sales funnel accordingly. You can see that in our reported sales funnel. As far as the sort of mature products that we have, the more discrete FPGAs, I think we went through sort of a trough in the second half of last year as people were digesting inventory, but we haven't seen that as far as this year. I think partly because I think companies sort of made a mess for themselves by forcing long-term sales agreements on customers and forcing people to take inventory when they didn't need it. We didn't do that, so I don't think we had such a drop as a result, because we weren't forcing customers to take things they didn't need.
And so I think we haven't seen that much of an issue as what you may have seen with other companies.
Great. We've got a minute left. The last thing I wanted to touch on, Brian, is just, you know, strategically looking ahead, you, you've done, you know, this SensiML, SensiML, I always get that wrong, but SensiML-
Sensible.
Sensible, okay. Sensible acquisition. You just announced this collaboration with YorChip. What's next strategically? What do you need to do, you know, sort of to scale this business further?
So what we're going to be doing there is, the whole vision here is that these next couple of years fuels the storefront, because five years from now, that storefront is going to be, I think, the dominant part of our revenue, and the dominant part of the gross margin dollars at that time. And so the things that we're doing strategically are all about how do we lay the groundwork with different IP or partnerships to make sure that we can turn those on at that time. And you'll see probably partnerships related to foundries and process technologies that are going to be very applicable for us to grow the chiplet side of the business, because I think chiplets actually gives us scale.
Yeah
... and ability to sell into many different application sets from one or two mask sets. And the industry is moving in the direction of chiplets today. You don't see a lot of it commercialized outside of companies just vertically integrating within themselves, but I think in the next few years, you're going to start to see sort of this open ecosystem around chiplets, where Company A can buy Company B's chiplet and integrate that. And, there's some groundwork that we need to lay for that. Part of that was YorChip, but there will be things related to process and foundries that we need to work with as well, that you'll see, that I think will be early markers for that.
I think that's really exciting. Unfortunately, we have to stop there. Thank you, Brian and Elias, for joining us today.
Thank you.
Thank you all.
Appreciate it.