Good morning and good afternoon, everyone. I'm Martin Yang from Oppenheimer. Today, we have the pleasure to be joined by Brian Faith, the CEO of QuickLogic, and Elias Nader, the CFO of QuickLogic. Thank you so much for being here, and please go ahead with your presentation.
Thanks, Martin, and thanks everybody for joining the presentation today. We're excited to be able to present our story to you, where we are with the business, and where we think it's gonna go. Hopefully, you can see the slides. We'll skip through the safe harbor statement. You can see that in the investor presentation that's posted on our website. So let's get right into it. First of all, QuickLogic is a programmable logic company. Programmable logic takes two different forms. This is all in the semiconductor space. So the traditional form of a programmable logic device is called an FPGA, field programmable gate array.
These are devices that are programmable by the customer, and they can use that programmability to extend use cases for the designs, change functionality to extend the life cycle of designs or systems, and it allows them the ability to uniquely program the hardware to do something specific for that customer. So typically, this can be used for aerospace and defense applications, industrial applications, high-end computing applications, things where the workloads are very unique to that customer. Over time, there's been a big push now, sort of a resurgence in customers doing their own unique chip design. Those are called ASICs, application-specific integrated circuits. And in that market, people are starting to look at: how do I take what functionality comes in a discrete FPGA and integrate it into my own ASIC? And that IP is called embedded FPGA.
So you can think of taking the functionality of a discrete FPGA, making it an IP that goes into somebody's ASIC. So QuickLogic has products in both of these categories. Our company was founded on FPGAs, discrete FPGAs, and a few years ago, we added to that initiative an embedded FPGA IP product category. That means that as customers traverse their needs, whether it be a discrete device or an IP for an ASIC, we actually can fulfill that need with these two product vectors that we have. Now, I mentioned aerospace and defense. That's actually been a mainstay market for programmable logic in general over the last 40 years of our industry.
But in recent times, people are looking at aerospace and defense as one of the fastest-growing segments for programmable logic, and I think we're very well-positioned to ride that wave through some of the initiatives that I'll talk a little bit more about in the slides today. So now let's take a little bit deeper view of this from a product perspective. So Australis, that you see there, that keystone in the middle, that is QuickLogic's specific capability that allows us to automate the development of both our devices and our IP so that we can be very agile in meeting the needs of the customer in a very time and cost-effective manner. Our primary business model there is on the top, which is embedded FPGA IP, remember, IP licensing.
But as customers engage with us, they oftentimes need some variation of that IP, some derivative of that, some customization of that, and we typically do charge for that in our design services arm, ultimately landing on a license of an IP to a customer or a development of a device that then we could sell to that customer or a set of customers through our device storefront. So you can see the different ways that this could pivot off of the Australis development flow. In addition to this programmable logic business, we also have an AI machine learning software business, and that is called SensiML, S-E-N-S-I-M-L, and I'll spend a little bit more time on that later in the deck. We're headquartered in San Jose, where Elias and I are calling you from right now, and again, uniquely positioned for growth.
So in the slides, you'll hear us talking about different market segments in a little bit more detail. But in total, we view that the market that we can serve with our technology is about a billion-dollar available market, and that does span different sectors that we'll go over in a minute. But this has not been happening overnight. This is something that we've been developing from a platform and a foundation over the last four years. It's driven a lot of top-line growth for us. I think the three-year CAGR, as you can see there, is 35% with our updated forecast for 2024. Puts four-year CAGR at 30% as well, and it's taken us to profitability.
More important, it's really taken us into avenues or product areas that we didn't have before, really with this IP and the storefront initiatives around aerospace and defense. So from a technology stack, I think this is really important because, again, programmable logic can take different forms for different people. For us, it all starts with the fact that we've been in this business for 30+ years, doing programmable logic devices, tens of millions of units, thousands of customers, and we truly know how to manufacture and design these devices and IPs at scale, meeting the needs of our customers. The Australis tool that we have there in gray allows us to develop and port this IP to different process nodes and foundries, and then vector that out to the sets of customers that we wanna license that to or to build devices for.
And then, as you can see in the market today, you probably hear a lot of news about the U.S. government investing in a lot of different chip or semiconductor initiatives. One of those very clearly is advanced packaging and chiplets. And what chiplets allow you to do is sort of take components of chips or small iterations of chips, much like Legos, and then stack those in an advanced package and have that be the solution that you take to market. This enables a different level of integration and flexibility for the end user to sort of piece together the functions they want in a much faster period of time, and we are squarely focused on how we're gonna participate in chiplets, both from an IP and a device perspective as well. Again, I'll cover more of those in future slides.
So let's talk about business model. The bottom part of the chart here you can see matches the opening slide that I had on the types of businesses that QuickLogic has from a product perspective. How we make money is shown on top of that. So as something that comes in, then we start doing design services for customers, that shows up as funded development, so more like NRE. We recognize that generally across the time of the contract. Again, ultimately, what we want as a product company is to license IP or sell devices. In the licensing of IP, the typical business model for us is there's an IP license that we recognize at the point of delivery.
There's also a royalty upon shipment of the devices that use the IP, and that can go on for years as long as that customer is continuing to manufacture those devices and ship them. In the case where we're running the supply chain and we're selling a device to a customer, an actual discrete device, that shows up as our device storefront revenue, and again, that's a long-tail market. As long as they need those devices, we're gonna continue to manufacture and sell them and support them. We have many examples today where people in the industrial, the aerospace, and defense industry are buying devices for literally decades. Some of the devices we sell today were first designed back in the late 1990s or early 2000s, and that sort of speaks to the longevity of these design wins once they actually go to production.
On the far right side, the SensiML AI/ML software, this is more like an IP contract in the sense that it has a SaaS subscription as you're using the service to create the AI/ML models, and then a royalty when you're actually deploying devices that use them. So when you blend this all together, you get a nice blend of having revenue in the near term with design services and IP license, and then a lot of gross margin in the long term when you have the royalties and the device shipments coming to fruition. So we like this balanced business model. Now, let's double-click a little bit here. I talked about the funded development, the leftmost category of the previous slide. Our largest contract there is very important for this company.
It's something that we've been executing on for almost 2 years now, and it's to design a strategic radiation-hardened FPGA, and our customer in this case is the Department of Defense. We are under contract to design something that's never been done before, but we're leveraging this 30 years of experience in developing programmable logic and selling to the defense industry, bringing that to bear here with these subs to us at the bottom of the chart to build this product. In total, this contract could be up to $72 million over 4 years. Again, we're almost at the 2-year mark. We just got awarded a third tranche in July, and we're excited about this. Not just because of the $72 million to design something that doesn't exist today, but what could happen after that.
This is critical, I think, for some of these strategic and space programs that are programs of record in the, the defense industry in the U.S., and we're really looking forward to wrapping up the development sometime in 2026, and then turning around and being able to be the company that sells this device to the defense industrial base for their programs of record. That's above and beyond any of the $72 million contemplated here on this slide. We're again, we're proud that we're chosen by the DoD to do this. We're executing on this and looking forward to not only the next two years, but really, what comes after that as well. As far as the served available market, you remember earlier on, I talked about it being a billion-dollar served available market. These are the primary markets that we have been selling to.
Aerospace and Defense is our largest. Second largest is Industrial. Probably third largest at this point is the Consumer side. But we're really interested in how this is gonna play out, where people are really becoming hyper-aware of security issues in systems, security issues in devices, and how they can design in more protections against these, cyber threats. And so there's definitely some use cases for programmable logic in the security side, and I think it's well-publicized that there are many use cases for the AI/ML acceleration part with programmable logic. Not every system can afford, from a cost or a power perspective, to use a, NVIDIA GPU. There's this notion about how to push more intelligence out to the edge and do it with sort of lightweight computing and acceleration.
Programmable logic is a great way to do that, and if you think about people that are doing their own custom ASICs, having the FPGA as an IP in that ASIC, again, doing lightweight acceleration, it's a great use case. So we actually have a win on that in a 12-nanometer core IP that we'll talk about more later. But when you wrap all this together, especially in the coming years, as we're executing on this strategic rad-hard program contract and some of the other IPs that we're doing today, we can see that we do have this billion-dollar served available market, available to us in the near future. So let's now double-click and talk about some of our near-term traction, and what leads us to this optimism about capturing a lot of this market.
Firstly, on the chiplet side, as I alluded to earlier, we're definitely laying down some investments now with the partner companies like YorChip, so that we can capture what we think is gonna be a very large opportunity to do programmable logic-based chiplets that people could have access to, and purchase, and integrate into systems that they're doing with advanced packaging. There's a lot of people that are gonna do their own ASIC. They're probably gonna invest $10s of millions, if not $100s of millions of dollars, doing their own ASIC. They may or may not want to have FPGA integrated as an IP forever into their ASIC. They might want it only in certain systems that they're gonna integrate, and having a chiplet-based approach gives them that flexibility to integrate programmable logic at a chiplet level when they need it, and not when they don't.
That's really the basis of what we're looking at with YorChip , and we're looking at having something sampling together as a collaboration by the end of 2025 for an FPGA-based chiplet. Even more near-term now, we've talked about two different contracts on the 12-nanometer node. One is at GlobalFoundries, one is at TSMC. Both of these have end customers that we are delivering to through 2024. The one on the 12LP is GlobalFoundries. That's a defense industrial base customer. That's on a 12-nanometer node. Again, GF 12LP is actually sort of a mainstay node for the defense industrial base, so this made a lot of sense. The second one there, TSMC 12, that is a customer that's doing a very low-power SoC. They're targeting both commercial and industrial IoT use cases.
That's one that we're looking at using the embedded FPGA from us to do some lightweight AI acceleration at the edge. And they're looking at that from the perspective of: how do they lower the power to do that acceleration? They don't wanna do it in pure software because that can be very power-consuming. They also don't wanna do pure ASIC, because the second that they do pure ASIC for a AI acceleration algorithm, the algorithm requirements are probably gonna change by the time the chip comes out. The FPGA on board gives them that forward flexibility to change the algorithms as they see fit. I skipped over Zero-Error Systems.
They have a radiation-tolerant, standard cell libraries, and they're looking at commercial space applications, so we're planning to do something with them for radiation-tolerant eFPGA IP, again, to target sort of the growing market of commercial space. And then lastly, we gave some airtime to this on our earnings call, yesterday, which is joining the Intel Foundry Accelerator IP Alliance and the US MAG Alliance. US MAG stands for United States Military Aerospace and Government Alliance. So this is a really strategically important, initiative for QuickLogic. If you look at the onshore foundries today in the U.S., 18A is by far the most advanced, process technology available.
We're doing our IP design right now for Intel 18A, in anticipation of closing customer contracts, so that as we're closing contracts, we are ready with that IP to deliver to the customers, and generate revenue, both for this fiscal year and beyond that. It's a, it's a very strategic initiative for QuickLogic. So this just gives you a snapshot of some of the key initiatives and IP contracts that we're working on, and, I'll refer back to some of these later in the presentation. So our slides are on our website in the investor page. You can go and browse the logos here later, if you wish. Some of these are partners, some of these are customers.
Again, as we're talking about things like the DoD, I think it's really important to recognize that this rad-hard contract that we're executing on, this is not the first time or first rodeo for us in the defense industry. For literally decades, we've been selling our FPGA devices, supporting them to all of the top 5 and 8 of the top 10 DoD prime contractors in the U.S., and to some extent, some of the contractors over in Europe as well. So one of which, which we're incredibly proud of, and if you're looking at the camera, you can see the banner behind us here on the wall, BAE Systems awarded us the Supplier of the Year Award in the FAST Labs technology category. And this is great.
It's exciting for us because I think it really, A, shows the value of the IP that we're bringing to the table here, and, B, it shows that as a supplier, we really know how to do business with these large defense contractors the right way, so that there's a lot of value, not only to the contractor, but to the U.S. government as well, who is ultimately funding a lot of these developments. So let's now turn to the funnel. In terms of the funnel, we look at it in two ways. So we have the book business, which in the last few years here, $37 million in new contracts related to the embedded FPGA IP initiative. On our pipeline, we talked about yesterday, the earnings call, $189 million is what it's got up to.
I think it's gone up on a net basis, every quarter since we started reporting this number, which is probably about six quarters ago or so. What we mean by pipeline, first of all, it's about a two-year horizon of the revenue potential of all of our qualified opportunities, where we have deep discussions with customers on the business level, on the technical level, often with proposals already out to these customers. We're not gonna close all the 189 to revenue. I think everybody knows that. We're refining our model as we go so that we can start coming out with the hit rates, so that we can actually use those in the future, so people get a sense of, okay, if you have your funnel, your hit rate is this, what does the projected forward revenue look like?
We're not ready to disclose that kind of number yet. I think at some point early next year we will be able to, but suffice to say, on a net basis, the funnel has been growing, and it's even been growing as we win new business, like this new rad-hard third tranche, $5.26 million. That fell out of pipeline into book business, so we had to more than replace that in the pipeline for it to grow. We have done that, and a lot of that has come from very recent proposals that we've been brought in to make proposals on for things like advanced IP and chiplet and/or device initiatives that we've been on. So all this is based on Australis, which I will get to in the next slide. So what is Australis?
I'm gonna take you on a little journey here back in time, but the old way that we used to do all of our IP, all of our device developments, really was allocating our entire engineering team for more than a year to take our architecture and port it to a new process. So think about if we had something at TSMC 40 nanometer and we wanted to port it to GF 40, or we wanted to go to GF 22 nanometer, or in these more recent cases, to 12 nanometer 18A, that was literally our entire team, all of them, over a year. So that really meant that the throughput to handle these new ports was less than one a year. Now, if you're just doing device developments, that's probably okay, 'cause you're not gonna do a new device more than once a year.
But from an IP business perspective, we had to have more scale to do more ports in more time with less people. That's how you truly create a scalable IP business. And so we found a way of doing this through automation, taking all the intellect and know-how that we have with our engineering team and applying some best practice and software automation to that to shrink that time down. So now, what we've been able to demonstrate is on the order of months, we can actually do a new process port, going again from a TSMC X node to a Samsung Y node and so forth, and we can do that with substantially less people.
So now the same engineering team has a lot more throughput and a lot more agility to meet customers kind of where they are in terms of their process and the foundry that they're choosing. So again, a lot more agility. Not only does that shrink the time to IP delivery for a customer, it also means that we can do it with lower cost, and when you combine faster to market, lower cost, you open up a bigger market, and that's really required for us to capture what we wanna capture out of this, this billion-dollar SAM. Now, for derivative IP, contracts, when we have already something supported on a given node, then it can be as little as days. If a new customer says, "Hey, you know, I want something much like that other one, can I get that?" Yeah, we can do that in days.
And they may not need it in days, but it kind of gives you a sense of the level of operating leverage we have with our flow and the team in place in order to handle that demand. Very, very scalable model. All right, so I'm gonna pause now on the programmable logic side, and we'll double-click a little bit more on SensiML. So again, SensiML is an AI software company. They were actually started inside Intel when Intel was doing low-cost IoT chips around the 2015 timeframe. Intel subsequently killed that chip, and so they had all this software and workflow to program AI models for that chip with no chip to run them on.
So the core team spun out from Intel, took the assets as part of that spin-out, immediately ported from the x86 Intel architecture to the Arm architecture that was much more pervasive in the edge IoT space, and you can see examples of those there at the bottom. We came across them as a partner company because we have a chip that actually has an Arm core and some FPGA in it, and we became partners, going after edge IoT applications. We acquired them, shortly thereafter, and we've let them run as a wholly-owned subsidiary company, continuing to try to grow the footprint of vendors and devices that use their software with their Arm, microcontrollers, primarily. Again, you can see examples there at the bottom of announced, partnerships and go-to-market strategies.
One microcontroller company has actually private branded or private labeled the SensiML Toolkit so that they can go to market with more of a complete solution to their install base and some of the new customers they're targeting, and we fully believe that this is the right way to go to market when you have a software product. You gotta go with hardware partners that, A, have a big footprint already of an install base, and B, have a large sales force to really carry that message as broadly as possible. So those are, again, ones that we've already announced via SensiML. In addition to that approach, within the last quarter, SensiML also announced an open-source strategy.
Now, it's not open-sourcing everything, it's open-sourcing elements of their toolkit that could benefit from more collective development from the community to improve those tools, to enable new features, things like generative AI, so that those have a bigger feature set or richer feature set, again, to expand the, the market. So now SensiML has a similar approach as QuickLogic. Some of the things are open source and benefit from open-source collective development, components, and there's still the option of having sort of a proprietary business model on top of that. That really suits the, the needs of, again, having critical mass on the development side and having a defendable product on the product side from which to generate revenue.
So I think our four years of figuring out the right way to do business with open-source components is actually playing out now, and SensiML is benefiting from that. All right, so let's get into some of the revenue growth over the last few years. So if you look back, now from 2020, end of then till now, in our forecast for the remainder of 2024, that would put our total revenue growth at about 30% CAGR over that four-year span. Most of that growth coming from the eFPGA IP and related developments. You can see, SensiML growing a little bit there in the yellow, and then 2023, sort of a correction from the smartphone market, in the gray line there, recovering a little bit this year, mature, recovering a little bit this year.
All in all, netting slightly under the $25 million revenue line for 2024 forecast. So pretty happy with the revenue growth overall over the last four years. I think we're incredibly proud of this slide because this speaks to the operating leverage that we get through the automation that is very unique to QuickLogic. Again, remember in the past, if we wanted to do all these designs, we'd have to hire like crazy in the OpEx side. We've hired very prudently, and with that automation, it's really taken us to the profitability line with essentially the same OpEx, but a lot more revenue, as you saw from the last slide.
Some of that is coming from the fact that it's more profitable to have IP, and royalties, and some of these other developments, but it certainly helps the fact that we've been able to keep OpEx effectively flat with that automation that we have. And so that really takes us to the punchline slide here. I think the additional traction that we've talked about throughout this year on the earnings call yesterday, our forecast for full year profitability again this year, as it was last year, it's gonna come from closing these contracts that are in the funnel, continued execution on this really large strategic rad-hard contract, more expansion in the SensiML revenue line, both from private label and individual customers. And again, that leading to the forecasted annual revenue growth this year.
We revised it on a call yesterday from 30%-15%. Still, that would net a 30% CAGR over the last 4 years. But I think more importantly, still forecasting annual profitability and then positive cash flow for the second half, as a result of, again, top line growth while we maintain really, a good optics on the OpEx side. So with that, I will say thank you, and I will turn it back over to you, Martin, to guide the Q&A.
Thank you, Brian. So first question, immediately following your quarterly results, can you maybe talk about the push-out, the nature of the push-out, and do you expect the... Are most of the push-out relating to the final payment, regarding the licensing portion?
Sure. So there's a few things in there we can double-click on. So let's start with Q2. Obviously, we're a little bit below the midpoint of the guidance. That was actually a few product lines that contributed to that. A little bit on the IP contract side and some on the device side, specifically aerospace and defense devices that we've been selling for a long time to a handful of customers, and some of which have pushed out the demand that they had told us about being Q2, now being in Q4. If we look at the balance of the year, we effectively took $3 million off the top line in our call yesterday for 2024 forecast, and that's really comprised of one particular customer that we've already done a custom FPGA for. We talked about this November 2022.
Further work on that is paused probably until 2025, in some part of the early part of the year, because there is a related subcontractor that is delayed on some deliverables that we need to have done before we can continue that work in earnest. So that pushes some of that expected revenue into 2025. The remainder of that $3 million is really comprised of a handful of, I would say, smaller IP deals, but collectively, them pushing to the right, some into Q4 of this year, some into early 2025. A little bit more color on those, aerospace and defense and industrial would be the market segments. Most of those are international customers.
One of the reasons for that falls into two categories: one, the unavailability of project funding yet to move forward, or two, other developments that make those engineering teams unavailable to start on those chip designs that would need our IP. Nobody's gonna buy IP if it's too early before it's needed for their chip design schedule. So that's really the bulk of those revenue changes that we talked about from prior earnings calls. Now, you mentioned cash payment. So cash usage in Q2 was higher than we expected, and really, that comes down to the timing of payments related to the strategic rad-hard contract. We forecasted things to be a little bit more into Q2 or second half. The sort of confluence events came together in the end of Q2, so higher cash usage as a result.
That primarily is due to the timing of the third tranche of the rad-hard contract that we signed. But as you heard on the earnings call yesterday for guidance for Q3, essentially flat on cash burn, up to maybe $500K, but then for the rest of the year to be positive on cash flow. So I think we're good on that front. Q2 was just a challenging one for us.
Got it. Next question, maybe a longer-term question on revenue composition.
Yeah.
You know, you have successfully built out a very strong pipeline based on open source toolchain. So maybe comment on both the near term and the longer-term revenue composition. Near term, do you expect revenues to be more coming from design services plus IP? And how do you expect. If that's true, how do you expect that revenue composition to shift or evolve into the longer term?
Yeah. So firstly, let me say, what we want in the long term, we want a lot of royalty contribution because that's 100% gross margin. We want a lot of device revenue contribution because that's also good gross margin dollars, long tail of revenue, long into the future. So as we're doing any design today for a customer, we are always thinking about: how do we best position the business for that growth in royalty and device revenue, device gross margins in the future? So with that being said, now let's zoom in to the here and now. Clearly, a bulk of our revenue today is in the funded development for the strategic rad-hard contract, that one that was a 2022 customer take out and IP license.
That's driving the bulk of it today because those elements are actually things that we could recognize as we're doing the work in the case of funded development or at IP delivery, in the case of IP license. Once we deliver that IP, we get revenue. Customer could take 1 year or 2, 3 years in order to finish their chip and go to production, at which point we take the royalty in the future. So we're always trying to structure these, these contracts and businesses to maximize the near-term revenue potential, but not give up any value that we want in these IP contracts. So in the next couple of years, most of the revenue will continue to be from funded development and IP license because of the nature they are nearer in the engagement....
Outside of that, especially once that strategic rad-hard development contract is completed, that's when you can really start to see device revenue, royalties driving that. And frankly, when the strategic rad-hard one is completed and we're selling that to the DIB or some of these other chiplet ones that we're talking about being completed, that's when you start to see a step function increase in revenue potential. Because the price of a device far exceeds any level of royalty that we get as a percentage of that. And so that's why we're always trying to balance those two as we're rolling out with these new proposals to customers.
Thanks, Brian. My next question is on an overview of your supported process nodes.
Yeah.
Can you give us, remind us, you know, how many different foundries and processes do you support? And also, in that context, how important is the upcoming qualification at Intel's 18A for you, for the longer term?
A lot of good questions in there. So from a foundry perspective, we've already talked publicly about the ones you see on this chart. So GlobalFoundries, SkyWater, UMC, TSMC, Honeywell's actually not there. Honeywell, Samsung, Intel Foundry. As far as the process nodes that we support today, you'll see a lot of the, what I'll call mainstream nodes, like 22, 28. The 12 nanometer ones that we talked about earlier this year, so GlobalFoundries 12, TSMC 12, those are also pretty mainstream, but starting to push into the more aggressive. Intel 18A is by far the most aggressive that we've worked with, most advanced. All of those really, I think, carry a higher value.
There are gonna be very sophisticated system on chips or ASICs that are designed with those process technologies, and that's what really excites us, I think, from a value perspective and time and market perspective, because of the, the value of those. That's not even talking about the strategic importance of some of these. If you think about the fact that we have a big play with aerospace and defense, clearly, the U.S. government is investing a lot in onshore manufacturing. And this is not just about creating jobs. That's important, but it's also just from a national security perspective, having the ability to manufacture advanced process technology wafers onshore is of critical importance. And so you see GlobalFoundries at 12, you see Intel Foundry at 18A. These are gonna be mainstay nodes for that market for probably decades.
And that's why, from a strategic perspective, it's very important that we have IP cores and the ability to do customer-specific variants of that on those, those process technologies. So that's in fact why when we had the opportunity to do Intel 18A, we jumped at it. I mean, that's gonna be around for a long time, high value. You can see online in research the types of entities that are doing chip design for that. I think even today or yesterday, there was an article that Draper put out about how important 18A was for the defense community and what they're doing around 18A to advance that from an advanced packaging perspective. So I think it's, it's gonna have a lot of potential, both for the government, industry, and then for QuickLogic as well.
So now, if you go to 18A and you're talking about these other nodes we've talked about, there's some air gap in between 12 and 18A. And as we get other customer opportunities coming in, our flow really is immutable to us, taking what we've done at those nodes and adapting it for whatever may have demand for in the future. We're very agile, and we're very driven by what the market and what customers want in order to do commit resources for this, and we're happy to do that if the business is there.
Got it. Thank you, Brian. Before I move on to the next question, a quick reminder to the online audience: if you have any questions, feel free to type in your questions through our online portal. Next question is regarding your recent announcements with distributor and fulfillment partners. Can you maybe dive deeper into those announcements? How do those announcement mean for your business?
Yeah. So I think, you know, anybody that's been involved in any business, certainly in the semiconductor industry, you wanna try to leverage external sales forces to the extent you can, because that really allows you to focus your OpEx on your core and use other resources to bring in these deals, help them close deals, help expand your sales force. So we're no different. I think we're at a point now where we're getting so much traction and interest, sort of outside the U.S., inside the U.S., with our technology, that we've hired the internal people that we feel we need in order to execute this, and now it's time to branch out and extend that through distribution partners. But you really have to be careful when you look at distribution partners. For us, this is a design-in product.
It is not a commodity, it's not a memory chip or a capacitor. You really need people that, A, have very deep roots serving a specific customer base, so they're a trusted supplier from a distribution perspective. The second is you need somebody that is gonna allocate enough of the technical resources that they can pre-qualify opportunities, so that by the time it comes to us and our direct team, they're gonna be pre-qualified, and we can then move them forward through the funnel. So with that background, we've talked about three new distributors on the call yesterday, and one of them, CTG, is about U.S.-centric. They have a long, long history of serving the defense industrial base, and I think they're very trusted in that sense.
They're accessing, you know, lots of different programs of record, lots of different program offices, and I think they're a great resource to help expand the reach that we have through our direct channel to new programs of record that we can then go in and, and open up opportunities for. On the international side, same thing. You can see both of those distributors have footprints in aerospace and defense. And so there are, you know, big pushes around the world, and there's a lot of investment going on, especially in aerospace, where they're trying to take products, design their own ASICs to serve their interests. And we're, you know, doing that.
We have, I think, a very good export lawyer that's helping us guide through all this, so we know that we're able to export and do it in the correct way to the correct companies and the correct geos. That actually drove the strategy of what distributors we wanted to bring on board to expand that footprint. Again, leveraging from all the success that we're seeing now in these recent contracts, now's the time to invest more in that from a go-to-market perspective and grow that top of funnel.
Thank you, Brian.
Yeah.
Next question is for Elias. So assuming we maintain a pretty robust, you know, 30% revenue CAGR, and how should we think about margin profile to in the medium to longer term? And how should we think about the revenue growth rate in relation to OpEx, and how much operating leverage is there in your long-term model?
I think on gross margin, you would look at it from what we said yesterday on the call, which is mid-60s for 2024. I believe I have stated in many calls in the past that no one should rely on a 73% or 78% gross margin that we had announced, because of certain mix and certain reclassification of OpEx and COGS. So going forward, I think our goal internally is to hit mid-60s and high 60s, and I think that's the way you kind of look at gross margin. In terms of revenue, how you look at revenue mix versus OpEx, we really are investing in headcount, in people, in engineers. We've hired a local North American director of sales, who we believe was needed for leveraging.
And second, we have hired a bunch of engineers for rad-hard technology in anticipation of many of the projects and agreements that Brian has talked about that are included in the sales funnel that we're excited about, and hopefully we'll land a few of them. So we've already started the investment process. We really don't have much to spend on, such as CapEx, and therefore, I believe OpEx will be flattish to maybe around 3.2 average in 2025. But for now, the remainder of 2024, I believe will be around the $3 million mark or below.
Thank you, Elias. And with that, I see no more questions from the audience, and that's... Well, before we close, is there any remarks you would like to make, Brian or Elias?
None other than I've already said. I think we're really excited about where we are, where the funnel is, where we are with some of these proposals that we're looking to close. And I would say just keep an eye open for announcements that we're making on that between now and sometime in the late fall, that would be contributing to revenue this year and early next year. So thank you for your time.
Great.
Look forward to seeing you at some point in the future.
Thanks, everyone, and thank you, Brian. Thank you, Elias. Let's wrap. Bye-bye. Good morning and good afternoon being here, and please go ahead with your presentation.
Thanks, Martin, and thanks everybody for joining the presentation today. We're excited to be able to present our story to you, where we are with the business, and where we think it's gonna go. Hopefully, you can see the slides. We'll skip through the safe harbor statement. You can see that in the investor presentation that's posted on our website. So let's get right into it. First of all, QuickLogic is a programmable logic company. Programmable logic takes two different forms. This is all in the semiconductor space. So the traditional form of a programmable logic device is called an FPGA, field programmable gate array.
These are devices that are programmable by the customer, and they can use that programmability to extend use cases for the designs, change functionality to extend the life cycle of designs or systems, and it allows them the ability to uniquely program the hardware to do something specific for that customer. Typically, this can be used for aerospace and defense applications, industrial applications, high-end computing applications, things where the workloads are very unique to that customer. Over time, there's been a big push now, sort of a resurgence in customers doing their own unique chip design. Those are called ASICs, application-specific integrated circuits. And in that market, people are starting to look at, "How do I take what functionality comes in a discrete FPGA and integrate it into my own ASIC?" And that IP is called embedded FPGA.
So you can think of taking the functionality of a discrete FPGA, making it an IP that goes into somebody's ASIC. So QuickLogic has products in both of these categories. Our company was founded on FPGAs, discrete FPGAs, and a few years ago, we added to that initiative an embedded FPGA IP product category. That means that as customers traverse their needs, whether it be a discrete device or an IP for an ASIC, we actually can fulfill that need with these two product vectors that we have. Now, I mentioned aerospace and defense. That's actually been a mainstay market for programmable logic in general over the last 40 years of our industry.
But in recent times, people are looking at aerospace and defense as one of the fastest-growing segments for programmable logic, and I think we're very well-positioned to ride that wave through some of the initiatives that I'll talk a little bit more about in the slides today. So now let's take a little bit deeper view of this from a product perspective. So Australis, that you see there, that keystone in the middle, that is QuickLogic-specific capability that allows us to automate the development of both our devices and our IP so that we can be very agile in meeting the needs of the customer in a very time- and cost-effective manner. Our primary business model there is on the top, which is embedded FPGA IP. Remember IP licensing.
But as customers engage with us, they oftentimes need some variation of that IP, some derivative of that, some customization of that, and we typically do charge for that in our design services arm, ultimately landing on a license of an IP to a customer or a development of a device that then we could sell to that customer or a set of customers through our device storefront. So you can see the different ways that this could pivot off of the Australis development flow. In addition to this programmable logic business, we also have an AI machine learning software business, and that is called SensiML, S-E-N-S-I-M-L. And I'll spend a little bit more time on that later in the deck. We're headquartered in San Jose, where Elias and I are calling you from right now, and again, uniquely positioned for growth.
So in the slides, you'll hear us talking about different market segments in a little bit more detail. But in total, we view that the market that we can serve with our technology is about a billion-dollar available market, and that does span different sectors that we'll go over in a minute. But this has not been happening overnight. This is something that we've been developing from a platform and a foundation over the last four years. It's driven a lot of top-line growth for us. I think the three-year CAGR, as you can see there, is 35% with our updated forecast for 2024, but it's four-year CAGR at 30% as well, and it's taken us to profitability.
And more important, it's really taken us into avenues or product areas that we didn't have before, really with this IP and the storefront initiatives around aerospace and defense. So from a technology stack, I think this is really important because, again, programmable logic can take different forms for different people, and for us, it all starts with the fact that we've been in this business for 30+ years, doing programmable logic devices, tens of millions of units, thousands of customers, and we truly know how to manufacture and design these devices and IPs at scale, meeting the needs of our customers.
The Australis tool that we have there in grey allows us to develop and port this IP to different process nodes and foundries and then vector that out to the sets of customers that we want to license that to or to build devices for. As you can see in the market today, you probably hear a lot of news about the U.S. government investing in a lot of different chip or semiconductor initiatives. One of those very clearly is advanced packaging and chiplets. What chiplets allow you to do is sort of take components of chips or small iterations of chips, much like like Legos, and then stack those in an advanced package and have that be the solution that you take to market.
This enables a different level of integration and flexibility for the end user to sort of piece together the functions they want in a much faster period of time, and we are squarely focused on how we're gonna participate in chiplets, both from an IP and a device perspective as well. Again, I'll cover more of those in future slides. So let's talk about business model. The bottom part of the chart here, you can see, matches the opening slide that I had on the types of businesses that QuickLogic has from a product perspective. How we make money is shown on top of that. So as something that comes in, then we start doing design services for customers. That shows up as funded development, so more like NRE. We recognize that generally across the time of the contract.
Again, ultimately, what we want as a product company is to license IP or sell devices. In the licensing of IP, the typical business model for us is there's an IP license that we recognise at the point of delivery. There's also a royalty upon shipment of the devices that use the IP, and that can go on for years as long as that customer is continuing to manufacture those devices and ship them. In the case where we're running the supply chain and we're selling a device to a customer, an actual discrete device. That shows up as our device storefront revenue. And again, that's a long tail market. As long as they need those devices, we're gonna continue to manufacture and sell them and support them.
We have many examples today, where people in the industrial, and the aerospace, and defense industry are buying devices for literally decades. Some of the devices we sell today were first designed back in the late 1990s or early 2000s, and that sort of speaks to the longevity of these design wins once they actually go to production. On the far right side, the SensiML AI/ML software, this is more like an IP contract in the sense that it has a SaaS subscription as you're using the service to create the AI/ML models, and then a royalty when you're actually deploying devices that use them.
So when you blend this all together, you get a nice blend of having revenue in the near term with design services and IP license, and then a lot of gross margin in the long term when you have the royalties and the device shipments, coming to fruition. So we like this balanced, business model. Now, let's double-click a little bit here. I talked about the funded development, the leftmost category of the previous slide. Our largest contract there is, very important for this company. It's something that we've been executing on for almost two years now, and it's to design a strategic radiation-hardened FPGA, and our customer in this case is the Department of Defense.
We are under contract to design something that's never been done before, but we're leveraging this 30 years of experience in developing programmable logic and selling to the defense industry, bringing that to bear here with these, subs to us at the bottom of the chart to build this, this product. In total, this contract could be up to $72 million over 4 years. Again, we're almost at the 2-year mark. We just got awarded a third tranche in July, and, we're excited about this. Not just because of the $72 million to design something that doesn't exist today, but what could happen after that.
This is critical, I think, for some of these strategic and space programs that are programs of record in the defense industry in the U.S., and we're really looking forward to wrapping up the development sometime in 2026, and then turning around and being able to be the company that sells this device to the defense industrial base for their programs of record. And that's above and beyond any of the $72 million contemplated here on this slide. So we're... Again, we're proud that we're chosen by the DoD to do this. We're executing on this, and looking forward to not only the next two years, but really, what comes after that as well. As far as the served available market, you remember earlier on, I talked about it being a billion-dollar served available market. These are the primary markets that we have been selling to.
Aerospace and Defense is our largest, second largest is Industrial, probably third largest at this point is the Consumer side. But we're really interested in how this is gonna play out, where people are really becoming hyper-aware of security issues in systems, security issues in devices, and how they can design in more protections against these, cyber threats. And so there's definitely some use cases for programmable logic in the security side, and I think it's well-publicized that there are many use cases for the AI/ML acceleration part with programmable logic. Not every system can afford, from a cost or a power perspective, to use a NVIDIA GPU. There's this notion about how to push more intelligence out to the edge and do it with sort of lightweight computing and acceleration.
Programmable logic is a great way to do that, and if you think about people that are doing their own custom ASICs, having the FPGA as an IP in that ASIC, again, doing lightweight acceleration, it's a great use case. And we actually have a win on that in a 12-nanometer core IP that we'll talk about more later. But when you wrap all this together, especially in the coming years, as we're executing on this strategic rad-hard program contract and some of the other IPs that we're doing today, we can see that we do have this billion-dollar served available market available to us in the near future. So let's now double-click and talk about some of our near-term traction, and what leads us to this optimism about capturing a lot of this market.
Firstly, on the chiplet side, as I alluded to earlier, we're definitely laying down some investments now with partner companies like YorChip, so that we can capture what we think is gonna be a very large opportunity to do programmable logic-based chiplets that people could have access to, and purchase, and integrate into systems that they're doing with advanced packaging. There's a lot of people that are gonna do their own ASIC. They're probably gonna invest $10s of millions, if not $100s of millions, doing their own ASIC. They may or may not want to have FPGA integrated as an IP forever into their ASIC. They might want it only in certain systems that they're gonna integrate. And having a chiplet-based approach gives them that flexibility to integrate programmable logic at a chiplet level when they need it, and not when they don't.
And that's really the basis of what we're looking at with YorChip, and we're looking at having something sampling together as a collaboration by the end of 2025 for an FPGA-based chiplet. Even more near term now, we've talked about 2 different contracts on the 12-nanometer node. One is at GlobalFoundries, one is at TSMC. Both of these have end customers that we are delivering to through 2024. The one on the 12 LP is GlobalFoundries. That's a defense industrial base customer. That's on a 12-nanometer node. Again, GF 12LP is actually sort of a mainstay node for the defense industrial base, so this made a lot of sense. The second one there, TSMC 12, that is a customer that's doing a very low-power SoC. They're targeting both commercial and industrial IoT use cases.
That's one that we're looking at using the embedded FPGA from us to do some lightweight AI acceleration at the edge. And they're looking at that from the perspective of how do they lower the power to do that acceleration? They don't wanna do it in pure software, because that could be very power-consuming. They also don't wanna do pure ASIC, because the second that they do pure ASIC for an AI acceleration algorithm, the algorithm requirements are probably gonna change by the time the chip comes out. The FPGA on board gives them that forward flexibility to change the algorithms as they see fit. I skipped over Zero-Error Systems.
They have a radiation-tolerant, standard cell libraries, and they're looking at commercial space applications, so we're planning to do something with them for radiation-tolerant eFPGA IP, again, to target sort of the growing market of commercial space. And then lastly, we gave some airtime to this on our earnings call, yesterday, which is joining the Intel Foundry Accelerator IP Alliance and the USMAG Alliance . U.S. MAG stands for United States Military Aerospace and Government Alliance. So this is a really strategically important, initiative for QuickLogic. If you look at the onshore foundries today in the U.S., 18A is by far the most advanced, process technology available.
We're doing our IP design right now for Intel 18A, in anticipation of closing customer contracts, so that as we're closing contracts, we are ready with that IP to deliver to the customers, and generate revenue, both for this fiscal year and beyond that. It's a, it's a very strategic initiative for QuickLogic. So this just gives you a snapshot of some of the key initiatives and IP contracts that we're working on, and, I'll refer back to some of these later in the presentation. So our slides are on our website in the investor page. You can go and browse the logos here later, if you wish. Some of these are partners, some of these are customers.
Again, as we're talking about things like the DoD, I think it's really important to recognize that this rad-hard contract that we're executing on, this is not the first time or first rodeo for us in the defense industry. For literally decades, we've been selling our FPGA devices, supporting them to all of the top 5 and 8 of the top 10 DoD prime contractors in the U.S., and to some extent, some of the contractors over in Europe as well. So one of which, which we're incredibly proud of, and if you're looking at the camera, you can see the banner behind us here on the wall, BAE Systems awarded us the Supplier of the Year Award in the FAST Labs technology category. This is great.
It's exciting for us because I think it really, A, shows the value of the IP that we're bringing to the table here, and B, it shows that as a supplier, we really know how to do business with these large defense contractors the right way, so that there's a lot of value, not only to the contractor, but to the U.S. government as well, who is ultimately funding a lot of these developments. So let's now turn to the funnel. In terms of the funnel, we look at it in two ways. So we have the book business, which in the last few years here, $37 million in new contracts related to the embedded FPGA IP initiative. On our pipeline, we talked about yesterday, the earnings call, $189 million is what it's got up to.
I think it's gone up on a net basis, every quarter since we started reporting this number, which is probably about 6 quarters ago or so. What we mean by pipeline, first of all, it's about a two-year horizon of the revenue potential of all of our qualified opportunities, where we have deep discussions with customers on the business level, on the technical level, often with proposals already out to these customers. We're not gonna close all the 189 to revenue. I think everybody knows that. We're refining our model as we go so that we can start coming out with the hit rates, so that we can actually use those in the future, so people get a sense of, okay, if you have your funnel, your hit rate is this, what does the projected forward revenue look like?
We're not ready to disclose that kinda number yet. I think at some point early next year we will be able to, but suffice to say, on a net basis, the funnel has been growing, and it's even been growing as we win new business, like this new rad-hard third tranche, $5.26 million. That fell out of pipeline into book business. So we had to more than replace that in the pipeline for it to grow. We have done that, and a lot of that has come from very recent proposals that we've been brought into to make proposals on for things like advanced IP and chiplet and/or device initiatives that we've been on. So all this is based on Australis, which I will get to in the next slide. So what is Australis?
I'm gonna take you on a little journey here back in time, but the old way that we used to do all of our IP, all of our device developments, really was allocating our entire engineering team for more than a year to take our architecture and port it to a new process. So think about if we had something at TSMC 40 nanometer, and we wanted to port it to GF 40, or we wanted to go to GF 22 nanometer, or in these more recent cases, to 12 nanometer or 18A, that was literally our entire team, all of them, over a year. So that really meant that the throughput to handle these new ports was less than one a year. Now, if you're just doing device developments, that's probably okay, 'cause you're not gonna do a new device more than once a year.
But from an IP business perspective, we had to have more scale to do more ports in more time with less people. That's how you truly create a scalable IP business. And so we found a way of doing this through automation, taking all the intellect and know-how that we have with our engineering team and applying some best practice and software automation to that to shrink that time down. So now, what we've been able to demonstrate is on the order of months, we can actually do a new process port, going again from a TSMC X node to a Samsung Y node, and so forth, and we can do that with substantially less people.
So now this same engineering team has a lot more throughput and a lot more agility to meet customers kinda where they are in terms of their process and the foundry of their choosing. So again, a lot more agility. Not only does that shrink the time to IP delivery for a customer, it also means that we can do it with lower cost, and when you combine faster to market, lower cost, you open up a bigger market. And that's really required for us to capture what we wanna capture out of this, this $1 billion SAM. Now, for derivative IP contracts, when we have already something supported on a given node, then it can be as little as days. If a new customer says, "Hey, you know, I want something much like that other one, can I get that?" Yeah, we can do that in days.
They may not need it in days, but it kind of gives you a sense of the level of operating leverage we have with our flow and the team in place in order to handle that demand. Very, very scalable model. All right, so I'm gonna pause now on the programmable logic side. I'm gonna go double-click a little bit more on SensiML. So again, SensiML is an AI software company. They were actually started inside Intel when Intel was doing low-cost IoT chips around the 2015 timeframe. Intel subsequently killed that chip, and so they had all this software and workflow to program AI models for that chip with no chip to run them on.
So the core team spun out from Intel, took the assets as part of that spin-out, immediately ported from the x86 Intel architecture to the Arm architecture that was much more pervasive in the edge IoT space, and you can see examples of those there at the bottom. We came across them as a partner company because we have a chip that actually has an Arm core and some FPGA in it, and we became partners, going after edge IoT applications. We acquired them, shortly thereafter, and we've let them run as a wholly-owned subsidiary company, continuing to try to grow the footprint of vendors and devices that use their software with their Arm, microcontrollers primarily. Again, you can see examples there at the bottom of announced, partnerships and go-to-market strategies.
One microcontroller company is actually private branded or private labeled the SensiML Toolkit, so that they can go to market with more of a complete solution to their install base and some of the new customers they're targeting. We fully believe that this is the right way to go to market when you have a software product. You've gotta go with hardware partners that, A, have a big footprint already of an install base, and B, have a large sales force to really carry that message as broadly as possible. So those are, again, ones that we've already announced via SensiML. In addition to that approach, within the last quarter, SensiML also announced an open source strategy.
Now, it's not open sourcing everything, it's open sourcing elements of their toolkit that could benefit from more collective development from the community to improve those tools, to enable new features, things like generative AI, so that those have a bigger feature set or richer feature set, again, to expand the, the market. So now SensiML has a similar approach as QuickLogic. Some of the things are open source and benefit from open source collective development, components, and there's still the option of having sort of a proprietary business model on top of that. That really suits the, the needs of, again, having critical mass on the development side and having a defendable product on the product side from which to generate revenue.
So I think our four years of figuring out the right way to do business with open source components is actually playing out now, and SensiML is benefiting from that. All right, let's get into some of the revenue growth over the last few years. If you look back now from 2020, end of then till now, in our forecast for the remainder of 2024, that would put our total revenue growth at about 30% CAGR over that four-year span. Most of that growth coming from the eFPGA IP and related developments. You can see SensiML growing a little bit there in the yellow, and in 2023, sort of a correction from the smartphone market in the gray line there. Recovering a little bit this year.
All in all, netting slightly under the $25 million revenue line for 2024 forecast. So pretty happy with the revenue growth, overall over the last four years. I think we're incredibly proud of this slide because this speaks to the operating leverage that we get through the automation that is very unique to QuickLogic. Again, remember in the past, if we wanted to do all these designs, we'd have to hire like crazy in the OpEx side. We've hired very prudently, and with that automation, it's really taken us to the profitability line with essentially the same OpEx, but a lot more revenue, as you saw from the last slide.
Some of that is coming from the fact that it's more profitable to have IP, and royalties, and some of these other developments, but it certainly helps the fact that we've been able to keep OpEx effectively flat, with that automation that we have. So that really takes us to the punchline slide here. I think the additional traction that we've talked about throughout this year on the earnings call yesterday, our forecast for full year profitability again this year, as it was last year, it's gonna come from closing these contracts that are in the funnel, continued execution on this really large strategic rad-hard contract, more expansion in the SensiML revenue line, both from private label and individual customers. And again, that leading to the forecasted annual revenue growth this year.
We revised it on a call yesterday from 30%-15%. Still, that would net 30% CAGR over the last four years. But I think more importantly, still forecasting annual profitability and then positive cash flow for the second half, as a result of, again, top line growth, while we maintain really, a good optics on the OpEx side. So with that, I will say thank you, and I will turn it back over to you, Martin, to guide the Q&A.
Thank you, Brian. So first question, immediately following your quarterly results, can you maybe talk about the push-out, the nature of the push-out, and do you expect the... Are most of the push-out relating to the final payment regarding the licensing portion?
... Sure. So there's a few things in there we can double-click on. So let's start with Q2. Obviously, we're a little bit below the midpoint of the guidance. That was actually a few product lines that contributed to that. A little bit on the IP contract side and some on the device side, specifically aerospace and defense devices that we've been selling for a long time to a handful of customers, and some of which have pushed out the demand that they had told us about being Q2, now being in Q4. If we look at the balance of the year, we effectively took $3 million off the top line in our call yesterday for 2024 forecast, and that's really comprised of one particular customer that we've already done a custom FPGA for. We talked about this November 2022.
Further work on that is paused probably until 2025, in some part of the early part of the year, because there is a related subcontractor that is delayed on some deliverables that we need to have done before we can continue that work in earnest. And so that pushes some of that expected revenue into 2025. The remainder of that, $3 million is really comprised of a handful of, I would say, smaller IP deals, but collectively, them pushing to the right, some into Q4 of this year, some into really 2025. A little bit more color on those, aerospace and defense and industrial would be the market segments. Most of those are international customers.
One of the reasons for that falls into two categories: one, the unavailability of project funding yet to move forward, or two, other developments that make those engineering teams unavailable to start on those chip designs that would need our IP. Nobody's gonna buy IP if it's too early before it's needed for their chip design schedule. So that's really the bulk of those revenue changes that we talked about from prior earnings calls. Now, you mentioned cash payment. So cash usage in Q2 was higher than we expected, and really, that comes down to the timing of payments related to the strategic rad-hard contract. We forecasted things to be a little bit more into Q2 or second half. The sort of confluence events came together in the end of Q2, so higher cash usage as a result.
That primarily is due to the timing of the third tranche of the rad-hard contract that we signed. But as you heard on the earnings call yesterday for guidance for Q3, essentially flat on cash burn, up to maybe $500,000, but then for the rest of the year to be positive on cash flow. So I think we're good on that front. Q2 was just a challenging one for us.
Got it. Next question, maybe a longer-term question on revenue composition.
Yeah.
You know, you have successfully built out a very strong pipeline based on open source tool chain. So maybe comment on both the near-term and the longer-term revenue composition. Near term, do you expect revenues to be more coming from design services plus IP? And how do you expect—If that's true, how do you expect that revenue composition to shift or evolve into the longer term?
Yeah. So firstly, let me say, what we want in the long term, we want a lot of royalty contribution because that's 100% gross margin. We want a lot of device revenue contribution because that's also good gross margin dollars, long tail revenue, long into the future. So as we're doing any design today for a customer, we are always thinking about: how do we best position the business for that growth in royalty and device revenue, device gross margins in the future? So with that being said, now let's zoom in to the here and now. Clearly, a bulk of our revenue today is in the funded development for the strategic rad-hard contract, that one that was the 2022 customer tape-out and IP license.
That's driving the bulk of it today because those elements are actually things that we could recognize as we're doing the work, in the case of funded development, or at IP delivery, in the case of IP license. Once we deliver that IP, we get revenue. Customer could take a year or two, three years in order to finish their chip and go to production, at which point we take the royalty in the future. So we're always trying to structure these, these contracts and businesses to maximize the near-term revenue potential, but not give up any value that we want, in these IP contracts. So in the, in the next couple of years, most of the revenue will continue to be from funded development and IP license because of the nature they are nearer in the engagement.
Outside of that, especially once that strategic rad-hard development contract is completed, that's when you can really start to see device revenue, royalties driving that. And frankly, when the strategic rad-hard one is completed, and we're selling that to the DIB or some of these other chiplet ones that we're talking about being completed, that's when you start to see a step function increase in revenue potential, because the price of a device far exceeds any level of royalty that we get as a percentage of that. And so that's why we're always trying to balance those two as we're rolling out with these new proposals to customers.
Thanks, Brian. My next question is on an overview of your supported process nodes.
Yeah.
Can you give us, remind us, you know, how many different foundries and processes do you support? And also, in that context, how important is, the upcoming, qualification at Intel's 18A for you, for the longer term?
A lot of great questions in there. So from a foundry perspective, we've already talked publicly about the ones you see on this chart. So GlobalFoundries, SkyWater, UMC, TSMC, Honeywell is actually not there. Honeywell, Samsung, Intel Foundry. As far as the process nodes that we support today, you'll see a lot of the, what I'll call mainstream nodes, like 22, 28. The 12-nanometer ones that we talked about earlier this year, so GlobalFoundries 12, TSMC 12, those are also pretty mainstream, but starting to push into the more aggressive. Intel 18A is by far the most aggressive that we've worked with, most advanced. All of those really, I think, carry a higher value.
There are gonna be this very sophisticated system on chips or ASICs that are designed with those process technologies, and that's what really excites us, I think, from a value perspective and time and market perspective, because of the, the value of those. That's not even talking about the strategic importance of some of these. If you think about the fact that we have a big play with aerospace and defense, clearly the U.S. government is investing a lot in onshore manufacturing, and this is not just about creating jobs. That's important, but it's also just from a national security perspective, having the ability to manufacture advanced process technology wafers onshore is of critical importance. And so you see GlobalFoundries at 12, you see Intel Foundry at 18A. These are gonna be mainstay nodes for that market for probably decades.
That's why, from a strategic perspective, it's very important that we have IP cores and the ability to do customer-specific variants of that on those process technologies. So that's in fact why when we had the opportunity to do Intel 18A, we jumped at it. I mean, that's gonna be around for a long time, high value. You can see online in research the types of entities that are doing chip design for that. I think even today or yesterday, there was an article that Draper put out about how important 18A was for the defense community and what they're doing around 18A to advance that from an advanced packaging perspective. So I think it's gonna have a lot of potential, both for the government, industry, and then for QuickLogic as well.
Now, if you go to 18A and you're talking about these other nodes we've talked about, there's some air gap in between 12 and 18A. As we get other customer opportunities coming in, our flow really is important to us, taking what we've done at those nodes and adapting it for whatever may have demand for in the future. We're very agile, and we're very driven by what the market and what customers want, in order to commit resources for this, and we're happy to do that if the business is there.
Got it. Thank you, Brian. And before I move on to the next question, a quick reminder to the online audience, if you have any questions, feel free to type in your questions through our online portal. Next question is regarding your recent announcements with distribution, distributor and fulfillment partners. Can you maybe dive deeper into those announcements? How do those announcement mean for your business?
Yeah. So I think, you know, anybody that's been involved in any business, certainly in the semiconductor industry, you wanna try to leverage external sales forces to the extent you can, because that really allows you to focus your OpEx on your core and use other resources to bring in these deals, help them close deals, help expand your sales force. So we're no different. I think we're at a point now where we're getting so much traction and interest, sort of outside the U.S., inside the U.S., with our technology, that we've hired the internal people that we feel we need in order to execute this, and now it's time to branch out and extend that through distribution partners. But you really have to be careful when you look at distribution partners. For us, this is a design-in product.
It is not a commodity, it's not a memory chip or a capacitor. You really need people that, A, have very deep roots serving a specific customer base, so they're a trusted supplier from a distribution perspective. The second is you need somebody that is gonna allocate enough of the technical resources that they can pre-qualify opportunities, so that by the time it comes to us and our direct team, they're gonna be pre-qualified, and we can then move them forward through the funnel. So with that background, we've talked about three new distributors on the call yesterday, and one of them, CTG, is about U.S. centric. They have a long, long history of serving the defense industrial base, and I think they're very trusted in that sense.
They're accessing, you know, lots of different programs of record, lots of different program offices, and I think they're a great resource to help expand the reach that we have through our direct channel to new programs of record that we can then go in and open up opportunities for. On the international side, same thing. You can see that both of those distributors have footprints in aerospace and defense, and so there are, you know, big pushes around the world, and there's lots of investment going on, especially in aerospace, where they're trying to take products, design their own ASICs to serve their interests, and we're, you know, doing that.
We have, I think, a very good export lawyer that's helping us guide through all this, and we know that we're able to export and do it in the correct way to the correct companies in the correct geos, and that actually drove the strategy of what distributors we wanted to bring on board to expand that footprint. Again, leveraging from all the success that we're seeing now in these recent contracts, now is the time to invest more in that from a go-to-market perspective and grow that top of funnel.
Thank you, Brian.
Yeah.
Next question is for Elias. So assuming we maintain a pretty robust, you know, 30% revenue CAGR, and how should we think about margin profile in the medium to longer term, and how should we think about revenue growth rate in relation to OpEx, and how much operating leverage is there in your long-term model?
I think on gross margin, you would look at it from what we said yesterday on the call, which is mid-60s% for 2024. I believe I have stated in many calls in the past that no one should rely on a 73% or 78% gross margin that we had announced because of certain mix and certain reclassification of OpEx and COGS. So going forward, I think our goal internally is to hit mid-60s% and high 60s%, and I think that's the way you kind of look at gross margin. In terms of revenue, how you look at revenue mix versus OpEx, we really are investing in headcount, in people, in engineers. We've hired a local North American director of sales, who we believe was needed for leveraging.
And second, we have hired a bunch of engineers for rad-hard technology in anticipation of many of the projects and agreements that Brian has talked about that are included in the sales funnel that we're excited about, and hopefully we'll land a few of them. So we've already started the investment process. We really don't have much to spend on, such as CapEx, and therefore, I believe OpEx will be flattish to maybe around 3.2 average in 2025. But for now, the remainder of 2024, I believe will be around the $3 million mark or below.
Thank you, Elias. And with that, I see no more questions from the audience, and that's... Well, before we close, is there any remarks you would like to make, Brian or Elias?
None other than I've already said. I think we're really excited about where we are, where the funnel is, where we are with some of these proposals that we're looking to close. And I would say just keep an eye open for announcements that we're making on that between now and sometime in the late fall, that would be contributing to revenue this year and early next year. So thank you for your time. Look forward to seeing you-
Great
... at some point in the future.
Thanks, everyone, and thank you, Brian. Thank you, Elias. That's a wrap.