Pleasure to host this presentation with QuickLogic. QuickLogic is a fabless semiconductor company specializing in embedded FPGA, hard IP, discrete FPGAs, and endpoint AI solutions. The company's unique approach combines cutting-edge technology with open-source tools to deliver highly customizable, low-power solutions for industrial, aerospace, consumer, and computing markets. Here presenting for the company today are President and CEO Brian Faith and CFO Elias Nader i will hand it over to Brian and Elias for the presentation and then come back to moderate Q&A. Brian?
Thanks, Quinn. Good afternoon, everybody t hanks, for staying at the conference. Good attendance in the room tonight. Some familiar faces, some new. So, as Quinn said, I'm Brian Faith, CEO of QuickLogic. I've been at the company for about 28 years, and more excited now than ever about what we have going on in the prospects for 2025.
So, just to level set, everybody, QuickLogic is a fabless semiconductor company a s Quinn mentioned, we have embedded FPGA IP. Let's talk about what that is. Programmable logic is a subsegment of the overall semiconductor industry, somewhere between $10 and $15 billion a year within the programmable logic segment. There are two types of products that companies do in this area o ne is called FPGA, field programmable Gate Array t hese are devices that are programmable by end customers.
Embedded FPGA IP is the same capability of an FPGA, but it's an IP that a customer would put into their fully custom ASIC. We have two different business models. Basically, the goal is to be programmable logic everywhere, independent of how customers choose to integrate that within their system. We were founded in 1988 as purely an FPGA company, only devices, selling them to aerospace and defense customers, instrumentation and test, industrial.
It wasn't until about five years ago that we started getting pull from customers that started to do their own ASIC development, saying, "Hey, I like the value proposition of this programmability of the FPGA, but if I'm already going to go commit $25, $50, $100 million doing a custom ASIC, I'd like to integrate that into my device.
That way, I don't have to do a big expensive ASIC and then still have to go put an FPGA beside it for my production." So, having that pull from the end markets really has sort of validated our approach that I'll go over in a second of how we design our IP in a very scalable way, and that launched into these different business models that we have today. So, to be clear, the way we monetize our technology, we can do design services for customers that want some level of customization of our products or IP a nd I'll go over some examples of that.
Ultimately, after we do that level of customization, we want it to land on one of two spots: either embedded FPGA IP as a license or a device that we do the supply chain management for and sell a finished good to the customer base, and we have examples of both that I'll walk you through in a second.
I will note, on the right-hand side is an AI/ML software. It is somewhat detached from the trapezoid, and that is a business that we are currently undergoing a strategic process on w e announced that last week, so we'll be giving it less airtime today and focusing on the real growth drivers for the business for this year and beyond.
If we look at the served available markets for our technology, we view it as around $1 billion, and that's comprised of device sales and also IP licensing and revenue, or excuse me, royalty. If you go back from the last couple of years, we really started to go down this path of IP licensing in earnest in around the 2020 timeframe.
In that timeframe, excuse me, Google was advocating a different approach to doing FPGA devices and software. You could imagine that Google wanted everything to be open source, and so they were actually putting a lot of investment, both in terms of FTEs and investment dollars, to sort of seed the FPGA market with open source alternatives to what was always a proprietary walled garden of software.
And we really resisted that for quite some time, but we got to appreciate the level of maturity of some of this open source technology. And so, instead of rejecting it, we decided to embrace that, and I'll go over the fruits of that labor in a bit. But that was really the catalyst that kicked off not just open source leverage of some of our toolkit, but this whole IP initiative that we've now launched to the market and is what has taken us to profitability in these last few years.
I'll go over a little bit on the vertical technology stack FPGA technology in general is very unique in the sense that a lot of people talk about the silicon, how fast, how much power, how many gates, et cetera, but actually, it's the combination of the silicon and the software that make it a very difficult technology to commercialize.
If you go back in time, and I'm going back into the 1980s now, earlier than some people in this room are born, no doubt, those companies that founded FPGA technology are largely still in business today. I'm talking about the AMD, the Xilinx division of AMD, the Altera division of Intel, Lattice, the Microsemi division of Microchip, QuickLogic. My favorite trivia question to ask, How many companies do you think tried to develop FPGA technology since those 1980s and now and are no longer selling FPGA technology? We have a 30 any other takers?
Keep going. This gentleman in the front said 40. It's actually over 50 i t's close to 60 companies that have tried to do FPGA technology and are not doing it today, and a lot of the reason why is because it's that unique combination of the silicon design having to co-work with the software to achieve that level of commercialization that you would expect.
And so, there are very few companies that have this sort of technology stack nailed. I'm proud to say that we are one of those companies, and that's what allowed us to not just have IP, but also devices and be on contract with the U.S. government for a very unique version of an FPGA that I'll go over in a second. In the future, we want to do chiplets so that any possible way that somebody could imagine using FPGA technology is something that we can cover in our go-to-market strategy o kay.
Investor conference, let's talk about how we make money as QuickLogic. Funded development, that's design services. This is an example of the government contract I'll go over in a second. We get paid by our customers to do development work. We do that so that we can get an IP license or we can do a device development w e don't just do services for the sake of services. If it's an IP license, it's very much like the ARM business model.
They've established and conditioned the market that value-added IPs, such as FPGAs and processors, should get an IP license when you deliver the IP, and then royalties when the customer ships products based on that, and we're no different.
And then, on the right-hand side, the device storefront, like a semiconductor company, like many of you know, if we run the supply chain and we do a device and we sell it to a customer, we get revenue and gross margin when we ship that product out the door to the customer. So, we like this business model because it actually balances very well revenue in the upfront of an engagement, in the case of IP, or revenue in the back end of an engagement for in perpetuity of products being shipped, either with device gross margins or royalties.
This balance allows us to, again, balance the sort of lifetime of the revenue and the gross margin contributions thereof. I mentioned services. Our largest service contract, our largest win ever as a company, and this is dating back to the late 1980s, is the DoD contract I'm pointing to here. We're under contract. Our customer is the Department of Defense t his is the first time that we've been a prime on a different project. This is for a strategic radiation-hardened FPGA technology.
To me, this is a very exciting project, not just because the contract ceiling of this is $72 million, and that's a large number for QuickLogic, but at the end of this contract, we will have a device, and our intention is to sell that device through our storefront to the defense industrial base.
So, all of the requirements for this have come from our sponsor talking with the end stakeholders for this within the defense industrial base that drove our statement of work, that continues to drive everything that we do on this on a weekly basis. We were awarded the fourth tranche of this in the end of last year, December 2024, and we're continuing to execute on the development of this with some very key partners, some of which have presented at this conference these last couple of days.
And we're, again, we're just excited that we were awarded this as the prime, the capabilities that this will actually deliver to the U.S. defense industrial base at the end of the development period, which should be about two years from now, and then the revenue potential for this for us as a company.
In my estimation, this type of capability is probably on the order of $200 million a year of potential revenue, and that could be for exactly what the press release says. Strategic systems could be things that are flying around in space, like satellites. It could be for strategic defense systems, as again, what the press release articulates s o, we're heads down focused on this while we start growing and diversifying on the IP portion of our business, which I'll go over in a second.
So, a lot of what I've talked about so far, aerospace and defense, that has actually been a mainstay market for QuickLogic since our inception in the late 1980s i t's actually a very typical market for FPGA technology in the market today, regardless of whether it's QuickLogic or our competitors or other folks. We have other presence, so industrial IoT instrumentation and test. We do have some consumer business. We do have one win in AI now, We're hoping to expand on that approach with some of the stuff we're doing this year.
And the goal for this year is really diversifying customers and diversifying markets, but we are quite happy with the fact that we have such a large footprint in aerospace and defense because that market tends to last for a very long time. I think the gentleman presenting in front of me was talking about automotive, similar characteristics. Long time to get in, but once you're in, you're in for quite a while.
I'll go back to this block on the lower left. As I mentioned earlier, Google was sort of a visionary in trying to bring more open source components into the FPGA tool chain, and we are the first company of all the FPGA companies to actually embrace open source components in our tool chain and in our development flow.
And what that means is that we get a much broader community of developers contributing to the code base so that we can come out with some amazing products without investing all of the OpEx ourself as a company. Furthermore, we can actually contribute back some elements to the open source domain, which again sort of is a snowball effect in a good way. I will go to the next slide now.
Let's talk about some of the progress last year specifically. Last year, big theme, diversification for us, both in terms of process technology porting for the FPGA and customer base, so we announced design wins for the IP license on 22-nanometer GlobalFoundries, 12-nanometer GlobalFoundries, 12-nanometer TSMC, and we kicked off our most advanced process node yet, which is Intel 18A. We kicked that off in earnest around the summertime of last year. This is by far the earliest we've ever engaged a foundry in a process.
This is in conjunction with a defense contractor, and we are planning for that defense contractor to be under contract very soon so that we can deliver this IP and get our first IP license on a very advanced technology node.
As part of that development on Intel 18A, we were included in some of their IP ecosystem alliances. This gives us broader visibility with other folks that are planning to use Intel 18A for their ASIC development so that we can possibly license our IP to them as well. So, good broadening last year. This year, we're going to take it up another notch in terms of diversification.
This presentation's on our website, so you can go look at these logos yourself. You'll see a good collection of foundries around the world, customers, and especially at the bottom here, reemphasizing again this defense contract w e already do business with all of the top prime defense contractors on our existing FPGA technology, and we're planning to expand that again once these other products complete their development cycle.
So, one award we're really proud of last year, BAE Systems. So, this is actually going back a couple of slides. I mentioned a 12LP IP license GlobalFoundries-based to a defense company. We got this award from BAE last year as a result of that. We're quite proud of this. I'm happy to be working with BAE and enabling them.
With some of the technology that they want to be able to get into their customer base as well s o, this is a good partnership between a company like QuickLogic and an established defense contractor, and I think it really speaks to the capabilities of us as an organization and from a technology perspective to win this award.
Okay, so if we think about sales funnel now, we've been growing the sales funnel. We started reporting numbers of the pipeline, definition of pipeline, qualified opportunities with customers o ur teams are engaged technically and on the business side.
These are not booked business yet. Not all of it will be booked, but we start to see some good momentum in converting these from just pipeline opportunities into booked business, and you can see the upper right there, $37 million plus in revenue contracts in June of 2021, which is really when we started to say our development model for IP was getting to that maturity stage, and all of these contracts are based on Australis.
That gets to this next slide, so how are we different than anybody else? As I mentioned earlier, it's tough to do programmable logic. It's tough to do it right, where you would actually trust getting into an airplane whose flight control and avionics are powered by that programmable logic.
I think we figured out how to do this over the last three decades, but automating that process so that we could scale it as an IP business was challenging. The first couple of years, we were doing complete full custom development, and that really throttled our capacity from a company because we had to allocate our entire engineering team for more than a year to do one process port.
Now, through this automation that we brought in with Australis, we can actually bring up a new process in three to four months for the first time, and then any customer IP after that, derivative licenses, much, much faster, talking about weeks and in some cases as little as days. So, that whole automation process is actually what has given us the scale that we have today.
Some people have asked, "How did you do that?" I'll be open about this. DARPA has actually invested in a lot of different semiconductor design techniques and software, one of which was done with the University of Utah, and their whole goal was to go from idea to implementation. Think about like a software update for an F-35, 24 hours, no human in the loop.
Big goal, typical of what you would expect from DARPA. So, they invested a lot in a program at the University of Utah, and they open-sourced that so that people could actually incorporate that into their workflows. So, we've done that. So, you can imagine we've blended now this three decades of experience with programmable logic design with this software automation capability, and that's what Australis is, and that's what's given us the incredible scale that we have today to execute on these concurrent programs.
And in fact, we have s o, in the last two years, I think we brought up six different process technologies. Remember, the reference was our whole team, one year. We've brought up, I think, six in two years. Two last year were two 12-nanometer nodes, and we kicked off the 18A that'll be finished in this quarter.
And again, it's not just the very first time we do the port, but it's all the ones after that where we get the real leverage of our model with our engagements and our customers. And you can see in the bottom right there, we're covering the space in terms of foundries. So, whether somebody needs a very ruggedized node from SkyWater or Honeywell, or they want to go super advanced with Intel 18A or anything in between, we've got that covered.
And by the way, if we don't have it today, it's three to four months to getting it in place with our flow s o, it gives us a lot of flexibility to be at the node that the customers need us to be in the right time. I'm not going to cover this slide in detail other than to say this is the company that is the software AI workflow company.
We announced we're going through a strategic process on this last week. So, they do AI software very well. You can label sensor data, create little models that run on microcontroller class devices. Really good company, really good product, strategically not aligned with where we're headed on the FPGA side, and so that is the reason for running that process.
If you start tying all this together from a revenue perspective, you can see here going back to 2020, really good growth last year 2024, we haven't reported on 2024 yet, but per our guidance on our call, it would be ending up around flat for 2024.
Growth up until then was averaging about 30% per year for the prior three years, and you could see the growth is coming from the eFPGA product, so the IP, the government rad-hard work, and we foresee that this is going to continue to be growing into the next couple of years until that rad-hard device is done, and then we can start selling that into the defense industrial base y ou'll start to see device revenues most likely surpass that, in fact, in terms of total revenue.
But I think good growth on the focus points, which is really around this eFPGA initiative. Now, the revenue growth was actually with relatively flat OpEx. So, with a good gross profit dollar generation from that revenue growth, relatively flat OpEx got us the profitability in 2023 a gain, we haven't had our 2024 final call yet, so we'll update on that on the call in February.
But you can see the leverage of the model through the automation, and this really comes from the value proposition of the Australis tool. So, that kind of gets us to the summary slide on 2025. So, we have a robust sales funnel. We go through a culling every quarter i f something's sitting in there too long and aging, we get it out. We've had good growth in the funnel otherwise.
I will spend some time on the second bullet so , in Q4 of last year, our largest competitor on the eFPGA business was acquired. The assets and the team were acquired by Analog Devices. And so that's presenting a very significant opportunity for QuickLogic because some of those customers now, actually all of those customers, if you're not ADI, no longer have access to that technology for new development because that is now a captive technology within Analog Devices.
So, I think the fact that Analog Devices acquired them speaks to the value of eFPGA t here were already users intending for it to be used in other devices, and there's a lot of customers now that we're getting inbound interest to QuickLogic as an alternative. They're committed to eFPGA, but now they need a supplier.
So, whereas we were sort of focusing on nodes that didn't really overlap with Flex Logix previously, now as customers are coming in with desire for the nodes, we're going to go and port to those, and we'll prioritize those based on customer demand, of course.
But I think that's a huge opportunity for us to really fill that void that they're leaving and really solidify QuickLogic as sort of the undisputed eFPGA leader in the market today, and I think that that's going to happen in 2025. In addition to that, continued execution on the strategic rad-hard FPGA, a lot of excitement there, both within the company and our partners.
To that extent, we are starting to engage with the defense industrial base now prior to the chip being done so that we can start getting devices and software into the hands of folks prior to the device being finished, as you would expect, and then again, for the rest of this year, 2025, you'll start to see more diversification from us in terms of end markets that we license to, in terms of process technologies that we support.
And that's all again around this whole initiative of becoming the undisputed embedded FPGA IP leader, and of course, doing that in a way that keeps us profitable and generating cash in 2025, so, with that, I will pause and see if there are any questions.
[audio distortion]
The valuation on the Flex Logix, I don't think that was disclosed.
Does that make sense to you? I mean, Analog Devices got mostly the analog space and buying that embedded FPGA business w hat are they using it for?
I've looked at the question from Quinn was, why? would ADI buy that, if I can paraphrase. They were a customer already t hey were a licensee. If I look at their product roadmaps, I don't see evidence of exactly where that is t hat tells me it was probably more of an internal use capability, not marketed as FPGA to the end market.
I think they're primarily DSP, high-speed data converter company i could see where they would want to have some level of programmability so you could start multiple SKUs to those customers. They do a lot in aerospace and defense and industrial where programmable logic is prevalent. If the company was for sale, it would be good to control their own destiny so that somebody else doesn't buy it and make it captive, and now they can't really have access to that anymore. I don't know the reasons, but these are ones I can just conjecture.
Do you remain positive for cash flow going forward after Q1, or is there seasonal factors?
With the large government contract, there's a lot of cash coming in and cash going out because we're the in-between between the government, ourselves, and all of our subs. So, I think our goal is to be, but let's wait for the call that we have in February to give a little bit more clarity on that for the year b ut that is clearly the goal. And I think we're getting to that revenue level where that's pretty achievable, especially with the fact that some of these more recent IP contracts that we're engaged on, like 18A, quite a bit higher value, which means quite a bit higher cash coming in for those IP blocks.
Brian, you're just one of the fourth phase, I guess, or the fourth contract award for that.
Yes.
Rad-Hard $72 million contract. How much of the $72 is now being allocated, and what's the purpose of the fourth phase? Is it still in the design? When you get [audio distortion]
Yeah, so for those listening online, Quinn's question is around what are the milestones? What's contemplated in these different tranches, especially the fourth tranche? It's somewhat opaque exactly what we're allowed to say, but I can tell you what we're allowed to say, so it's intended to be a four-year development phase in total from start to finish. It comprises a test chip and a final chip.
And we are hoping that we can do a press release similar to the very first press release that gives a little bit more detail beyond that we're in the fourth tranche, including things like specific roles of subs. I would like to be able to say when a test chip is going to be available, but we need to go through the process and the right channels to do a press release of that nature.
I will say that the timelines of each tranche, those are not pre-planned by us t hey're decided by our customer, the government, and so this most recent tranche in December, this fourth tranche, we actually can't disclose exactly how long that is. It definitely gets substantially into 2025.
And again, I'm hoping for the press release that we can put out will give a little bit more clarity on maybe milestones that we would all understand a little bit better than just saying we have a tranche, but I think the key is continued execution and knowing that we are allowed to say there's supposed to be a test chip and a final chip t he nearest material milestone would be having that test chip, and that's what we're gearing towards right now. Kevin, you look like you might have a question.
That's all good.
You good? Okay.
[audio distortion]
The relative profitability of the IP space in general is quite good. The first port is always the most expensive when you're bringing up a new technology node. That's already profitable for us. Any subsequent customer license is quite a bit more profitable because most of the engineering is already at sunk cost at that point. If you look at our blended model overall as a company, we've been around 60% plus or minus depending on the quarter.
The device side on the mature down here is pretty good. These are relatively old devices, kind of a cash cow business. This gray area, these were devices we did more for consumer. So, you would assume they're going to be a little bit lower margin, lower than corporate gross margin averages. The green is pretty much like right there right now because a lot of these are the first ports. Again, as we do more of those additional derivative ones this year on the green, you'll start to see some good gross margin percentages on that.
[audio distortion]
I think let me answer it in this way. The whole model with automation is designed so that as we grow with new customers, with new markets, it's not going to be a drag i t's just going to be upside t here's only so many process nodes we really need to support. We had the ones that we had, now we'll port to some ones that Flex Logix had, and I think you'll see nice operating margin at the bottom line, especially again because a lot of this is a sunk R&D. These next licenses should be more profitable.
We are designing, we're very careful. Elias does a great job of fiscal discipline within the company, spending, hiring where we need to, and we actually have grown engineering in the last year, but being really mindful to time that with customer needs and deliverables so that we're not getting too far ahead of our skis on the spending side. Yeah.
[audio distortion]
What's the break-even point on an annual basis?
[audio distortion]
Yeah.
[audio distortion]
So, I think on a, let me just do it at a quarterly granularity for a moment. Depending on product mix, you can see it's been about $6 million is the sort of profitability point, depending on the product mix within a quarter.
Six, six and a half, yeah. Quinn, you had a question.
I did. You asked the strategic process for SensiML y ou had a chart there that showed a pretty thin sliver of revenue s o, it doesn't look like there's a lot of revenue associated with it at least doesn't look like it's material to your business i s there significant costs? Like if you divest that asset, does your OPEX come down meaningfully, or is it a similar pretty small fraction of the OPEX?
It's about $1.2 million a year in OpEx. But not only do you get the OpEx out, but you get EPS about $0.06-$0.09 for the year added back in. So, it's a plus.
Yeah. Plus financially and a plus from a focus perspective.
Correct.
Double down on the things that are really growing.
Any other questions?
All right.
Brian, Faith, thank you very much for coming.
Thank you.
Thanks, everybody. Appreciate it.