Home / Transcripts / QuickLogic Corporation (QUIK) · March 3, 2026

QuickLogic Corporation (QUIK) Earnings Call Transcript

March 3, 2026

NASDAQ US Information Technology Semiconductors and Semiconductor Equipment earnings 61 min

Earnings Call Speaker Segments

Alison Ziegler executive
#1

Thanks, operator, and I appreciate everyone being with us. Joining us on the call are Brian Faith, our President and Chief Executive Officer, along with Elias Nader, our Senior Vice President and Chief Financial Officer. Before we continue, please note that certain remarks made by QuickLogic during this call constitute forward-looking statements subject to risks and uncertainties, which include, though are not limited to, statements about our future profitability and cash flows, expectations concerning our future business and anticipated revenue growth, plus statements about the timing, milestones, and payments tied to our government contracts. A range of factors could cause actual results to differ, among them delays in market acceptance of the company's new products, the capacity to turn design opportunities into customer revenue, our capability to replace revenue from end-of-life products, the level and timing of customer design activity, market acceptance of our customers' products, the risk that new orders might not generate future revenues, our ability to bring to market and manufacture new products built on advanced wafer technology in a timely manner, our ability to properly market the low power, competitive pricing, and short time to market of our new products, fierce competition from rivals, our capacity to recruit and keep qualified personnel, shifts in product demand or supply, broader economic conditions, political events, international trade disputes, natural disasters, and other business interruptions that might disrupt supply or delivery of, or demand for, the company's products, as well as changes in tax rates and exposure to additional tax liabilities; for a more thorough discussion of the risks, uncertainties, and assumptions that could lead to those differences, please see the risk factors covered in QuickLogic's most recently filed periodic reports with the SEC. QuickLogic undertakes no obligation to update any forward-looking statements or information, which speak as of the respective dates of any new information or future events. On today's call, we will be reporting non-GAAP financial measures. I'd like to point you to the earnings release we issued today for a detailed reconciliation of our GAAP to non-GAAP results and our other financial statements. We've also posted an updated financial table on our IR webpage containing current and historical non-GAAP data. Please be aware that QuickLogic uses its website, the company blog, corporate X account, Facebook page, and LinkedIn page as channels for distributing information about its business. Such information may be considered material information. QuickLogic may use these channels to satisfy its disclosure obligations under Regulation FD. A copy of the prepared remarks made on today's call will be available on QuickLogic's IR webpage shortly after today's earnings call ends. I'd now like to hand the call over to Brian. Go ahead, Brian.

Brian Faith executive
#2

Thanks, Alison. Good afternoon, all. I appreciate everyone being here for our Q4 2025 earnings call. During the year, some contracts slipped, and that pushed 2025 revenue well below what we had anticipated. Even so, we hit a number of concrete milestones that position us nicely heading into 2026 and afterward. Evidence of that is our projection of close to 50% sequential top-line expansion in Q1. We're in late-stage talks on sizable deals for eFPGA hard IP cores with extremely high density, plus our storefront business model is picking up speed, which we expect will begin contributing meaningfully to revenue in 2026. Let me spend a few minutes bringing you up to date on these items and others; on February 18, our press release announced that QuickLogic had secured a $13 million tranche tied to our continuing U.S. government contract, which started back in 2022. Recognition of revenue from this tranche starts in Q1; consistent with what I said on our prior earnings call, this tranche supports higher quarterly revenue recognition compared with 2025; alongside that U.S. government contract, QuickLogic used its own funds to develop an SRH FPGA test chip. In August of last year, we handed design files to GlobalFoundries so it could fabricate our SRH FPGA test chip on its 12LP process. The chip was built to satisfy the particular needs of certain large DIBs whose programs currently in development are strong fits for this device. This outlay puts us in an excellent position as the sole provider today of a U.S.-made FPGA covering the entire range of radiation hardness needs. We got samples of our SRH FPGA test chip earlier in Q1, and a January 14th press release announced orders for our SRH FPGA dev kit, which lets DIBs assess the test chips. To me, that's a robust demand indicator and our first concrete move toward what I expect will be hundreds of millions of dollars in possible storefront business for our discrete SRH FPGA over the years ahead. Past the discrete SRH FPGA market, we're using this test chip to reach a far wider audience. Besides applications needing strategic radiation hardness, which are the likeliest to be designed with our storefront discrete SRH FPGA, plenty of other applications have less demanding radiation specs and might rather embed our SRH eFPGA hard IP into ASICs. DIBs already employ GlobalFoundries' 12LP fabrication process for differing degrees of radiation hardness in ASIC designs. By showing off our SRH FPGA test chip, which is likewise made on 12LP, we're putting QuickLogic in position to serve discrete SRH FPGA needs and to give DIBs the assurance required to embed our SRH eFPGA hard IP in upcoming ASIC designs. In certain instances, those DIBs might also choose to use our storefront services for their ASIC designs. To put it briefly, by building on what we achieved in 2025, we feel we're in a strong position to take on both discrete and embedded FPGA projects spanning the entire range of radiation hardness needs. Thanks to the architectural improvements we rolled out last year, which can be carried forward to 12LP, our SAM in these markets has grown considerably to encompass the profitable uses for very high-density, discrete, and embedded FPGA. On our previous conference call, I mentioned that a mid-seven-figure eFPGA hard IP contract involving Intel 18A had slipped to 2026 because government funding was delayed. From what we've discussed with this DIB, we stay very confident that once funding comes through, we'll win this contract. Although when the funding will arrive is still unclear, our talks with this DIB have grown to cover the possibility of QuickLogic delivering storefront services for the customer-designed ASIC that will incorporate our eFPGA hard IP. We anticipate finding out more about the possible storefront expansion and the award's timing over the next few months. Amid this funding delay and the conversations about broadening our role, we've collaborated closely with this DIB on numerous projects. As a result, we've secured three smaller Intel 18A contracts worth well over $1 million combined, and a fourth is in the works that would push the total to almost $2 million. Intel 18A test chips were the subject of the first two contracts. We provided IP for both during 2025 and look forward to getting an allotment of test chips for our internal evaluation in the next quarter. The third contract covered a 1 million LUT feasibility study, which we finished in Q4. A fourth contract, expected to be awarded before this quarter ends, builds on the architectural improvements created through the 1 million LUT study; for this contract, we'll supply hard IP for a very large Intel 18A eFPGA core that the customer intends to integrate into its ASIC, with tape-out targeted for the second half of 2026. The architectural improvements we created for the 1 million LUT study can be applied across every advanced fabrication node, which we define as 12 nanometers and smaller; these improvements cut power consumption, boost performance, and shrink the silicon area needed for a given size block of our core FPGA technology. In industry parlance, the improvements deliver a material gain in our PPA. Now that these architectural improvements are available, we can serve the profitable markets needing very high-density eFPGA cores in ASIC designs and very high-density discrete FPGAs. This substantially widens our SAM for eFPGA hard IP and discrete devices, such as our SRH FPGA, chiplets, and other storefront possibilities. Beyond these DIBs contracts, we're collaborating closely with a large commercial customer on a new Intel 18A contract worth several million dollars. Initially, we thought this contract would be granted late in Q4. But the customer opted to increase the eFPGA core size within their ASIC so they could get more programmable flexibility. That's a positive development for QuickLogic, though it pushed the award date back. Right now, our projection is that this contract gets awarded sometime in Q2. Back on our November 2025 conference call, I mentioned we'd shortly be revealing a broadening of our work with a DIB focused on cybersecurity for strategic and tactical weapon systems. Then on December 8, a press release went out saying Idaho Scientific picked our eFPGA hard IP for forward-leaning, hardware-based cryptographic solutions aimed at mobile, IoT, infrastructure, and defense systems. Idaho Scientific brings a deep track record of using FPGA technology to create strong security systems that adjust fast to shifting external threats, avoiding the weaknesses built into software-based approaches. When our eFPGA hard IP gets integrated into its secure system-on-chip processors, it can boost cryptographic security even more and reach new markets far faster, with reduced risks and reduced costs. In April of last year, we disclosed an eFPGA hard IP contract with a fresh defense industrial base customer worth $1.1 million, set to be fabricated on the GF 12LP process. This application makes use of a sizable block of our eFPGA hard IP for critical functions — a pattern we're noticing in designs aimed at advanced fabrication nodes.

Brian Faith executive
#3

Thanks to collaboration with this DIB and the end customer it serves, we're taking that big eFPGA core and applying it toward a fresh contract in the seven-figure range, which we hope to make public before this year is out. That said, because several parties are involved in this deal, wrapping it up has taken more time than we had anticipated. Looking at our present projections, we expect the award to come sometime in the current quarter. Under this new agreement, we'll receive test chips, and those will go into an evaluation kit. That kit, presently planned for late 2026, is going to work with the common third-party development environments that both DIBs and commercial customers rely on. What that does is let those customers speed up system-level evaluations and designs capable of using either a storefront version of the discrete FPGA or our eFPGA hard IP inside an ASIC. Alongside all of that, we're looking into whether we can use the FPGA as a chiplet co-packaged alongside a microcontroller from one of our partners; interest in this idea is already coming in from a number of those partners; during 2025 we finished the first phase of our digital proof of concept chiplet program, a strategy meant to speed up our storefront chiplet initiative. Inside the company, we call it POC. Backed by our large strategic partners, we took our existing eFPGA hard IP along with readily available third-party IP and pushed this program ahead quickly while investing very little. Given the continuing discussions around communications and protocol layers for chiplet interfaces, this POC, combined with our decades of experience in FPGA bridging, puts us in a strong position as a possible answer for advancing chiplet designs to meet what looks like considerable pent-up demand. At the recent Chiplet Summit, we were asked to give a paper on our POC, and we'll also present at Intel Foundry's Partners talk at the coming GOMAC, alongside Cadence and Trusted Semiconductor Solutions. What we took away from our Chiplet Summit presentation reinforces our confidence that chiplets will gain traction in 2026. Interoperability gaps are the main obstacles right now. In our view, a storefront FPGA chiplet is the sensible answer for a programmable bridge. Earlier in the year, Epson granted us permission to share its case study, which backs up our assertions that employing FPGA technology to run algorithms cuts power usage while keeping programmability, compared with processing and software. We put those results out in a blog post dated January 13th. Epson's SoC was initially architected so that workloads ran entirely in software. As the need for additional features and real-time responsiveness increased, power usage turned into a constraint. Epson's engineering team saw that shifting compute-intensive functions into dedicated hardware might yield substantial efficiency improvements, yet that hardware approach would have to adapt when algorithms change. That left only one practical answer: an eFPGA core built into the SoC. Leveraging our proprietary Australis eFPGA IP generator, we managed to deliver a customized hard IP core rapidly, one designed specifically for the SoC application and aimed at TSMC's N12e fabrication technology. One more hurdle we faced: this was the very first eFPGA hard IP we had done for N12e. All the way from handing off the design through validating it on silicon, the IP went into Epson's SoC smoothly, with no respins and no late-stage design changes required. Epson could boot, configure, and validate the eFPGA subsystem right away, which sped up its timeline and lowered risk. Once final testing wrapped up, Epson verified that the design cut total power consumption by 50%. For systems that run on batteries, that is an enormous difference. Because this design worked out so well, we feel we are positioned strongly for upcoming opportunities both with Epson and with other firms that have comparable needs. As you surely saw in our 8-K, we recorded a sizable impairment charge related to SensiML. That stems from the usual accounting rule requiring an asset held for sale for a year or longer to have its value impaired. Over the past year, we talked with microcontroller companies about divesting SensiML, and with one of them the talks reached due diligence, though they ended with no agreement. Right now we are talking with a large company where SensiML software could be quite valuable for new AI and drone projects; we cannot give any assurance that this discussion, or others, will lead to a transaction. With that, I'll hand the call to Elias so he can present the financial data.

Elias Nader executive
#4

Thanks, Brian, and good afternoon to all of you. Our total revenue for the fourth quarter came to $3.7 million. That represents a 35% decline versus Q4 2024, while it rose 84% compared with Q3 2025. Revenue from new products during Q4 reached $2.8 million, while mature product revenue amounted to $0.9 million. New product revenue fell 39% relative to Q4 2024 and climbed 199% against Q3 2025. Mature product revenue declined from $1 million in the fourth quarter of 2024 and from $1.1 million in the third quarter of 2025. Our non-GAAP gross margin for Q4 stood at 20.8%. A couple of items pushed the non-GAAP gross profit margin below what I had projected: $473,000 related to inventory reserves, plus $135,000 of contracted professional services costs booked to COGS that weren't on our radar when we held the last conference call. Beyond those, the rest of the variance mostly reflects professional services contributing more than we expected, compared with IP and mature product revenue. Q4 non-GAAP operating expenses came to about $3.5 million, which was $500,000 above the midpoint of our outlook, driven by certain executive incentives booked during the quarter. For comparison, non-GAAP operating expenses were $2.9 million in Q4 of 2024 and $2.9 million in Q3 of 2025. On a non-GAAP basis, our net loss was $2.9 million, which works out to $0.17 per share. For comparison, in Q4 2024 we posted non-GAAP net income of $0.6 million, or $0.04 per diluted share, while the third quarter of fiscal 2025 brought a non-GAAP net loss of $3.2 million, or $0.19 per share. The difference between our GAAP and non-GAAP figures stems from non-cash stock-based compensation expenses, plus the non-cash impairment charge tied to SensiML that Brian discussed. Stock-based compensation in Q4 totaled $700,000, compared with $900,000 in Q4 2024 and $800,000 in Q3 2025. Three customers each accounted for at least 10% of total revenue during the fourth quarter. When Q4 ended, our total cash position was $18.8 million, including $15 million drawn from our credit facility. At the end of Q3 2025, the figure stood at $17.3 million, $15 million of which came from our credit facility. The $1.5 million gain in net cash reflects $3.2 million brought in through our ATM in Q4. Now for our fiscal first quarter guidance and outlook, with that quarter wrapping up on March 29, 2026. Based on backlog and what customers are forecasting, we're guiding to total revenue of $5.5 million ±10% for Q1. Within that total, we expect $4.5 million to come from new product revenue and $1 million from mature product revenue. For the full year, our expectation is mature product revenue of roughly $4 million. Based on the revenue mix we expect for Q1, non-GAAP gross margin in that quarter ought to land around 45%, plus or minus 5%. When we look at the whole of 2026, our model assumes a non-GAAP gross profit margin of 57%. As we see it, a number of factors are going to weigh on that non-GAAP gross profit margin across the opening two quarters of 2026. Our modeling for that stretch assumes services revenue accounts for a large share of total revenue. Supporting those services means paying for leased software tools, and on top of what our own teams can handle, we'll bring in outside engineering services as well. At present, a substantial portion of those expenses sits in COGS under our model. At this point, we can't say precisely what portion of those expenses will be capitalized, but we do anticipate that some will be. During the first half, we'll take on sizable costs related to big contracts whose recognition comes late in the quarter, and the revenue that offsets them won't show up until the quarter after that. Our models treat these dynamics as a drag on non-GAAP gross profit margin across Q1 and Q2. For fiscal 2026, the significant costs in our model include three multi-project wafer, or MPW, tape-outs. Every one of those three tape-outs relates to products we plan to sell through our storefront program. Customer contracts will fully absorb the costs tied to two of those tape-outs. A pair of these agreements is in place already, while a third is being hammered out and is quite far along. Our expectation is that at least some portion of what the third tape-out costs will be offset through contracts. Should we lock in those contracts before the tape-outs happen, that would represent additional upside beyond what we've baked into the full-year model. Keep in mind that because of how this industry works, there are times when we have to decide whether certain expenses belong under COGS or OPEX, or whether to capitalize particular costs; those decisions involve labor and tooling tied to IP products — IP contracts, sorry. As a result, you may see swings in gross margins and operating results from quarter to quarter, though these typically even out when you look at the operating line. Taking that into account, we anticipate non-GAAP operating expenses for Q1 to come in at roughly $3.2 million, give or take 5%. For the entire year, we're looking at non-GAAP operating expenses of about $13.5 million; that projected 14% increase in non-GAAP OpEx compared with 2025 is meant to back the revenue growth we expect in 2026. Once interest and other income are factored in, we're projecting a Q1 net loss of around $800,000, which works out to roughly $0.04 per share. The primary gap between our GAAP and non-GAAP figures stems from non-cash stock-based compensation expenses; for Q1, we anticipate that compensation to be about $800,000, roughly in line with both Q4 2025 and Q1 2025. Just a quick note that stock-based compensation will fluctuate throughout the year, and it could shift from one quarter to the next depending on when grants are issued. Before this call, we brought in roughly $3.2 million in Q1 through our existing ATM. We expect cash usage in Q1, after subtracting what we raised via the ATM, to come to about $1.4 million. The expected schedule for payments tied to our prime U.S. government contract is weighing on our projected Q1 cash use. Those payments, spread across the year, should give cash flow a boost in the back half. Just to give everyone a heads-up, we're in the process of lining up a new bank to work with, since our aim is to get better terms that bring our costs down and, on purpose, cut our line of credit in half, from $20 million down to $10 million. Thanks for your time; with that, I'll hand things over now to Brian for his closing comments.

Brian Faith executive
#5

Thank you, Elias. Thanks to hard work, commitment, and many long hours, everyone at QuickLogic delivered a number of strategic milestones during 2025, and those achievements have put us in an exceptionally solid position as we begin 2026. I appreciate everything you have all done. Our ongoing execution on our primary U.S. government contract has resulted in that contract growing to a possible $89 million. We added GlobalFoundries along with its 12LP fabrication process — a process that many DIBs rely on today for various radiation hardness needs — and, most recently, we received a $13 million tranche. Separately from that contract, QuickLogic self-funded a strategic radiation-hard, or SRH, discrete FPGA test chip. We currently hold test chips in-hand, and we also have orders for our SRH FPGA dev kit, which will let DIBs assess our test chip against the complete range of radiation hardness requirements. That meaningfully speeds up our capacity to secure both the discrete SRH FPGA designs we are able to storefront and designs more appropriate for embedding our SRH eFPGA hard IP into ASICs. To speed up our storefront business model even more in 2026, we intend to carry out three multi-project wafer, or MPW, tape-outs during this year. Each of those three tape-outs involves chips we plan to sell through our storefront program. Customer contracts will completely offset the cost of two of those tape-outs. One such contract is already booked, while another sits in the final stages of negotiation. As for the third tape-out, we expect contracts to cover it at minimum partially. Working under a revenue-generating agreement with a customer, we created architectural improvements to our core eFPGA technology, and these allow us to serve the profitable very high density markets in both discrete and embedded designs. Those improvements were first created for Intel 18A, and they can be extended to every advanced fabrication node. Considering the strong base provided by the U.S. government contract we recently won, our expectation for continuing mature business of roughly $4 million in 2026, and the several pending contracts now in late-stage negotiation, we feel we are well-positioned to achieve revenue growth of between 50% and 100% in 2026. With that, I will hand the call over for questions.

Richard Shannon analyst
#6

Brian, Elias, I appreciate you taking my questions.

Elias Nader executive
#7

Richard, hello.

Richard Shannon analyst
#8

Hey, thanks. Brian, it seems like you kept the best part for the end, with the guidance for this year. I suppose I'll begin there and ask for some assistance in working through how the dollar growth breaks down, the contributions as we move from 2025 into 2026, and that 50%-100% figure. Could you break that out by storefront contribution, defense versus commercial, and whatever other categories you'd like to use, please.

Brian Faith executive
#9

Like I mentioned, $4 million of that comes from our base mature business, and we feel quite good about that for the year at this point. Then there's the $13 million tranche tied to the U.S. government contract, which brings it to $17 million. When you look at that 50%-100% range, clearly we have to get well into the twenties to reach it. Our expectation is that additional defense-related contracts will come in, whether for one of those MPW test ships I referred to, and/or IP. That would go to defense contractors for use in their ASICs. As I noted, since we've upgraded our architecture to handle higher LUT counts, we're seeing substantial interest in that kind of architecture for some of these process technologies that are tried and true for U.S. defense companies such as 18A and 12LP. Then, going a bit further beyond that, we see other commercial IP opportunities, and I mentioned one during the call that we believed shifted into 2026 from 2025 with that specific commercial customer, mainly because they were considering making the IP core bigger to handle more capability. There were numerous architectural discussions and trade-offs happening that kind of naturally moved that IP contract into what we now forecast as 2026. That particular IP license would involve a non-defense customer. Does that help shed some light on the question, Richard?

Richard Shannon analyst
#10

Yes, that applies in this case. I likely ought to have framed this as a question with several parts, but perhaps I just want a bit of clarity regarding what separates the top and bottom of that range. I believe I caught some of it, yet I would really like to hear you assemble the whole picture, please.

Brian Faith executive
#11

Of course. When you consider the bottom of that range, that would certainly be the base $4 million business, the present tranche we hold for the government contract, and I would mention a pair of IP licenses, one of which would be helpful for one of these MPW tape outs. The top end, or even going beyond the top end, would come about as we add more IP licenses on top of that, and possibly additional funding on the government contract.

Richard Shannon analyst
#12

All right. That's useful. Perhaps a few more questions from my side here. At a high level, whenever we consider Strategic Rad Hard, I'm posing this both in terms of FPGAs and the chance to create store fronts for ASICs that would incorporate your IP. How should we think about, or what do we think about, the timing of wins with any of these DIBs for, you know, what I believe are substantial programs that I believe was the intent of this program from the start here. Give us some clarity on what you anticipate occurring this year as opposed to in subsequent years.

Brian Faith executive
#13

During this year, we anticipate evaluations happening, utilizing our test chips, whether ours or the one funded by the government, and then reaching some form of architecture comprehension with these DIBs by the close of this year, this fiscal year. In the next year, commencing actual development work with those chips. To be clear, this year is very much a year of evaluation. Every one of these companies is extremely risk-averse from a technical standpoint, so they require the time to examine the test chip and confirm that they grasp it and feel at ease with the tools that come with it, which means our software tools and the device and the dev kits themselves. Which means finishing this year with their favorable feedback and sort of a thumbs up that they wish to proceed with architecture insertion in the next year.

Richard Shannon analyst
#14

All right, that's a useful way to look at it. Perhaps we can go back briefly to how you're thinking about this year. You brought up a sales figure, and Elias provided additional figures as well. I couldn't combine them to figure out if we'll end up with positive net income or positive cash flow this year. Could you walk us through your thinking on both, at the bottom and the top of your sales guidance range?

Elias Nader executive
#15

Well, here's what I'd say — we definitely expect positive cash flow during the second half of the year. Not in the first half, Richard.

Richard Shannon analyst
#16

All right. And what about net income or EPS? How does the bottom line look there?

Elias Nader executive
#17

Same. I believe we'll land at the high end during the year's second half, and not the first half either. That's my expectation.

Richard Shannon analyst
#18

Alright.

Elias Nader executive
#19

So, when it comes to net income, we expect to be in the black during the latter six months of the year.

Richard Shannon analyst
#20

All right, that's useful. One final thing from my side, and then I'll hop onto the queue. Brian, you brought up aiming at 3 MPWs in 2025, and I believe a few of the specifics you gave about that have slipped my mind. Could you maybe walk us through how this works, and are these basically continuations of the ones you landed in the prior year, or are they new prospects you anticipate carrying forward? How are we supposed to view them? Also, I don't recall — did you say which process node or whether the founders of those will be involved? Thanks.

Brian Faith executive
#21

Yeah. No process technology was mentioned for these. I won't be doing so. They rely on process technologies that are already in our portfolio. There's no need for us to perform a real physical port onto a new process in order to deliver on them. What we're attempting to get across is that 2 of these will be entirely covered by customer contracts, while 1 would be covered only in part by the contract. The important point being that end customers will be tied to each of the 3. Those customers are the ones pushing the definition of these, and in certain instances, as we said, they are either partly or completely funding the development work on them across the year.

Richard Shannon analyst
#22

All right, that sounds fine. I'll hop on the call, everyone. Appreciate it.

Brian Faith executive
#23

Thank you, Richard.

Neil Young analyst
#24

Hi everybody, I appreciate the chance to pose a few questions.

Brian Faith executive
#25

Hi there, Neil.

Neil Young analyst
#26

Hi, the first question I have relates to that win you announced in the press release involving a high-performance data center. Maybe you could elaborate a little — let us know what the application is, and frankly, any extra information would be interesting. Thanks.

Brian Faith executive
#27

Absolutely. It's a design at 12 nanometers, and what it involves is an eFPGA IP core. In this particular instance, that eFPGA IP core makes up a significant portion of the die area, so it's not just some kind of insurance policy — it's genuinely providing functionality that they built in from the very beginning as a crucial part of what this chip does. Because this isn't a chip built on a three or four nanometer process, it clearly won't function as a GPU-class device. Still, data center printed circuit boards contain plenty of surrounding components that sit alongside those kinds of devices. What we're talking about here is one such example — you could describe it as a peripheral chip — which remains significant and essential to overall functionality even though it isn't at the heart of the compute. We keep moving forward with that effort, we keep working with the customer, and ideally we'll be backing their tape out sometime later in the year. That's the great part about it.

Neil Young analyst
#28

All right.

Brian Faith executive
#29

Neil, I appreciate you raising that point, since this is likely the biggest IP agreement we've signed in recent memory for something that isn't a defense application. Many folks have kept asking us, "Will you ever go past defense alone?" Our answer was yes. It's nice that we can now discuss this specific instance, as it's plainly a non-defense use, and we're hopeful it marks the beginning of additional non-defense uses too. The second one I want to bring up is Epson, correct? We devoted more time to Epson on today's call because of that blog. That's another illustration of a non-defense purpose. Getting to more commercial customers has taken some time, you know, but I believe we're beginning to notice somewhat more momentum and curiosity there, now that these other process technologies are supported.

Neil Young analyst
#30

Thanks. That clears things up. My second question relates to how competition shapes up. You mentioned that during the next year and the one after, the potential storefront business could be quite big for this discrete, Strategic rad-hard FPGA. I wanted to know whether the competitive landscape looks any different from the traditional eFPGA IP you've described. Any details on that competition side would be useful, just to get a better handle on it.

Brian Faith executive
#31

Sure. Let's zoom out to 50,000 feet and ask what the whole programmable logic umbrella looks like. It includes eFPGA and FPGAs. The FPGA rivals most people are familiar with — or maybe I should say peers, since some aren't truly rivals — are Xilinx, Altera, Microchip's FPGA division, Lattice, Efinix, and Achronix; these are roughly the firms making discrete FPGA devices. Among them, if we ask which are U.S.-based, defense-focused, and build devices that fit into this category with some degree of radiation hardness, you can narrow it down and note that Microchip today has devices from its Actel acquisition from years back that provide this rad tolerant, and to some degree rad hard, capability. Xilinx offers some rad tolerant parts, and I believe one rad hard device. Altera has some, though I'll admit I haven't reviewed their product lineup lately. Lattice would presumably like to enter defense but currently doesn't really offer anything there. Achronix, I don't believe, does. Efinix, I think, is mainly concentrated on Asia. So you've already narrowed things quite a bit to only a handful of players doing any serious level of radiation hardness or tolerance. Then if you shift the bar and say it must be manufactured on shore and satisfy strategic levels, I'd challenge anyone to visit the websites of the companies I just listed and identify a device that clears those requirements. I believe it's a null set. I think we're positioned very well in that regard, as we keep executing on this program. The other side of the programmable logic umbrella, as I said, is eFPGA IP. None of the firms I just named have a genuine eFPGA IP business. They'd rather sell you devices, since they likely see IP as cannibalizing device sales. On the IP side, apart from QuickLogic, only a handful of companies have done IP over the past few years. One is Flex Logix, acquired last year by Analog Devices and made captive, so licensing is no longer part of their business. Another is a French startup named Menta, which offers licensable IP. There are also a few extremely small academically-oriented companies that I won't spend time on today. When you're putting us side by side with a single competitor, Menta, once more, I'll circle back to the point that, well, we've been around longer as a business. All those major corporations are our customers; we offer the full range of IP across two devices. Ours is an American firm, with goods manufactured here in the States; at that level, I'd say we bring plenty of differentiation relative to Menta. What matters more, though, is that once you examine the technical specifics, you find Menta operates as a soft IP outfit. By soft IP, I mean that in licensing IP to a client, what they receive isn't merely soft IP — they also inherit an enormous pile of additional effort required to convert it into hard IP prior to inserting it into their ASIC. That enormous pile of effort brings with it substantial risk, plus significant time and expense. As a hard IP provider, which is what we are, we handle every bit of that. There's nothing they must design on their own. Their only concern is figuring out how to architect and deploy this IP. The engagement model differs enormously between ourselves and Menta, and I believe that shows up in the deals we're publicizing, and in who chooses us. It likewise shows up in what our IP typically sells for, correct? We aren't dealing in soft IP priced at $20,000 that arrives bundled with tons of extra labor. What you license costs considerably more, yet we absorb that labor and that risk on the customer's behalf. That's the enormous distinction separating us from Menta when it comes to eFPGA IP. I hope this serves as a useful update regarding competitive dynamics.

Neil Young analyst
#32

All right, that's really useful. Thanks. That covers everything I wanted to ask.

Brian Faith executive
#33

Happy to help.

Tyler Burmeister analyst
#34

Hi, everyone. I appreciate you giving me the chance to pose a couple of questions.

Brian Faith executive
#35

Of course.

Tyler Burmeister analyst
#36

To start, it's really encouraging to see that next portion of the U.S. SRH development program — that $13 million you received — plus the news that the effort has grown to include the GlobalFoundries process. This may build a bit on something asked earlier. I believe you mentioned that, at the high end of expectations for this year, additional funding could come through. So, as far as you're allowed, I'd like to know what upcoming milestones we ought to anticipate or watch for under that program.

Brian Faith executive
#37

Tyler, that comes up often, and sadly I'm not able to share programmatic specifics about this effort. What I can do is return to a point I was cleared to make when this contract was first awarded, back in 2022 — namely, that the overall scope of the contract covers two devices: a test chip and a final chip. I think it was in a December press release, when we revealed the contract ceiling had grown to $89 million and that GlobalFoundries was being added, that we were able to state we had actually taped out a test chip under that contract. So you can mark one of the two off the list. It seems a natural extension that, with additional funding — particularly the rise in funding this year compared with last year — and given that one chip is done while the second is not, we are now starting development on that second chip. Regrettably, I won't be able to provide truly specific information about what goes into the chip, when we tape it out, or when it will be released, but all of it aligns with what we owe the government under this contract.

Tyler Burmeister analyst
#38

Yeah. That's ideal. Thanks for the additional detail; the guidance for the full year was excellent, and I'm grateful for the specifics on it. Just assembling everything, a robust Q1 plus several initiatives converging around the same period. Could we reasonably anticipate some unevenness across the year, perhaps Q2 declining on a sequential basis? Or would you say revenue can increase quarter over quarter, more or less in a straight line?

Brian Faith executive
#39

Our view is that Q1 will be the lowest point this year, in fact. There's a hint for you. The remaining quarters should exceed Q1. Some unevenness is possible, and the reason I say possible is that when contracts you're handling run $2 million to $3 million apiece, particularly when it's IP and revenue is booked upon delivery, then timing naturally gets uneven regarding when we secure the contract and complete delivery within a given quarter, correct? Unevenness could stem from that, though I believe we're attempting to convey this outlook: Q1 is in fact the lowest point this year, and it will rise from here.

Tyler Burmeister analyst
#40

That works well, and I'm grateful for it. Okay, I have nothing further. Thank you, everyone.

Brian Faith executive
#41

Thank you, Tyler.

Gus Richard analyst
#42

Yes, thanks for taking the questions.

Brian Faith executive
#43

Hey, Gus.

Gus Richard analyst
#44

To get a sense of where things stand for Q3, there was mention from you folks of a commercial contract worth $3 million that you anticipated recognizing as revenue in Q4, and it doesn't appear that happened. Does that fall within the guidance for Q1?

Brian Faith executive
#45

No, that isn't the case. During this call, we stated our expectation, or forecast, that it would be contracted in Q2, which means that $3 million is excluded from the Q1 guide. Another point I'd mention, Gus, is that originally we referred to Q3, and afterward during Q4 we noted it might fall there, or might not, and obviously it slipped. This is the situation where we had indicated that what they are now considering is a bigger eFPGA core. Since bigger cores are under consideration, they essentially want to spend additional time on technical feasibility and diligence prior to signing a contract. To be very clear, it is not included in the Q1 guide.

Gus Richard analyst
#46

Okay; that contract was upside guided, if I heard you correctly.

Brian Faith executive
#47

I'm sorry, could you say that again?

Gus Richard analyst
#48

The value of the contract, it was increased.

Brian Faith executive
#49

Okay, so what we actually indicated was that the core itself is bigger. What we didn’t indicate is that its value has gone up. When it comes to how much eFPGA logic they’re looking for, that is absolutely greater than what they had first envisioned.

Gus Richard analyst
#50

Thank you. I understand.

Brian Faith executive
#51

Yeah.

Gus Richard analyst
#52

Moving on, I'd like to ask about that test chip you taped out, received samples for, and are now taking orders on for the test development boards. When should shipments begin? What revenue do you anticipate bringing in, and across how many customers?

Brian Faith executive
#53

There are a number of questions there. Let me go through them one by one. To be clear, we have discussed 2 test chip tape outs up to this point, correct? In public settings, we've mentioned the one funded by the government, and we've mentioned the one we funded ourselves. Since my obligations to the government prevent me from providing updates on the government one, I can only discuss the self-funded one. Regarding the self-funded one, the chips did arrive in Q1. I would say the fab ran a little bit later than our original plan for that. Right now our engineering team is working with those chips and moving through the validation process. As I've mentioned previously, after validation is complete, we'll ensure those units ship to satisfy test chip or dev kit orders. I think we noted earlier that ideally we'd release it before Q1 wraps up. Should it slip to Q2, that's acceptable as well, since the test chip tape-out was deliberately scheduled this way — we built in substantial time cushion so the DIB could receive these and run whatever evaluations they require to feel confident. Regarding your final question, Gus, it concerned the customer count. Since this kind of component is radiation-hardened, the pool of permitted buyers is quite small. Obviously, they must be located in the United States. Its purpose is really tied to strategic defense systems, and only a small number of players actually design any sort of subsystem for those devices—or rather, those systems. As you know, we're aiming for fewer than 5, since those 5 are truly crucial when it comes to these major systems.

Gus Richard analyst
#54

Understood, thank you. I appreciate that clarity. My final question concerns gross margins. How should we view the path of gross margins across the year? Non-GAAP came in at 45% during the first quarter. Does that rise at all in the second quarter, or does it look more like a steady climb? What is the way we ought to think about this?

Elias Nader executive
#55

For the first quarter, we mentioned 45%, plus or minus. The second quarter will probably land in a similar spot, roughly flat. When it comes to Q3 and Q4, I expect a significant boost in gross margins. I should mention that during my tenure here, forecasting and measuring this particular element has proven the toughest challenge, largely due to how we capitalize some COGS and shift other items into OpEx, among other factors. It has been quite a demanding task, though we are making progress. On the whole, for the entire year, I anticipate a solid 57% where gross margin is concerned, which is what I stated in the script.

Gus Richard analyst
#56

Got it. I understand. Thank you so much.

Elias Nader executive
#57

No problem at all. Thanks.

Rick Neaton analyst
#58

I appreciate you fielding this. My only item concerns chiplets. You've mentioned the bridging solution that was previously employed alongside programmable logic. In what ways do you envision chiplet use cases leveraging programmable logic? Which final applications are being considered for some of these? Regarding the latter portion of my inquiry, which deals with bridging — do you mean bridging happening on the chiplet, or bridging linking one chiplet to another?

Brian Faith executive
#59

Okay. Two questions there. One is really the use case.

Rick Neaton analyst
#60

Yeah.

Brian Faith executive
#61

Regarding what these chiplets ultimately get used for, and then I suppose the other thing is, in what way are they divided up inside those packages? Does everything sit in a single chiplet, or does it take several chiplets to address the issues? Correct?

Rick Neaton analyst
#62

Right.

Brian Faith executive
#63

Okay.

Rick Neaton analyst
#64

Is the connection you're making happening among layers within one chiplet, or does it span several chiplets that each have multiple layers?

Brian Faith executive
#65

Yep. Okay.

Rick Neaton analyst
#66

I'm just curious.

Brian Faith executive
#67

There are a few points I can expand on here. Let's begin by looking at the end markets and the applications for chiplets. I believe this has come up in prior discussions, but aerospace and defense represents a very significant opportunity for chiplets, since those customers would rather not develop numerous custom ASICs when they can sidestep that path — their production quantities tend to be fairly modest, and pursuing bespoke ASIC designs is quite expensive. If they're able to achieve heterogeneity within the package, that goes a long way toward reducing their development program expenses. With FPGAs in that scenario, you can essentially ask where FPGAs are deployed in those systems right now, and picture that turning into a chiplet that links up with the other components those FPGAs currently connect to in those systems. Typically, those systems contain some kind of large processor — perhaps a flight computer. Today's FPGA technology is extremely valuable for handling signals arriving from sensors, performing preprocessing on them, and packaging them into packets so that the CPU or SoC itself can work on them without needing to duplicate much of what the FPGA already handles. Keep in mind that FPGAs excel at real-time, massively parallel computation; that's roughly the broad defense application for them. Once again, you can picture that substantial software has already been developed within the defense sector for particular processor architectures. FPGAs are already widely deployed. Putting those die or capabilities into a single package actually reduces the A in PPA — that's area, correct? Many of these systems are pursuing greater miniaturization, and they're exploring combining these components into one package to achieve that. That's the primary application there. Beyond defense, I'd mention there's considerable interest around security, or things that are somewhat shielded for this post-quantum computing era, essentially to safeguard systems from a cyber standpoint. eFPGA, or FPGA, works well for that, because should anything get hacked down the road, if the hardware itself can be programmed, you can then reprogram whatever algorithm resides in those to respond to that threat at that moment; so there's interest there when it comes to making these systems more trusted — and by trusted I don't mean from a defense angle, I mean trusted in the sense that you can be confident it's executing what it's meant to execute and nothing beyond that; that's another compelling application emerging for eFPGA chiplets. The same reasoning applies: they already possess substantial software and infrastructure. How can they modify that existing infrastructure by introducing a bit of programmability rather than redesigning every ASIC? That's a solid application for an eFPGA chiplet. When it comes to bridging, these examples show that we aren't swapping out the entire SoC with this eFPGA. Essentially, you're putting a discrete FPGA's functionality into the very same package that holds the SoC or the ASIC within their system architecture. Now, when it comes to the actual bridging piece, chiplets remind me of the early days when USB or PCI first saw widespread adoption during the PC and laptop growth era — that worked out; it succeeded because everyone could agree on one common interface, correct? PCI is PCI, USB is USB. You build a peripheral around that, you show up to the plug test, it all functions, and everyone's satisfied. With chiplets, things get far messier, since you've got all these varying UCIe flavors, for instance, plus a whole bunch of wires, BOW, and none of them work together. That idea of, like, universal compatibility simply isn't there. Should you go deeper into the specifics, if you'd like to, there's a distinction between the physical layer and the protocol layer, correct? Picture it this way: the physical layer is like me jotting a note down on paper and handing it over to you, right? I hand over the paper. Mm-hmm. You take it, and we're fine. But if the language I'm using differs from the one you know, it'll appear as gibberish the moment you attempt to read that paper. That exact scenario plays out with UCIe, where the physical layers might line up, right? UCIe running on both ends. Yet the protocol that folks layer on top of UCIe varies. That gets made worse when people build ASICs at varying moments in time. One aspect of our eFPGA is this line of thinking: well, that hardware can actually be programmed. So long as the physical layers link up and so long as our FPGA holds sufficient gates, you can likely program them to introduce some degree of compatibility — or, to use my paper comparison, something like a translator. We're optimistic that a few of those will genuinely materialize thanks to that value proposition, and we're beginning to receive some encouraging responses to that notion from what we picked up at the Chiplet Summit last week, and I believe from what we'll pick up at GO MAC next week in Louisiana when we're there presenting. I'm not sure. Does that clarify things?

Rick Neaton analyst
#68

Okay.

Brian Faith executive
#69

with the use cases?

Rick Neaton analyst
#70

Yeah. No, I appreciate the color. Thanks for your time and answering my questions.

Brian Faith executive
#71

Yeah, sure thing; happy to help.

Richard Shannon analyst
#72

Hi, everyone. I've got one final question. Brian, once more, touching on strategic rad-hard, and I suppose I'd like to broaden that to rad-hard perhaps, based on what you said during today's call. What's the count of separate programs you're pursuing? I realize an exact figure won't be shared, but hopefully you can phrase it along the lines of a couple, a few, several, over a dozen, something like that. Just give us a rough feel for how many programs you're bidding on. Thanks.

Brian Faith executive
#73

The near-term ones at the top tier of radiation hardness, I'd put those at fewer than 5 major programs, though within every major program there are multiple subsystems we'd want to be designed into. What probably matters to you is the overall count of socket opportunities, to use that terminology, and that figure would be, I'm not sure, somewhere between 10 and 20 in total; that applies to the highest radiation level, which is where we've concentrated since that's where differentiation is greatest. Once you begin loosening the radiation hardness specs, clearly many new applications open up in the space realm, and you'll find tens of applications there. Our early emphasis, particularly for these initial dev kit orders, will be on those at the higher radiation levels, where nobody is competing with us right now.

Richard Shannon analyst
#74

Understood — that context really helps, Brian. I've got nothing else to ask. Appreciate it.

Brian Faith executive
#75

Thank you, Richard.

Brian Faith executive
#76

Thanks. We’re going to give a technical talk about our chiplet POC during Intel Foundry’s Partners session at the coming GOMAC on March 10th, alongside Cadence and Trusted Semiconductor Solutions. During April, we’ll have a booth at HEART, which is a separate government conference focused on radiation effects, and we will also exhibit and present at IP-SoC Days in Silicon Valley, once again in April. We appreciate your backing and your presence with us today, and we’ll speak with you on the next call. Thanks. Bye.

Operator operator
#77

Thanks for joining; that brings today's call to a close. You're free to hang up now. We appreciate everyone taking part.

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.