QuickLogic Corporation (QUIK) Earnings Call Transcript
November 11, 2025
Earnings Call Speaker Segments
Thank you, Van, and I'm glad all of you could join us. On the line with us today are our President and Chief Executive Officer, Brian Faith, as well as our Senior Vice President and Chief Financial Officer, Elias Nader. Before we continue, I want to remind you that some of what QuickLogic says today will be forward-looking in nature, and such statements come with risks and uncertainties. That covers, without being limited to, remarks regarding our profitability and cash flows down the road, what we expect from our business going forward, the schedule, milestones, and payments associated with our government contracts, how the company's ATM program will be utilized, and whether we will be able to complete a successful exit from SNL. Several things might make the real outcome come out differently. These include the possibility that the market takes longer to embrace the company's newly launched offerings, along with how well design prospects get turned into money from customers. Other elements at work: whether we can swap in new revenue to make up for products reaching the end of their life, how much design work customers do and when they do it, whether buyers accept what our customers sell, the chance that fresh orders never translate into later sales, and whether we can roll out and manufacture new items built on cutting-edge wafer technology promptly. Other elements at play include how well we can promote the low-power advantages and aggressive pricing of our new offerings, along with how quickly we get them to market. We also face fierce rivalry from other companies, and we depend on bringing in and keeping talented staff. Shifts in supply or demand matter too, as do broader economic circumstances. Beyond that, there are political developments, trade conflicts between nations, natural calamities, and other disruptions to business operations that might interfere with the supply, shipment, or need for what the company sells — plus adjustments to tax rates and the possibility of owing extra taxes. Should you want a fuller explanation of the risks, uncertainties, and assumptions behind these potential differences, please see the risk factors covered in QuickLogic's latest periodic filings with the SEC. QuickLogic won't take on any duty to revise forward-looking statements or information, since those reflect conditions only as of the dates they were made, even if fresh information or later events come along. During this call, our figures will be presented on a non-GAAP basis. If you'd like a full reconciliation between our GAAP and non-GAAP numbers, along with our other financial statements, that can be found in the earnings release we put out today. Beyond that, a revised financial table now sits on our IR web page, offering both current and historical non-GAAP figures. One more thing to keep in mind: QuickLogic treats its website, Hy blog, corporate Twitter account, Facebook page, and LinkedIn page as channels through which it distributes information regarding its business. That kind of data could count as material, and QuickLogic might rely on these outlets to meet what Regulation FD requires it to disclose. Soon after this earnings call wraps up, the prepared remarks delivered on it will be made available on the company's investor relations page. At this point, I'll hand things over to Brian. The floor is yours, Brian.
Thanks, Alison. Hello, everyone, and I appreciate you joining our Q3 2025 earnings call. We've made tremendous strides since the call we had back in August. On that previous call, I mentioned that we had allocated considerable engineering manpower toward accelerating storefront design wins for Radot, our strategic FPGA, and toward expanding our addressable market to include eFPGA Hard IP designs of very high density targeting leading-edge fabrication nodes. I'm delighted to share that our engineering organization has delivered outstanding results, and those objectives are now coming to fruition. We expect storefront revenue to begin being recognized in early 2026, and we believe it will contribute meaningfully to our total revenue for 2026. The level of interest we're seeing from big players in the defense industrial base — the DIBs — around that SRH test chip we put funding behind has come in well above what I'd expected. Our capacity to serve the highly profitable markets for very high-density discrete FPGAs and ASICs needing large eFPGA blocks has grown a great deal. The new contracts and engagements involve much bigger eFPGA blocks and use advanced fabrication processes. eFPGA and customer designs are now contributing substantially more value. We're also gaining ground across commercial market sectors. Given all this progress, the pace at which new contracts are closing keeps speeding up, and it's gotten to where license revenue could top NRE revenue for the first time this quarter. Our view is that these trends will only pick up speed from here. Before diving into the hard numbers that back up what I've been saying, I'd like to spend a moment giving some context around the revenue guidance Elias is going to walk through in his presentation. Drawing on our backlog along with forecasts our customers have given us, we're aiming for total revenue of $6 million in Q4. Most of the contracts underpinning that outlook are either already booked or have been indicated by customers as expected to be awarded over the next several weeks. That said, one contract worth close to $3 million, for a commercial application aimed at an advanced fabrication node, has been projected by the customer to be awarded toward the end of the quarter. Assuming the award lands at or quite close to the projected timing, we'd be in position to book a substantial chunk of that revenue during Q4, which would let us hit the $6 million target we've set. Our conviction about securing this deal is extremely strong. Even so, keep in mind that it might slip into Q1 2026, which would mean lower overall revenue for Q4 2025. Because of that possibility, Elias is going to give a guidance range that's wider than usual. With that said, let's go through what we've achieved. During early August, we handed over design files to GlobalFoundries so they could manufacture our SRH FPGA test chip on their 12LP process. The purpose behind this test chip was to satisfy what certain big DIBs need — those organizations currently running development programs that fit well with this particular device. Test chips should reach us near the beginning of Q1 2026, and shortly after that, we expect our SRH dev kit to become available for shipping to customers. Funding for this came out of QuickLogic's own pocket, and it is entirely separate from our U.S. government contract. Our decision to invest money and resources into creating this test chip reflects our belief that doing so is critical for securing strategic design wins and accelerating our storefront business model. Since the last time we held an earnings conference call, I've sat down personally with several of the DIBs that partnered with us during development, and I truly can't stress enough how much promise our SRH storefront initiative holds; in earlier meetings, PowerPoint decks were all I had to put in front of them. Now that the test chip is in fabrication, though, you can genuinely feel a higher level of excitement. In fact, we've already got commitments for SRH dev kit orders that should come in by the close of this month. To me, this represents the first concrete move toward the hundreds of millions of dollars in possible storefront business we might capture over the years ahead. Showing off our SRH FPGA test chip matters for reasons far beyond the storefront designs we think it will help us land. Within the defense industrial base, FPGA ranks as the #1 spend category for semiconductor devices, with custom ASICs sitting a close second. Put together, we figure these 2 categories account for about half of the DIB semiconductor TAM. We anticipate that a lot of the new strategic designs needing differing degrees of radiation hardness will rely on either discrete FPGA devices we're able to storefront or eFPGA hard IP we're able to license into new ASIC designs. Launching a standalone SRH FPGA test chip built on the 12LP node demonstrates the broader potential of our eFPGA hard IP for ASIC use cases, and it will meet program needs ranging from radiation-tolerant all the way to strategic rad-hard. I want to highlight 3 especially important items. To start, DIBs already depend on the 12LP manufacturing process from GlobalFoundries for radiation-tolerant and SRH ASICs. Next, once parts become obtainable, government agreements require domestic production for strategic efforts. Given where we stand today, we will be the only provider of strategic rad-hard FPGAs and SRH eFPGA hard IP produced within the United States by an American company. Third, when I talk with engineering managers at big DIBs, they've told me straight out that one big plus is being able to do design work with our Aurora FPGA user tools spanning our SRH discrete FPGAs, our eFPGA hard IP, and ASIC designs. During our last conference call, I noted that this year's low point for revenue recognition on our U.S. government SRH FPGA contract would land in Q3. Thanks to money from the tranche we're in now, that contract's revenue recognition will come back up sharply in Q4. Beyond that, as the following tranche provides funding, we look for quarterly revenue recognition to climb during 2026. On the previous earnings call, I said we expected to win a contract in the mid-7-figure range from a DIB in Q4, aimed at Intel 18A. Regrettably, funding has slipped, which moves that contract to 2026. Our confidence that this contract will be awarded to us is very high, yet right now the customer has only limited visibility regarding when funding will arrive. As we wait for the money on this 7-figure deal, it's important to point out that this strategic customer has already granted us several contracts in 2025. Back in April, we delivered custom eFPGA hard IP for that client's initial Intel 18A test chip. Per the agreement, we expect to receive our share of those test chips during the first quarter of 2026, which will let us carry out internal verification and characterization. Following that, we secured a second Intel 18A test chip deal worth somewhere in the mid-six figures. During Q3, we handed over eFPGA hard IP tailored to a particular customer for this test chip. Outside of the Intel 18A test chip agreements, I mentioned on the last earnings call that this same customer granted us an award covering a feasibility study worth 1 million let, and we're on track to hand that over next week. We anticipate a subsequent order in the coming few weeks connected to that feasibility study, which would let the customer tape out an Intel 18A proof-of-concept device of very high density during the second half of 2026.
The architectural modifications from this feasibility study are applicable to every advanced fabrication node — our definition being 12 nanometers and under. Thanks to those modifications, we're now able to go after the profitable markets needing extremely dense eFPGA blocks within ASIC design, plus extremely dense discrete FPGAs. That broadens our SAM considerably for eFPGA hard IP along with discrete devices, covering our SRH FPGA, chiplets, and additional storefront possibilities. Earlier in the year, we kicked off our digital proof-of-concept chiplet effort as a way to speed up our storefront chiplet push. In-house, we call it POC. Backed by our major strategic partners, we've drawn on our current eFPGA Hard IP plus third-party IP that was readily available to advance quickly while investing minimally. Consistent with the projection I gave during our previous conference call, we finished the first stage of the digital FPGA chiplet POC — the point where eFPGA IP links up with UCIe IP and the interface logic needed for those IPs to talk to each other. That digital simulation of the POC exists today and can be built out further to satisfy various customer needs. Alongside our ecosystem partners, we're in discussions with potential customers across defense, aerospace, industrial, and commercial sectors. Once outside funding is secured, our intention is to proceed into the subsequent stages of the FPGA chiplet POC. Those stages will involve adding more IP — programmable GPIOs, AXi Bus, DSPs, data converters, and interfaces like PCI Express — to satisfy particular customer needs. We feel confident our POC effort will generate storefront revenue during 2026. On October 2, we revealed a fresh $1 million eFPGA hard IP contract tied to a high-performance data center ASIC that will be produced using TSMC's 12-nanometer process. Within that ASIC, our eFPGA Hard IP is set to serve as the main IP in the design. This contract nicely demonstrates our success on multiple points I raised earlier. The demand for bigger eFPGA blocks, eFPGA's rising value contribution within customer designs, securing contracts for designs aimed at advanced fabrication processes, and our expanding traction in commercial market segments — we'll shortly be announcing broader involvement with a DIB focused on cybersecurity for strategic and tactical weapon systems. That DIB builds secure system-on-chip processors that take advantage of the improved security that eFPGA alone can deliver. Operating these processes and hardware is intrinsically more secure compared with software approaches. Since eFPGA sits at the center of the designs, the hardware is modifiable so it can react to emerging threats and refreshed algorithms. We take pride in having been selected as a trusted eFPGA hard IP supplier for these designs. Back in April, we told you about an eFPGA Hard IP deal worth $1.1 million with a fresh customer from the defense industrial base, and that one is going to be built on the GF 12LP process. This use case leans on a big block of our eFPGA Hard IP for critical functions — a pattern we keep noticing, especially when designs are aimed at advanced fabrication nodes. Working alongside this DIB and the customer at the end of its chain, we are taking that large eFPGA core and turning it into a brand-new 7-figure contract that we hope to reveal within the next few weeks. Under this new agreement, test chips will come to us, and those will go into an evaluation kit. That kit is going to work with the usual third-party development environments that DIBs and commercial customers both rely on. Thanks to this, such customers can speed up system-level evaluations and designs capable of using either a storefront edition of the 12LP test chip or our eFPGA Hard IP inside an ASIC. We expect evaluation kits to be ready in late 2026. Having said that, I will hand the call to Elias so he can present the financial data.
Thanks, Brian, and good afternoon to all of you. Our total revenue for the third quarter came in at $2 million, which matched the midpoint of the guidance we provided. Compared with Q3 2024, total revenue declined 52.5%, and versus Q2 2025 it fell 45%. When rounded to the closest $100,000, revenue from new products during Q3 reached $1 million, while mature product revenue came to $1.1 million. New product revenue dropped 73.1% relative to Q3 2024 and declined 67.3% from Q2 2025. Mature product revenue rose compared with $0.7 million in the third quarter of 2024 and increased from $0.8 million in the second quarter of 2025. Our non-GAAP gross margin for Q3 stood at negative 11.9%. That compares against a non-GAAP gross margin of 65.3% in Q3 2024 and 31% in Q2 2025. The main factors behind the weaker Q3 gross profit margin are unfavorable absorption of fixed costs stemming from lower revenue, along with $300,000 in R&D costs that were allocated to COGS. Non-GAAP operating expenses for Q3 totaled roughly $2.9 million. That figure came in about $300,000 under the midpoint of our outlook, thanks to the COGS allocation I just referenced. This compares with non-GAAP operating expenses of $3.3 million in the third quarter of 2024 and $2.5 million in the second quarter of 2025. Our non-GAAP net loss was $3.2 million, or $0.19 per diluted share. That compares with a non-GAAP net loss of $0.9 million, or $0.06 per diluted share, in Q3 2024, and a non-GAAP net loss of $1.5 million, or $0.09 per diluted share, in the second quarter of fiscal 2025. The gap between our GAAP and non-GAAP results stems from noncash stock-based compensation expenses, impairment charges, and restructuring costs. Stock-based compensation for Q3 was $0.8 million. Stock-based compensation totaled $1.2 million in Q3 2024 and $0.8 million in Q2 2025. Impairment charges amounted to $0.3 million in Q2 2025. During the third quarter, 3 customers each represented 10% or more of total revenue. As Q3 closed, total cash stood at $17.3 million, which includes utilization of $15 million from our $20 million credit facility. That compares with $19.2 million, including usage of $15 million from our $20 million credit facility, when Q2 2025 closed. Net of roughly $200,000 raised through our ATM in July, cash usage during Q3 came to approximately $1.9 million; this was mainly driven by tape-outs and wafer costs tied to our internally financed SRH FPGA test chip. Beyond those one-time costs, there were also expenditures related to revenue contracts and repayments for finance tooling and equipment. Now let's turn to our guidance and outlook for the fiscal fourth quarter, which ends on December 28, 2025. Given our backlog and customer forecast, we are targeting total revenue of $6 million for Q4. A large number of the agreements underpinning this guidance are either signed already or customers have told us they expect to award them over the next several weeks. That said, one customer has indicated that a contract worth close to $3 million, for a commercial use, will be awarded late in the quarter. Should that award land on the forecasted date or very close to it, we would be in a position to record a significant share of the associated revenue in Q4. Doing so means we hit our $6 million target. Our confidence in securing this contract is quite high, though keep in mind it might slip into Q1 2026, which would leave Q4 revenue at $3.5 million. Because of that, we are guiding total Q4 revenue to a range of $3.5 million to $6 million. At $3.5 million, we would expect the total to break down as $2.5 million from new products and $1 million from mature products. At $6 million, we would expect $5 million from new products. Given the anticipated Q4 revenue mix, non-GAAP gross margin for the fourth quarter should come in around 45% at $3.5 million of revenue and 68% at $6 million of revenue. When we are at the bottom of the range, the main driver of the weaker gross profit margin is less favorable absorption of fixed costs. Considering the span of our Q4 outlook, our full-year 2025 non-GAAP gross profit margin should be 38%, plus or minus 5%. We expect Q4 non-GAAP operating expenses of roughly $3 million, plus or minus 5%. On that basis, we model full-year 2025 non-GAAP OpEx at approximately $11.3 million. Keep in mind that, because of the way our industry works, we may from time to time need to put certain expenses into COGS rather than OpEx, or capitalize certain costs. Those classifications involve labor and tooling tied to our IP contracts with customers, and they can create variability in our quarterly gross margins and operating results, though this typically evens out at the operating line. After interest and other income, at the bottom of the revenue range we project a Q4 non-GAAP net loss of about $1.9 million, or $0.11 per share. At the top of our revenue range, we project a non-GAAP net profit of roughly $600,000, or $0.04 per share. The biggest gap between our GAAP and non-GAAP results comes from noncash stock-based compensation expenses. For Q4, we expect that compensation to be about $800,000, which matches Q3 2025 and is slightly lower than Q4 2024. As a reminder, stock-based compensation will move around during the year and can differ quarter to quarter depending on when grants occur. Even at the bottom of our revenue guidance range, we expect positive cash flow in Q4. That said, the timing of payments under our U.S. government contract could work against this outlook. Since we brought in roughly $2 million through our current ATM during October, we're fully ready to handle any late payments tied to the U.S. government contract. Thanks. Having said that, I'll hand things over to Brian now so he can give his closing remarks.
I appreciate that, Elias. Over the past several months, we've recorded substantial advancement, and we're using that advancement to generate concrete outcomes. I discussed those outcomes previously. So for the next few minutes, I'd like to walk you through the industry shifts behind these outcomes. Once you grasp that, I believe you'll see why interest in FPGA technology keeps growing, and why an expanding number of businesses are placing bigger eFPGA blocks at the center of fresh ASIC architectures. Across commerce and design alike, the dominant movement is toward smart systems. Intelligence in smart systems comes from algorithms. When implemented in hardware rather than software, algorithms can run far quicker while drawing significantly less power. Compared with software, hardware-based processing also offers built-in resilience to cyber attacks. The difficulty is that across a product's life cycle, algorithms require refreshing; consequently, hardware needs programmability to keep pace with shifting algorithms. That's what has created demand for bigger eFPGA blocks at the center of ASIC designs, in contrast to earlier scenarios where modest eFPGA blocks typically served as programmable connectivity bridges. As a result, both the necessity for eFPGA and its value proposition are on the rise. Typically, intricate smart system designs aim at cutting-edge fabrication nodes. For ASICs, that translates into steeper fixed costs and extended design cycles. To offset those steeper fixed costs in a favorable way, ASIC designs need to achieve longer life cycles than was previously the case. Designs incorporating eFPGA can adjust to shifting algorithms, changing functional needs, and outside developments that aren't apparent while the design cycle is underway. Such adaptability extends the life cycle of ASIC designs and gives program managers confidence to shift ASICs into production sooner and with reduced risk. That in turn shortens design cycles and brings down development costs. Finally — and this matters greatly — numerous programs now in development need to satisfy demanding environmental criteria, spanning from radiation-tolerant all the way to strategic rad-hard. Our internally funded effort on an SRH FPGA test chip is built to cover that entire spectrum of criteria and speed up our capacity to chase design wins. Because we're relying on the identical onshore 12LP fabrication process that DIBs have employed for SRH ASICs, we're maximizing our odds of securing discrete FPGA designs we can storefront, along with eFPGA Hard IP contracts that customers can build into ASIC designs. Strengthening our standing further is the reality that customers can run designs using our Aurora user tools for each of these. That strategic DIBs have embraced this QuickLogic investment so warmly is confirmed by the commitment we hold for SRH dev kit orders, which we expect to receive by the close of this month. As we get ready to hand things over to Q&A, I'd like to pause briefly to acknowledge Veterans Day and say thank you from the bottom of my heart to everyone who has worn our country's uniform; this occasion matters to me personally, since a number of people in my own family have served, and I hold great admiration for what veterans and their loved ones have given up. That's a value we hold dear here at QuickLogic, particularly given that the technologies we build play a part in keeping our country defended and secure. Operator, at this point I'd like to invite any questions from the line.
This is Neil Young filling in for Quinn Bolton; so the thing I was hoping to ask, Neil, as I mentioned, standing in for Quinn. My apologies for that. So how is the government shutdown affecting your business? From what was said in the prepared remarks, it sounds like some projects have slipped. Have any cancellations come up? And then, with the shutdown still going on, even though it's supposedly meant to wrap up shortly, why are you confident the USG strategic radiation Arden FPGA program bounces back in 4Q? And then I've got one more after that.
Yeah. I'd say, first off, let's take a step back. For decades, programmable logic has played a major role in the defense industrial base, and that isn't going to shift. It shows up throughout roughly 75% of defense systems. And like I said before, a really big share of what the DoD spends overall on semiconductors; so that demand isn't disappearing. The real issue is, once you dig into the details of these programs, will the money be available given the budgets and so forth? So from the programs we currently have under contract, I don't see any slippage there. Elias brought up during his remarks the cash usage for the quarter, or rather, the net cash gain in the quarter; we tapped the ATM in October as a precaution in case something like this arose regarding funding. So should funding slip, we're fine with that. Should there be no slippage, we'll end up with solid positive cash flow for the quarter. Beyond that, looking at other contracts in the pipeline, I mean, you can find all this publicly that many of the new RFIs or RFSs or RFPs the government was putting out for different development programs, a number of those were in fact halted; and I believe that's mostly due to some of the workers who were pushing that being furloughed. Again, I don't expect those to vanish for good. It's more that once the government gets funded and folks return from furlough; those will go full steam ahead; so a delay might hit some of those new programs, but not the ones we're fully executing on today. I simply don't see that changing since this isn't an experimental technology. These are actual programs of record that require this today, and continuing forward on that. Does that cover your first question?
Right, and now for my second thing I'd like to raise; it seems as though storefront revenue is expected to take a substantial jump in 2026. If you can, I was hoping you might be able to give us a sense of the size of the range of storefront revenue that you believe is achievable? If that's not possible, then maybe you could share what might push results above your own internal projections, or conversely, what might push them below those projections.
Certainly. I'll begin with the what, and afterward I'll get to the why. On the what side, I mean, I would describe significant for us as roughly 10% of total revenue. And without providing the precise figure, since we haven't released numbers for 2026 as of yet, we believe the storefront revenue tied to these developments we've been discussing will be meaningful, which is to say it will fall within that 10% range. And yes, I do believe next year's revenue will be considerably above this year's total revenue. Now, as for why do I feel that way? I think if you look back at what I said at the start regarding the strategic router initiative, I couldn't begin to count the meetings I've had over the last quarter since the previous conference call, in person with these DIBs, who see what we're doing. They appreciate the fact that we've completed this tape-out that we mentioned. And even right now, plenty of calls and e-mails are coming in asking when they can get their hands on this. So once you begin to see people advocating for the technology, and you're aware of the projects under development, public projects. The strategic defense system is undergoing a major modernization. That's all public knowledge. And then if you add into that, this idea of hypersonics and golden dome, many of these programs are going to require some level from strategic radar down to radiation tolerant; and the part we've got in the fab now is built to address those needs. So as we get it out, we start moving to these orders for dev kits, we start getting those out, hopefully, by the end of Q1, I think we're going to be in a real prime position to monetize that and start turning talk that I've had for 2.5 years into actual revenue and bottom line contribution. But it's not just one entity here. We're talking about all the major DIBs that we've been talking to. I think there's good demand for that; so that's why we think it's going to be meaningful for next year. Does that answer your question?
The audio quality on my end is quite bad for my response. I hope you're able to hear me. I apologize if you cannot. Alright. Here we go. I suppose at least one of us is able to. There's a great deal of detail on this call, and some genuinely fascinating things are happening. Allow me to pose a broad, high-level question regarding your new initiative involving the GF12LP process. How should we think about the opportunity for FPGAs as opposed to ASICs that would incorporate your hard IP? And are the timing dynamics for each of these notably different from one another?
To begin, I would say regarding 12, that process sees very widespread use across the defense industrial base. The reason behind that, I believe, is that it was the most advanced node that GlobalFoundries offered, and GlobalFoundries is a U.S.-owned and U.S.-operated company. So if you wanted something fabricated domestically by an American firm, that was the most advanced option available. Global has subsequently introduced 12LPus, an upgraded iteration of 12LP; however, when you consider what goes into producing an ASIC or an SoC, you require a great deal of IP to be on hand, plus extensive test data characterization across all of that IP before you can feel confident proceeding with your ASIC. And when it comes to the defense sector, that sector is highly risk-averse, which is appropriate given that they are the ones designing the system. So there's an abundance of IP on 12LP that exists, it's available now, it's recognized, it's comprehended, with characterization data. And the government, again, all of this is publicly fungible. The government has been assisting organizations in doing ASICs on 12LP. Plenty of IP can be accessed. There are government-funded multi-project wafers and similar programs aimed at promoting development on that node. So from that perspective, I believe you will continue to see a lot of 12LP. It has appeared previously, and it will appear going forward. So the question then becomes, for us, alright, our IP is on 12LP. We can now construct devices from it, or alternatively we could build or license it to others doing their own ASICs. And I believe we've already discussed IP licenses we hold on ASICs, and you've heard the timing on that; so people will begin taping those out and moving to production, ideally, within the next few years; so there's a near-term licensing opportunity. There's also a back-end royalty opportunity for us there. And we certainly intend to monetize that to the tune of several million dollars annually. On the device side, things get more interesting because we've clearly taken our commercial 12LP and produced a rad-hard version of it. The purpose there is to pursue this strategic rad-hard FPGA and to remove that element. If you jump ahead to the point where we're able to produce the actual product die for that, once it's available, that will represent a major step function rise in the revenue potential for us specifically, since devices of that kind will always carry a far higher ASP than a royalty contribution would; therefore, in my view, 12LP is extremely important to us. And it's a land and expand strategy at this stage. Our aim is to license it to the largest number of parties possible. We want to secure this strategic FPGA for revenue. And I believe that's the foundation of what could amount to hundreds of millions of dollars in revenue. I'm not sure whether I've fully answered your question. If I haven't, simply let me know.
For the most part, yes. I'm simply attempting to orbit this for a moment from a really broad vantage point. Let me pose yet another fairly big-picture question, Brian, and that is: when you set the chance you've just taken on with GFS 12LP against what you've been pursuing in rad-hard alongside the other foundries you've made announcements with, how do those stack up? I suppose, looking at things in aggregate, the time frame is yours to choose; but what's your sense of how big each of these opportunities is relative to the others?
When you say the other founders, do you mean SkyWater and Honeywell, or someone different?
Yes, those are the correct items.
All right. So without diving into programmatic specifics, I'd say the 12LP chance is bigger for us since it includes the strategic router FPGA. IP licensing is also a possibility there. And one advantage—which you're already aware of—is that with smaller process technology, transistors become denser, capability increases, you can pack more onto a die, and that raises the value; we laid that out regarding where our eFPGA architecture is headed, but the same holds at 12-nanometer and 12LP. The greater the number of transistors and functions we can place on that die, the greater the worth of the part becomes. And again, I think what's compelling about 12LP here is that we can run a great deal of capability on our FPGA, 12LP, and perhaps someone no longer even needs to pursue an ASIC for 12LP. That's enormous. If we can begin assisting people in meeting a mission's requirements without having to go build a custom ASIC, you're discussing saving a customer or the government literally tens of millions of dollars across years of development expense and schedule; and that, I believe, is the true payoff of putting our FPGA on 12-nanometer, since it's strategic right at the core and so capable for the node. Now, when you mention those other foundries, those are older process geometries that they've brought up; so what you can accomplish on the die will differ. Capability-wise, what you can do will differ—not worse, simply different. But I think the larger revenue pool for us will be what we can achieve at 12-nanometer on these for now; and I'd rather not say more about that, because going further would reveal too much programmatic detail.
Yeah, I follow that. At a broad level, Brian, this is really useful for thinking things through. You said you anticipate orders for the new dev kits — I believe by the close of this month — with shipments landing at some point next year. Could you tell us roughly what quantity of dev kits and what number of customers you're expecting to send those to? And after customers have them in hand, what does the design cycle look like — what are their next steps?
Look, I won't provide figures — that likely doesn't shock you. Still, the volume will be large enough to matter as a meaningful revenue figure, so it won't simply be a rounding item on the income statement. Elias discussed what we laid out in Q3 for that. We deliberately purchased sufficient die so we can supply enough to those customers eager to try these out — whether as a dev kit or simply as bare devices on their own boards. From an evaluation standpoint, here's how it goes: once again, this is an extremely careful, risk-averse community we're discussing. They will insist on running their own tests on these items, and typically that requires a couple of quarters to carry out all your design exercising plus the various environmental tests that must be performed on those devices. You've likely seen the TRL levels — technology readiness level. Our aim is to reach customers as fast as we can, and TRL5 is the point where they can genuinely claim they've taken the part and put it through rigorous testing that reflects the environment where it will operate. So we're hoping to help our customers get past that stage at some point in the middle of next year, and from there begin capturing actual programs of record with us and advancing, hopefully alongside them, into fairly late design stages. And once again, that's exactly why timing matters so much. That's why we took the leap of faith to design it ourselves and finance our own tape-out, aware that MPWs rarely come around, and we wanted to ensure we were on one that still left us enough time for the part to arrive, be verified, and reach the chip's hands so they can begin experimenting with it in their own labs rather than merely relying on our data.
Final thing from me, and then I'll step aside. There's a whole lot going on with everything you've got in motion, and that's really encouraging to watch; at the same time, QuickLogic is, clearly, a pretty small outfit. With such a wide array of customers heading your direction before long, it feels like some extra backing may be required. When we look at next year, how should we frame the spending we ought to expect, whether that shows up as OpEx or as items charged into COGS? And if everything breaks right — you draw plenty of interest and plenty of engagement around those dev kits you're shipping out — how should we think about where this might end up?
Let me take that first, and Elias can add his thoughts on top. From where engineering sits, and from the go-to-market side too, we've clearly pinpointed a handful of essential hires. A few of those are already visible on our site right now, and the whole point is securing the proper people so the devices reach users' hands as fast as we can manage. That's all there on our site — engineering, field application engineering, that sort of thing. Once we shift from the test chip over to the real product chip, costs will grow. Other items will need funding as well; and my sense is we've got a solid view of exactly what those will amount to. It won't be anything wild for next year. My expectation is that we stay very conscious of our financial position as a business, and that it ties back to bringing these customers aboard with Test chips, so that whatever we invest flows from understanding what the customer is after, understanding the problem our solution addresses, and in certain cases, possibly even securing money from customers who co-invest alongside us — giving them skin in the game while it reduces the upfront cost for QuickLogic to bring this to market.
Right, that's accurate. Also, Richard, let me jump in here — take, for instance, the 3 new people we're trying to bring on board, every one of them engineers. Because of that, on the OpEx side, headcount is certainly easing off. So even with every one of those additions Brian is talking about, my expectation is we'd be sitting at roughly $3.5 million in OpEx each quarter, likely in the next year, though beginning around Q2. As things stand right now, I'd say we're fine at somewhere below $3.
About your Q4 guidance, I want to get this straight. What you're putting forward is a choice between two outcomes — either we land at $3.5 million, give or take, or we land at $6 million, give or take? Is that the idea?
That's right, and the reason is a timing problem. Take the scenario where the order arrives and pushes us up to $6 million — it would show up toward the end of the quarter, and we might manage to book some pieces of that revenue. But suppose it arrives and we can't ship within that quarter — say it lands on closing day or the following day — then it's unquestionably Q1. So really, Rick, what we're dealing with is a matter of timing. That's the reason we took such care to lay out the gap between $3.5 million in revenue and $6 million in revenue, and honestly, that gap comes down to a single order. Given that, timing is the whole story. Which makes it very hard, right now, to respond to someone who asks, fine, could you book all 100% of it? Plainly the answer is no, should it arrive during Q4. That's exactly why Brian and I came to an agreement: should that happen — and we do expect that order to land, at minimum within Q4, let's simply hope it does — then at least we retain the chance to surpass the top end of the range.
For 2025, what are you projecting your share count to be?
So, 17,090,252 shares. That's every share currently outstanding.
All right. Last question. So your closing share count would be 17. Back in August, when you talked about how much revenue you expected to lose in 2025, you used the word modest. Yet your Q4 guidance now points to annual revenue falling 20% to 30% versus 2024. What shifted since August to bring about what I'd call a substantial double-digit percentage drop in revenue?
Rick, that's the difficulty that comes with big IP contract values, since we're discussing ASPs in the $3 million range for these. So during the call, we pointed out that one of them is clearly in 2026. So it shifts out of this year and into the next one. And then there are some other, smaller ones that add to that. But once again, when your IP contracts run $3 million, if they fail to land in the year, the fiscal year, you're going to see a large percentage-wise swing from the revenue levels we're sitting at today. Once that turns into more of the usual pattern and we land more of these higher-value contracts like the ones we're discussing now, that begins to even out some of that lumpiness. But at the stage we're in right now, it's nearly impossible to avoid a material hit to that percentage if something slides out.
Absolutely. I appreciate everyone being here with us today. With any luck, we'll see some of you at an event we have coming up — that includes the Craig-Hallum Alpha Select 1-on-1 Conference, taking place in New York on November 18, the Semiconductor-focused Annual New York Summit, likewise in New York on December 16, or the Annual Needham Growth Conference, which happens in early January 2026. Thanks so much, and enjoy the rest of your day.
Thanks so much, everyone, for joining in. That brings this call to a close. You can now hang up, and I hope the rest of your day goes great.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete QuickLogic Corporation transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to QuickLogic Corporation earnings transcripts and 132,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.