Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript
February 10, 2026
Earnings Call Speaker Segments
Thanks, operator, and good afternoon to all of you. Joining us on this call are Ford Tamer, who serves as Lattice's CEO, and Lorenzo Flores, our CFO. Our plan is to walk through a financial and business review covering 2025, along with the outlook for 2026. Should you not yet have our earnings press release, you can locate a copy in the Investor Relations area of our company website, latticesemi.com. Before we continue, a reminder that on today's conference call we might make projections or other forward-looking statements about future events or the company's future financial performance. We want to caution you that these statements are predictions grounded in information available at present, and that actual results could differ materially. Please refer to the documents the company files with the SEC, among them our 10-Ks, 10-Qs, and 8-Ks. Within those documents you will find and can identify significant risk factors that might cause actual results to differ materially from what our projections or forward-looking statements contain. The official guidance of the company for 2026 is included in this call and constitutes it. Should we ever communicate any material changes to this guidance at any point following this call, our intention is that those updates will occur through a public forum, for example a press release or a publicly announced conference call. Throughout this call we will refer primarily to non-GAAP financial measures. Through the disclosure of certain non-GAAP information, management aims to give investors additional information allowing further analysis of the company's performance and underlying trends. For historical periods, reconciliations of these non-GAAP financial measures to GAAP financial measures have been provided, and you can find them in the Investor Relations area of our website, latticesemi.com. I'll now hand the call over to our CEO, Ford Tamer.
Thank you, Rick. Good afternoon, everyone, and welcome to this call where we'll discuss results for the fourth quarter and the full year 2025. When 2024 came to a close, we shared our plan, and throughout this past year we executed it precisely as we had outlined. That outcome came from honoring our commitments, stabilizing the business, restoring channel inventory to normal levels, improving execution, and positioning Lattice to capture two of the industry's most powerful secular trends: data center AI and physical AI. Now that 2026 is here, the thing that excites me most is seeing low-power FPGAs adopted broadly at an accelerating pace, becoming the everywhere companion chips. Since the Super Bowl is probably still top of mind for a lot of people, here's a sports comparison. Think of the main processors — GPUs, custom AI accelerators, CPUs, and MPUs — as the most valuable players, the MVPs, on the system. Those MVPs bring a lot of power, but a single player can't win a game, much less a championship, without the rest of the team around them. That team, in this case, is Lattice. What we bring are the FPGAs — the companion chips you find everywhere — which handle a large share of the essential system jobs. Those jobs include boot, fire sequencing, security, control, IO expansion, board and rack management, leak detection, power and cooling, bridging, sensor aggregation, sensor fusion, preprocessing, and plenty of other important system functions. We deliver this across both data center AI and physical AI. Our reach spans every key sector — communications, compute, industrial, automotive, aerospace, defense, medical, and consumer. We also serve several of the globe's quickest-expanding use cases: security, rack management, communication, quantum cryptography, humanoids, industrial automation, logistics, robotaxes, space, and AR/VR wearables. Those capable companion chips deliver widespread interoperability throughout all of these essential functions, markets, applications, and varied suppliers. Lattice can be seen as the Switzerland of data center and physical AI applications. Another major advantage of having a companion chip approach is that it generates growth that's both sustainable and diversified across every market and application, and that's assisting us in hitting our financial goals. This companion position also turns into something foundational as AI workloads keep raising system-level complexity and create pressure for quicker time to market. With AI servers splitting apart into processor boards, networking cards, security cards, power and cooling modules, storage, and other specialized blades, FPGAs turn up all over the data center. That same trend is picking up speed in physical AI as well. Intelligence is shifting nearer to the sensors, which is where data gets generated. One robot might carry several kinds of vision sensors—cameras, LIDAR, radar, and infrared among them—which means you need fusion, aggregation, and preprocessing. On top of that, a humanoid robot may contain dozens of motors, and those demand control that is both highly precise and very low latency. That's where companion FPGAs come in: they sit alongside the sensors and actuators, keeping high bandwidth vision streams in sync with real-time motion and producing deterministic responses. Rather than relying on software executing on a microcontroller, Lattice FPGAs handle this directly in hardware, which makes it simpler to promise identical cycle accurate responses on every occasion. Finally, certain use cases call for FPGAs to serve as the primary compute engine as well—areas such as signal processing, real-time networking, and what people now refer to as far-edge AI. At Lattice, when we talk about Far Edge AI, we mean contextual AI that sits close to the sensor and operates using compact, self-contained models. That momentum continues to grow. One recent illustration: we landed a design win for a human-machine interface, or HMI, use case in industrial robotics, and interest keeps rising for applications that stay under one tops and under one watt. After our most recent worldwide sales conferences, we came away with considerable momentum that strengthens our confidence heading into 2026. The pace of design wins is accelerating across both data center AI and physical AI, and we're excited about what's coming with our Tier 1 customer deployments. With that said, let's move on to the numbers. During Q4, revenue came in at $145.8 million, which is 9.3% higher sequentially — our best sequential result in seven years — and 24.2% higher year-over-year. Revenue for the full year reached $523.3 million, matching expectations. Turning to what's next, our Q1 revenue guidance sits at $165 million at the midpoint, which works out to more than 37% year-over-year growth and shows how confident we are in a robust recovery and quickening momentum. At the midpoint, our Q1 EPS guidance of $0.36 works out to almost 65% growth year-over-year, since we anticipate continuing to produce earnings growth that outpaces revenue growth. As you can tell, plenty of positive operating leverage is working in our favor; we run on mature nodes, so our capital spending stays more reasonable than that of competitors working on advanced nodes. That lets us generate EPS growth considerably faster than revenue growth. New products continue to be a crucial engine for growth over the long term. During 2025, revenue from new products climbed roughly 70%. We're still on track for new product revenue to hit the mid-20% range as a share of total revenue in 2026, and we enter the year with strong momentum as adoption of Nexus and AVANT keeps widening. Considering how large the opportunities in front of us are, Lattice is investing for the future as appropriate. Our slogan for 2026, "Go Big, Be Great," captures both our ambition and our dedication. We're putting investments into silicon, software, systems, operations, and infrastructure to back growth at scale and to broaden our leadership in small and mid-range FPGAs. To sum up, 2025 brought disciplined execution and meaningful progress; we steadied the business, generated tremendous momentum in data center applications, moved our product and software roadmap forward, and markedly enhanced operational performance, which includes normalizing channel inventory. We stepped into 2026 with great confidence. Several factors underpin that confidence: a robust backlog, sustained demand from data centers, a recovering industrial sector, growing adoption in companion applications, and a steady stream of new products entering the market. We remain focused on innovation that sets us apart, strengthening relationships with customers, and delivering value to shareholders over the long term. At this point, I'll turn things over to Lorenzo, who will walk through our fourth quarter and full-year results in detail. Lorenzo?
Thanks, Ford, and hello everyone this afternoon. We'll kick things off with a quick look at how we did in fiscal 2025, covering the fourth quarter as well, and then we'll move into what we expect for the first quarter of 2026 along with how we're currently thinking about fiscal '26. When it comes to the whole of 2025, we're happy to share that Lattice met expectations, with revenue, gross margin, operating profit, and EPS all landing in line with the full-year outlook we gave, the fourth quarter included; we pulled this off while finishing up our transformation and putting money into areas we think will fuel profitable growth over the long haul. Revenue for all of 2025 came to $523.3 million, a 2.7% rise, matching what we had anticipated. The increase came from wins on our platform in communications and computing, where revenue climbed 28%. That was partly countered by an 18% drop in industrial and automotive revenue, which we had anticipated as we worked through normalizing channel inventory over the course of the year. Our non-GAAP gross margin keeps showing the value we bring to customers, and it grew 190 basis points to 69.3% in 2025. We boosted both revenue and gross margin during 2025 while also trimming non-GAAP operating expense by about 1%, bringing it to $213.5 million. Once restructuring wrapped up, we directed investments toward talent, infrastructure, and technology, giving us a stronger and more efficient platform for driving growth over the long term. Consequently, during 2025, non-GAAP operating margin grew 340 basis points, and EBITDA margin rose 320 basis points, reaching 35%. Non-GAAP EPS came in at $1.05, up 17%, which shows the leverage in our business model since earnings grew quicker than revenue. Other figures moved in the right direction as well; during 2025, GAAP net cash flow from operating activities rose to $175.1 million, compared with $140.9 million in 2024, while GAAP operating cash flow margin climbed to 33.5%, versus 27.7% in 2024. Free cash flow for all of 2025 totaled $133 million, with a free cash flow margin of 25.3%, compared with $120 million and 23.5% in 2024. Turning to Q4 in particular, our performance began to accelerate, and that will carry on into 2026. Revenue in Q4 reached $145.8 million, which was 9.3% higher than the prior quarter and 24.2% higher than the same period a year ago. The improvement came from a record showing in communications and computing, which rose 25% from the previous quarter and 60% year over year. As Ford mentioned, we're clearly gaining from the extraordinarily strong growth in data centers. Gross margin in Q4 was 69.4% on a non-GAAP basis, a slight decline from Q3. Even with our mix changing, our gross margin still reflects the value and differentiation our products deliver to customers. Non-GAAP operating expense rose to $56.4 million, roughly 5% higher than the prior quarter and 7% higher year over year. As Ford and I have said before, we see major opportunities both near term and long term, and we're putting money into extending our future leadership in the small and mid-range FPGA markets as well as our companionship program. In Q4, non-GAAP operating margin grew 170 basis points to 30.7%, and EBITDA margin rose 90 basis points to 36.5%. On a non-GAAP basis, fourth-quarter earnings per share rose 14% versus the prior quarter, reaching $0.32, which matched the guidance we provided. For 2025, GAAP net cash flow from operating activities climbed to $57.6 million, compared with $47 million in Q3, while our GAAP operating cash flow margin came in at 39.5%, versus 35.3% in Q3. Q4 free cash flow totaled $44 million, a 30.2% free cash flow margin, improving from $34 million and 25.2% in Q3; this continues to be an area of focus for Lattice, and we anticipate that the upward trend in free cash flow margin will persist. These gains are being delivered even as we deliberately direct CapEx toward R&D and projects aimed at operational improvement. One final note on the results: we hit our overall goal for channel inventory and are positioned well to capture growth across every end market we serve. Let me now move to capital allocation. Our balance sheet is still solid, since operations have strengthened our cash flow generation, and we carry no debt; should the need arise, capital is readily available to us; this puts us in a good position to handle macroeconomic uncertainty and to invest for future growth. Because of the strength of our balance sheet and our business model, giving capital back to shareholders continues to be a central piece of our capital allocation approach. Across the full year 2025, we bought back roughly 1,800,000 shares, equivalent to $100 million of the company's common stock. With that repurchase program finishing in Q4, our Board of Directors gave authorization for the company to buy back another $250 million of its outstanding common stock; turning now to guidance, which includes our usual level of detail for Q1 2026 along with extra commentary to help frame expectations for the full year 2026. Our outlook for both Q1 and the full year 2026 captures the extraordinary demand strength we are seeing right now. For Q1 2026, we anticipate revenue rising to a range of $158 million to $172 million. At the midpoint, that represents 37% growth over Q1 2025 and 13% growth versus the prior quarter. On a non-GAAP basis, gross margin should come in at 69.5% plus or minus 1%. Non-GAAP operating expense should land between $59 million and $61 million. The bulk of OpEx growth will come from R&D and mirrors disciplined investments meant to fuel future growth. For Q1, the income tax rate should be 46% on a non-GAAP basis. Non-GAAP EPS should fall in a range of $0.34 to $0.38 per share. As for the full year, we enter 2026 enthusiastic about the catalysts creating multiple growth opportunities in front of us. I want to walk you through our financial framework for the year. We have said before that we have high confidence in delivering at least 20% year-over-year growth, and how the year has started suggests we now have better visibility into growth beyond that level. We intend to provide more specific guidance as the year progresses. Looking ahead to the entire 2026 fiscal year, we anticipate gross margin landing within the same band we've outlined for Q1, though there could be some movement across the year depending on how the customer mix shifts. Turning to OpEx, we still see plenty of room to extend our position as a leader in small and midrange FPGAs. Those investments, mostly in R&D, are what push OpEx higher in Q1. Q2 should bring a further step up, followed by more modest growth in the back half. We keep a tight grip on the knobs and levers that govern OpEx and will make adjustments as needed to reach our profitability targets. For income tax, we expect a rate of between 47% on a non-GAAP basis. Now that channel inventories have normalized, 2026 revenue ought to track consumption more closely, creating a strong setup for growth in 2026 and beyond. We'll keep partnering with our channel to serve end customer demand while trimming channel inventory where it makes sense. On internal inventory, we're intentionally building stock to support this growth environment. Given the strong growth potential we see and the long life cycles of our products, the risk of obsolescence strikes us as low. To wrap up, we stay focused on executing our strategy and making investments that strengthen our leadership in small and mid-range FPGAs. We have high confidence in driving revenue growth while growing EPS at a faster rate. Operator, that brings our formal remarks to a close. We're ready to take questions now.
Thanks. Ford and Lorenzo, nice work on wrapping up the year so well. For my first question, Ford, could you take us through the server growth, which was clearly very strong, and the data center dynamics you've discussed, plus the MVPs in that space? When we think about how you view this business moving ahead, we've been hearing quite a bit regarding server CPU momentum from a few of the CPU vendors. There's also been some supply tightness in the picture. So perhaps you could lay out those dynamics, along with how your growth breaks down between unit volumes and dollars of content per server? Thanks.
Thanks, Ruben. Indeed, we're seeing robust demand for as far out as we can tell; the CapEx increase is fueling it, our attach rates are fueling it, higher ASPs are fueling it, and a greater count of applications we keep discovering on these servers is fueling it. At a broad level, then, when we consider server units climbing from roughly 15.3 million units in total to approximately 16.5 million units in 2026, our attach rates have climbed consistently with each generation. During the 2024 period, we sat in the mid-1s. As for attach rate, we were in the mid-2s a year ago, and this year we'll exceed three units per server — FPGA units per server. Multiply that by server unit counts, with our attach rate rising, and our ASP also rising from about $3 to more than 4. Once you combine all of it, you can see this is producing a very healthy overall gain in that business.
Fantastic. I appreciate all those specifics, Ford. Lorenzo, perhaps a quick follow-up regarding the Q1 outlook, which is quite impressive. Could you take us through the underlying segment-level assumptions? It seems we are at last witnessing a bit of a rebound in industrials and automotive, and clearly, communication and compute keep expanding, but maybe you could break that down for us, please. Thank you.
Yes. I believe, taken as a whole throughout the company, we are simply extremely optimistic regarding the underlying trends we are observing. Naturally, the OMS and compute portions of our operations are the frontrunners in terms of our momentum. Yet, as we mentioned, we have reduced our channel inventory close to our objective, and we anticipate seeing expansion there too.
Hi, thanks for taking my question, and congratulations on the strong outlook. I realize how much effort went into reaching this point. Let me try the revenue side. It doesn't seem like you're very willing to provide full-year guidance for 2026, but three months ago, you said comms computing might rise 20% to 40%, and industrial might rise 5% to 15%, if my memory serves. You're beginning this year much stronger than those figures. So in terms of magnitude, can you offer any additional detail? And how should we think about which of the two grows more quickly, and where total revenue growth might land for '26?
Yes; things are very strong. Our confidence has grown regarding demand in both segments, Melissa. So first, thanks for asking; and indeed, the demand picture today is far stronger than it was three months back. This spans both the comms and compute and industrial markets; so in comms and compute, as we mentioned, we're very well booked for this year and are now booking into '27, which gives great visibility; and in industrial automotive, where we ship to true demand. Channel inventory sits below three and will be in the twos during Q1. So, we feel confident about growth in both segments.
Got it. So to build on that, a large portion of your prepared comments focused on the physical AI opportunity. I think the companion piece and the attach rate in data centers are fairly straightforward for us to track, but estimating the potential size of the physical AI opportunity is somewhat more difficult. So I'm not sure whether you'd prefer to frame it as a TAM or through some sort of units framework, but is there any way you can help us gauge how large this physical AI piece might be? Because it does appear that you're seeing a lot of momentum there.
Yes, we're seeing tremendous momentum there along with this companionship strategy. We've got a really solid relationship with NVIDIA, for instance—we jointly went public with the Holoscan reference design, which takes various video feeds and routes them through the FPGA and then into NVIDIA's processors. On the partnership front with NXP and their microprocessors, we've been advancing nicely. Back in January, we held a gathering in Europe that brought together the leading industrial, automotive, telecom, and aerospace defense players from across the continent, alongside us and NXP, to kick off the year. Beyond that, we're gaining ground with a number of additional collaborators—some in sensors, some in analog, and others in microprocessors. So keep an eye out for more on that. There's a lot of room for growth ahead, since we're picking up share in markets where we previously didn't have much of a presence. Aerospace defense is the clearest illustration — back in '24 we had barely any content there, and now we're looking at '27 shaping up to be a really strong year. Every region — the US, Europe, Asia, all of them — is seeing some impressive design wins in aerospace defense tied to our new red hard and red tolerant designs. On the industrial robotics side, we keep expanding, both in industrial robotics and in AMRs. When it comes to humanoids, the opportunity is enormous. In vision, we're winning everywhere. So in humanoid vision, our results have been excellent. We secured a pair of marque design wins along with one for motor control. Our view is that this serves as a foothold for educating customers about the way we'll deliver improvement, since FPGA hardware-based solutions offer latency, determinism, parallel performance, and accuracy that far outstrip what alternatives provide. As a result, we're quite enthusiastic about the market potential humanoids hold for us as we head toward '27. Robotaxis have likewise seen substantial advancement on our part. Specifically, we've landed a robotaxis win that, once again, has us very excited as we move toward '27. Across physical AI, we're also gaining ground in medical and telecom. Because our applications are far more varied, and spread so widely among our 11,000 customers, plus the extensive reach of our channel, pinning down as precise a number for the server side isn't really possible. Still, we're genuinely enthusiastic about what industrial can bring, particularly once advanced arrives in 2027.
Hi everyone. I want to give 2026 a try. When I look at the segment guidance the two of you mentioned during the previous quarter — 20% to 40% for comms and compute, plus 5% to 15% for industrial and automotive — it seems that given how strong the first quarter guide is, you're likely on pace for at least the top of those ranges, maybe even beyond. Is that the correct way to view growth across those segments? Or would you say one segment is meaningfully better than what you thought ninety days ago?
Yes, and thanks.
Hi, Quinn. I appreciate you bringing that up — as you mentioned, this came up 90 days ago. During our commentary, you should have noticed, all of you should have noticed, that our confidence and the foundation for our outlook have grown quite a bit over those 90 days. So the general perspective holds that comms and compute will be the fastest-growing among our end market sets, but we're also witnessing gains across the board in our industrial end markets too. And in every geography.
Yes. Quinn, one thing worth noting is that even using the top end of the guidance we provided last quarter, you still fall short of the revenue we need to reach this. So you can figure that we're at the upper end or beyond it — the high end on both markets.
Yeah. Given that you beat by nearly $20 million versus consensus, I'd figure you're likely toward the upper end. Okay.
So then, Ford, I suppose — you said while answering one of the questions that you're booking, that comms and compute are essentially fully booked out through 2026. You're beginning to book into 2027. Could you address, first, lead times in general? Have they been extending? Do you feel at ease with those lead times? And since you're already taking bookings into 2027, does anything suggest you might be seeing double ordering start to appear, given how far ahead you're booking now? Yeah. Quinn, as you'd expect, that's a really critical question. As a management team, together with our channel partner and our customers around the world, we've devoted a great deal of time to really examining it; and right now, we don't think double ordering is happening. What I'd say gives us confidence is that these orders get scheduled across the year. So they aren't all piling into Q1. They're spread across '26 and into '27. So you're able to see orders five, six quarters ahead. Clearly, lead times have grown for us and for everyone else in the industry. With our customers, we've been extremely transparent and proactive. And we've been flagging these longer lead times. And I believe that by flagging them, we're doing our customers a service. We're noticing competitors who might not be as transparent, and we are being extremely transparent and making certain we collaborate closely with both our suppliers and our customers to handle the situation. And, you know, since this summer we've been ordering substrate and placing additional orders. So for roughly the last six months now, we've been raising our supply and our orders with our suppliers, which puts us in solid shape to get through the coming few months.
Hi, everyone. Thanks for asking. I might just add to what Quinn said regarding M as well. I think INA came in a bit below expectations during the prior quarter. Do you believe some inventory burn occurred there? And as we look ahead, is this the bottom for growth, with improvement starting from here, possibly even alongside some inventory build on top of that? What would you all consider the typical quarterly rate for INA, absent inventory burn, that we could use as a baseline? Thanks.
Sure. Let me share the framework we use. First, the total inventory figure we referenced was, of course, our combined number at year-end spanning every one of our end markets and our distributors.
And looking back to our starting point, it first declined in comps compute, and after that, as you mentioned, in industrial and auto.
At this point, what we've observed is pressure, in certain situations, throughout every end market coming from our channel partners to accumulate inventory. Therefore, we will keep managing that downward, with the ultimate goal being not truly channel inventory, but rather reaching appropriate levels of customer support and responsiveness. Consequently, there will be moments when we may need to increase it, to meet anticipated demand, and there will be moments when we'll keep managing it downward tactically; so, I don't believe it would really be that helpful to provide you with another figure to use as a baseline.
All right. Thanks. Then for you, Ford, perhaps a question with two parts. I believe what you provided was data on server CPUs, but could you discuss FPGA to XPU, or processor—AI processor—attach rates? For instance, what they have been and what you expect going forward? And finally, you've mentioned Lattice becoming a bigger company, maybe through buying new IP and capabilities. Could you update us on that, especially since TI is acquiring Silicon Labs?
Yes. Chris, that's a good question. What I discussed was server units, not server CPUs. So our estimate, based on market research of the server unit TAM, is 15.3 million units in 2025 rising to 16.5. Within that overall TAM, we've previously stated that our attach rate during 2025 was roughly 12% for AI servers, which is what we observed. And that appears to be about right. Those 12% AI servers equated to approximately a high teens share, call it 20%, of revenue. That should give you a sense that AI servers account for a higher revenue percentage than traditional servers. Lately, we've observed a solid improvement in traditional servers. In fact, traditional servers are even expanding more quickly than anticipated. So we're seeing decent growth across the board. Regarding the inorganic side, we have in fact finished four small tuck-ins that weren't publicly discussed. Those are all small tuck-ins of the IP and software variety. We're still pursuing a few additional tuck-ins of this kind. Right now we're investing in software tools and IP solutions enablement. And we'll keep doing the small tuck-ins. A larger acquisition is something we'd like to pursue, but we want to be patient and choose the right one. Every star must align. It needs to make strategic sense and be accretive, and the teams all have to want to work together; so many different factors are involved, but we're not rushing. As you can see, the business is quite strong, so we'll act at the appropriate time.
Yeah. Hey, everyone, I'd like to add my congratulations about those impressive results. Hey, as you keep picking up share and companionships in the data center, are you noticing any greater competitive pressure or aggressive pricing coming from certain rivals who are attempting to win back these sockets?
So Kevin, our strategy was built around the customer. So the customer always comes first for us. We ensure we meet with our customers about how we can deliver a solution offering the greatest benefit. Allow them to expand their revenue, boost their success in the marketplace. And we're extremely, extremely focused on giving them low power, small size, low cost, fast boot time, longevity, and a whole bunch of attributes that we think set us apart in the small and mid-range FPGA versus competitors. So we've watched our market share increase, and we anticipate that it will keep increasing. And so within that small to mid-range FPGA, we feel very strong about our market share and convicted about where we are.
Understood. Alright. Yes. That's completely logical. So from where we sit, when you consider what remains of 2026, I mean, you folks are operating at maximum capacity. You know, what's causing you sleepless nights? What might possibly throw the Lattice story off course?
At this juncture, I'd say the thing causing us sleepless nights is our very own pursuit; we've set some highly ambitious plans regarding where we're putting our investment dollars and what we must carry out and deliver against. And we're intensely concentrated on meeting those plans. We need to guarantee we've got inventory available for our customer; we need to guarantee we remain a strong partner in what we refer to as the three S's.
The, uh, solutions, support, and supply, plus supply. So those three S's really matter to us. Particularly when conditions are like what we're seeing now, with supply being scarce. So a great deal of our time goes into making sure we deliver the right supply. And our aim is to keep being strong partners. So I'd say what concerns us most is our own ability to deliver against plans that are quite ambitious.
Yeah, that's completely logical. Alright, great. Thanks for the detail. Congratulations once more on the results.
Thank you, Kevin.
Hi everyone, I appreciate you letting me ask something. During the prior quarter, I believe you discussed design wins reaching an all-time high when measured by lifetime value. I'd like to know what the totals looked like for Q4. Also, when you reflect on fiscal year 2025 and add up the lifetime value from every one of those design wins, what rate of revenue growth can that sustain?
Yes. Thanks, Gary; design wins have kept reaching unprecedented levels for us. Quarter after quarter, we keep setting new records on design win. Q4 gave us yet one more record-breaking design win quarter; consequently, we keep accumulating a bigger count of design wins that will enter production during this year and the following year in '28. That therefore positions us for growth across multiple years, underpinning no less than 25% growth.
Thanks for that, Ford. Whether this matters much or not, I'm not sure, but I picked up on the fact that the revenue range in this quarter's guidance sits at $14 million, whereas it's usually $10 million. Is there something worth flagging there? Could you perhaps walk through the puts and takes between the low end and the high end of that guidance range?
No; honestly, there are a few dimensions to it. For one, there's simply the quantitative angle — as revenue climbs, you're working with a comparable band surrounding the midpoint, and that lands you on a bigger figure. Beyond that, as we considered what the revenue potential might be while putting it together, you had varying possible scenarios regarding what we'd be delivering to customers and when during the quarter that would flow through.
Honestly, I'd tell you the midpoint of that is something we feel genuinely, genuinely at ease with. And then, I suppose the thing that matters is, you know, what the shape of things looks like as the quarter unfolds. Still, that's something we're quite at ease with too.
Hi, everybody, and thanks for the chance to ask. Nice work on the quarter and the outlook. Several of my colleagues attempted to pull more forward-looking detail and didn't get anywhere, so I'll flip the direction and look backward instead. Thinking about 2025 in comms and computing, could you break down what fueled the acceleration — we went from under 60% growth in 1Q to more than 90 in 4Q? Some kind of ranking among the major sockets would help, or perhaps how GPU, CPU, and ASIC stack up against each other, plus roughly how big a drag the PC softness was early in the year. Whatever insight you can offer about how you all are positioned going forward would be really valuable, I think.
Absolutely. At a broad level, when we describe the way we try to characterize expansion across comms and compute, we separated server growth from wired comms growth. Server grew 85% compared with the prior year. That makes it clear the server piece was the fastest-expanding part of comms and compute. Next, comms as a whole rose 61%. That figure, though, takes in wireless, which means wired comms is running above it. Therefore, the second growth driver you'd be able to identify is wired communications. And looking inside those two—assuming server ranks first and wired comms ranks second—the AI server component represents the quickest-expanding portion of the server business. Regarding that AI server component, we went over the different elements involved. First, we talked about CapEx: it rose as the year progressed, and by year-end it stood well above where it had been at the start. That same pattern has shown up over the last three months. Josh, it appears this is accelerating even further. When December came around, our assumption for 2026 CapEx among the top five hyperscalers was $500 billion. By January, we figured it would land nearer $580 billion, and after the latest earnings reports, the figure sits at $740 billion. So across December, January, and the start of February alone, that estimate has climbed quite a bit. The same thing happened as we moved into and then out of '25 — the CapEx projection continued rising all year long. Second, we hold design wins across every GPU and AI accelerator, and with each new generation, the amount of our content goes up. When you move from one NVIDIA generation to the next, and as a CPU generation grows, along with a generation from Meta, Amazon, and Microsoft, our attach rates increase each generation. So throughout the year, as new generations come out, that's item number two. ASP is the third item. Going forward on ASP, more components are being designed in for security. As a result, we have several new use cases; for example, we started the year with a strong use case in boot, power sequencing, control, IO, connectivity — that's the core FPGA work. Throughout the year, we started getting designed into additional attestation work as well as more post-quantum cryptography. Leakage detection is another use case that we pinpointed in the second half of the year, and it's now being deployed. That's because our latency for spotting the leak and shutting it down was so strong. You know, deployment occurs as soon as they find out about it, since it's an absolutely essential piece of this AI infrastructure; so those attach rates keep rising. And as those attaches continue rising and we shift into these larger security-focused use cases, that's driving up the attach rate for larger FPGAs, which then leads to a higher ASP. The way these factors stack on top of one another is what drives that number progressively upward. And that momentum is going to keep showing up through 2026 and 2027. Alright. So we feel great about the direction this takes us from here.
Thanks for providing all those details. One more question — perhaps my calculations are off, but when I take that 70% growth figure for new products and pair it with what I believe was a mid-teens result in 2024, I end up with new products already accounting for well above 20% of exposure in 2025. Any guidance you could offer regarding where new products will end up in 2026? I believe your comment was something above 20%. Is some of this simply that pre-Nexus components are returning now that those issues are fixed? Or should we continue to anticipate that new products will expand faster than Prenexus components? Thanks.
Yes. Your arithmetic checks out. You calculated it properly. Nice work. We did clear 20; as a total percentage, we're sitting in the low twenties. So that's right. That 70% year-on-year growth lands us in the low twenties as a share of total revenue in 2025. And for '26, we anticipate that moving into the mid to high end of the 20% range.
Okay. Thanks.
Thanks. Nice work on a strong quarter, everyone. I hopped on a bit behind schedule, so I’m sorry if this has come up already; Ford, my question concerns the supply side. I realize you mentioned demand has been far stronger than anticipated. So I’m just curious whether you’re encountering any supply limitations on your end or shortages of other components, like memory, and so on? Are any of those shortages affecting your business?
Yeah; the first portion of that was covered before. I’ll go over it once more. The second portion is fresh, Srini. So I appreciate the question. First, regarding supply, as I mentioned, we remain dedicated to our customer across three S’s: support, supply, and solutions; and since the summer we’ve been placing larger material orders. And I’ve watched lead times climb steadily from the first half of last year, then the second half of last year, then the end of last year, and now they’re rising extremely, extremely quickly and continuing to rise. We’ve been quite transparent with our customers and given them advance notice about this. And, you know, we’re in fairly decent shape, though there are clearly spots such as substrate and assembly that are extremely tight everywhere. And we’re collaborating with our customers, our suppliers, and our channel partners to ensure we make the correct allocations and ensure nobody gets left hanging. But right now it’s certainly an extremely tight situation. On your second point, the one about memory, we’ve posed this question many, many times now to our customers, and they reassure us that they hold supply agreements with their memory suppliers, and that the forecast going into Lattice accounts for that memory supply situation; so clearly we’ve devoted considerable time to this to ensure that, you know, our demand was solid, and we’re being assured by every major cloud service provider, OEM, and ODM that they have supply locked in through their supply agreements.
Understood. And, Ford, from a strategic standpoint, you appear somewhat more willing to consider bigger M and A. So I'm just wondering what the requirements would be in your case, since, well, I'm not sure there are many FPGA companies available for you to consider. So I'm just curious what some of the measures and end markets might be, and so on. Plus, one thing people in the industry worry about is that, well, any kind of bigger deal would need regulatory clearance, especially from China. In your view, is the landscape shifting on the regulatory front?
Yeah. Thanks, Srini. So I already gave an answer to that. I'll give it once more: we have in fact completed several small tuck-ins. We completed 4 small tuck-ins involving software tools and IP that we have not announced publicly, though that's because they are small. And we're working on a couple additional ones. So these small tuck-ins are something we really favor, since no approval is required and they are pre-strategic, pre-accretive, and bring some very important skills in IP to our toolset and our team. So we've completed them; we will keep doing them. And no regulatory approval is required. As for the large M and A, we've been transparent that we would pursue one down the road, but we're not rushing. A whole bunch of factors have to align. The stars have to align, and the first is strategic. Second, it has to be accretive. Third, the business model has to make sense. And fourth, the teams have to get along and we need to have some synergies, both top line, bottom line, product roadmaps; so all these factors, and you regulatory is one extra one. So we're being patient. Right now, the good news is the business is extremely strong on our organic side; so we're not in a hurry to do one.
Thank you, Ford. Best of luck.
I appreciate all of you being with us on today's call. During this quarter, we'll be present at these investor events: the Susquehanna Technology Conference, taking place in New York City on February 26, and the Morgan Stanley Technology Media and Telecom Conference, held in San Francisco on March 4. That brings our call to a close. We're very grateful that you took part, and enjoy your evening.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Lattice Semiconductor Corporation transcript - plus 131,000+ 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 Lattice Semiconductor Corporation earnings transcripts and 131,000+ 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.