Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript
November 3, 2025
Earnings Call Speaker Segments
Thanks, operator, and good afternoon to all of you. Joining me on this call are Ford Tamer, who serves as Lattice's CEO, along with Lorenzo Flores, Lattice's CFO. Together we'll walk through a review of both financial and business matters for Q3 of 2025, plus the outlook for Q4 of 2025. Should you not yet have our earnings press release, you can locate a copy in the Investor Relations area of our corporate site, latticesemi.com. I want to remind all listeners that throughout today's conference call, we might offer projections or other statements that look forward, concerning upcoming events or the company's future financial performance. We caution that these statements amount to predictions grounded in information available at present, and that actual outcomes could differ materially. We point you to the documents the company submits to the SEC, among them our 10-Ks, 10-Qs and 8-Ks. Within those documents, significant risk factors are contained and identified that might lead actual results to diverge materially from what our projections or forward-looking statements describe. This call includes, and constitutes, the company's official guidance for Q4 of 2025. Should we, at any point following this call, share any material changes to that guidance, our intention is that such updates will occur through a public forum, whether a press release or a publicly announced conference call. During this call, we will refer primarily to non-GAAP financial measures. Through disclosing certain non-GAAP information, management aims to give investors additional information enabling further analysis of the company's performance and underlying trends. For historical periods, we provided reconciliations of these non-GAAP financial measures to GAAP financial measures, which are available in the Investor Relations section of our website at latticesemi.com. I'll now hand the call over to our CEO, Ford Tamer.
Thank you, Rick, and good afternoon to all of you joining our third quarter earnings call. Having just completed my first 12 months as CEO, my confidence in Lattice's upward path has never been stronger. The solid results we posted in Q3 and the guidance we're providing looking ahead both speak to how well our strategy and execution are working. Backed by an exceptional team and a strong pipeline of innovation, we're positioned to take advantage of the constantly growing investments being made in AI and data center infrastructure. At last month's Open Compute Summit, interest in Lattice's low-power data center offerings kept building. We also saw momentum accelerate for Lattice's security and board management solutions, which we supply to hyperscalers, neo-cloud and server and communications OEMs and ODMs; and at that same event, we showed off our leadership position in post-quantum cryptography, or PQC. I'm happy to share that adoption of Lattice's PQC technology is speeding up as well, driven by the NIST mandate that systems achieve CNSA compliance. We closed out Q3 with even greater confidence, as shown by fresh use cases for Lattice products throughout our core markets, along with an acceleration in design wins, which were on track for a record year in 2025. This momentum underscores our differentiated value proposition — low power, small size, fast boot time and high reliability — and lays the groundwork for rapid growth into 2026. Broadly speaking, we keep seeing data center investments widen across all payers and applications within the AI infrastructure tsunami. Lattice is gaining from corresponding revenue growth, reflected in rising bookings now and into 2026. That gives us more confidence to invest further for growth in 2027 and beyond; we strongly believe that delivering above-average revenue growth aligns with why our shareholders put their money into Lattice. For Q3, revenue came in at $133.3 million, a 7.6% increase over Q2. That marks the highest sequential growth in over 4 years. Our Q4 revenue guidance of $143 million at the midpoint works out to 22% year-on-year growth. That estimate is the biggest increase in almost 2 years and demonstrates our belief in a strong recovery and upward momentum. Turning to end markets in the third quarter of 2025, Communications and computing rose 8% sequentially and 21% year-over-year, reaching a record level. Growth in the computing subsegment is being fueled by our expanding footprint and more use cases in both general purpose and AI-optimized servers. And growth in the communications subsegment continues to be fueled by wired data center infrastructure, including network interface cards, switches, routers and security appliances. As expected, the Industrial and Automotive segment was up 6% sequentially. That growth rate is tempered as we kept strategically shifting under true demand to normalize channel inventory. As we mentioned on earlier calls, we remain on track to normalize channel inventory by year-end, which sets us up for renewed growth into 2026. Based on customer feedback and design win activity, we're confident that we're gaining share across smart factory, robotics, medical and aerospace and defense applications. Even though macroeconomic and industry conditions still affect us, multiple elements are strengthening our belief in what lies ahead. Lattice's addressable market keeps expanding because infrastructure capital expenditures are rising quickly, we hold a diversified position across the biggest and fastest-growing applications, attach rates are climbing, average selling prices are going up thanks to new products, our small and mid-range FPGA portfolio's application footprint keeps broadening; and lastly, AI usage keeps increasing. Considered together, we think these dynamic forces are growing the dollar content Lattice captures per customer system; we also keep winning with pre-Nexus, Nexus and Avant products. Customers keep picking Lattice instead of our rivals, which our design win growth demonstrates, and those wins are on pace for a record in 2025. New product revenue keeps expanding at a strong rate, and we are on track to surpass the 2025 goal we projected in prior earnings calls. Finally, we estimate AI usage as a percentage across our products will land in the high teens in 2025 and mid-20% range in 2026. To sum up, Q3 was a strong quarter defined by consistent execution and strategic progress; we stay focused on delivering differentiated innovation, deepening customer relationships and driving long-term shareholder value. Our confidence in the outlook has grown, led by our leadership position and capacity to seize the compelling opportunities before us to drive accelerating growth in 2026 and beyond. I'll now hand the call to Lorenzo for a detailed review of our Q3 results and Q4 guidance. Lorenzo?
Thanks, Ford, and good afternoon, everyone. We'll kick things off with a quick look at how we did financially in the third quarter of 2025, and then we'll cover what we expect for the fourth quarter. We're happy to say that Lattice once again came through as expected, with revenue, gross margin, and operating profit all landing right where our outlook said they would. Revenue came in at $133.3 million, which is up 7.6% from the prior quarter and 4.9% higher than the same period a year ago. All told, that's the biggest revenue figure we've posted in 5 quarters, and we anticipate further growth in Q4 as well as in 2026. Communications and computing revenue hit a fresh record, rising 21% year-over-year and 8% sequentially; we look for robust growth in this end market during Q4 and 2026. On a non-GAAP basis, gross margin came to 69.5%, which is an expansion of 20 basis points from the previous quarter and 50 basis points from a year earlier. These results continue to show how durable our business model is and how much value and differentiation our products deliver to customers. Non-GAAP operating expense landed at $53.9 million, matching our guidance. OpEx was unchanged versus a year ago. The roughly 4% sequential rise in operating expense mirrors our strategy of putting money into products, infrastructure, and talent that will further reinforce our leadership position and let us drive accelerating growth. Our confidence in our 2026 revenue expectations has grown. In line with those expectations, we accrued stock-based compensation expense for PRSUs this quarter. That accrual was the main reason our GAAP operating expenses went up. Non-GAAP operating margin expanded 150 basis points to 29%, and EBITDA margin likewise rose 150 basis points to 35.6%. Non-GAAP EPS came in at $0.28, the midpoint of our guidance, marking 17% growth both year-over-year and quarter-over-quarter. GAAP net cash flow from operating activities for the third quarter of 2025 rose to $47.1 million, up from $38.5 million in Q2, while the GAAP operating cash flow margin reached 35.4%, up from 31.1% in Q2. Free cash flow for Q3 was $34 million, with a free cash flow margin of 25.5%, up from $31.3 million and 25.2% in Q2; this remains an area of focus for Lattice, and we expect to keep generating robust free cash flow. We're reaching these free cash flow levels even as we deliberately invest in CapEx to support our product road map and operational improvement projects. We're encouraged by the ongoing improvement in industrial and auto channel inventory; we're still tracking to plan and anticipate inventory normalization by the end of 2025. As we mentioned last quarter, comps and compute channel inventory has already normalized. Let me now move to capital allocation. Our balance sheet is still strong. We carry no debt and have ready access to capital to back both organic and inorganic growth opportunities, and we're well positioned to handle macro uncertainty. Thanks to our balance sheet strength and our business model, returning capital to shareholders continues to be a key piece of our capital allocation strategy. Over the course of the quarter, roughly $15 million in common stock was bought back through the buyback program we already have in place. Across the first 9 months of 2025, we've repurchased common stock totaling about $86 million. On the current authorization, $14 million remains available, and when we meet with our Board of Directors in December, we'll be reviewing the next authorization. Let me now move on to the guidance for Q4. That guidance captures how our business is recovering and lays the groundwork for further growth as we head into 2026. For Q4, revenue should increase, landing somewhere between $138 million and $148 million. Taken at the midpoint, that amounts to 22% growth in revenue compared with Q4 a year ago. That would mark the biggest year-on-year increase in close to 2 years, and it's backed by the best booking patterns we've witnessed in no fewer than 6 quarters. Gross margin should come in at 69.5%, give or take 1%, on a non-GAAP basis. Between $54.5 million and $56.5 million is what we expect for non-GAAP operating expenses. On a non-GAAP basis, the income tax rate for Q4 should land somewhere between 3% and 5%. Non-GAAP EPS should rise to somewhere between $0.30 per share and $0.34 per share. To wrap up, Q3 represented yet another solid quarter for Lattice. Our results came in as our guidance indicated, and with our standing in high-growth markets, we anticipate even more acceleration ahead. We're pursuing operational gains in the near term, putting money into reinforcing our leadership position in small and midrange FPGAs, and getting the company ready for a 2026 that's even stronger. Operator, that wraps up our formal remarks. At this point we can open the call for questions. Operator, Lorenzo Flores here at Lattice. I want to let everyone listening know that in our press release, we discovered a typo connected to the guidance for operating expenses. What I gave in my prepared remarks — between $54.5 million and $56.5 million — is the correct figure, and right after this call we'll fix the press release, which states between $54 million and $55 million. With that said, let's begin our questions. Thank you.
Kevin, thanks for the congratulations on our results. Let me start with your first question. You mentioned that our confidence in the 2026 outlook has grown. Does that increased confidence depend on the inventory normalization we expect in the industrial and auto channels during Q4? Or is it possible that the robust growth in comms and compute that we're witnessing could more than compensate for any possible softness in the macro industrial and automotive areas?
Ford here, thank you, Kevin. From our Q3 results and Q4 guidance, you can observe that revenue growth in our comms and compute business keeps speeding up, and it will speed up even more as we move into 2026. When it comes to our comms and compute business as a share of total revenue, it rose from 35% of total revenue in 2023 to 45% of total revenue in 2024, and it's expected to exceed 55% of total revenue in 2025, and we anticipate it will climb to roughly 60% of revenue as we enter 2026. Given that expansion, this segment is becoming extremely important to how we accelerate revenue growth. Also, during 2025, we've grown server and communications more quickly than the CapEx growth underneath. So compared with 2024, our server business is up 85% year-to-date, and compared with 2024, our communications business is up 63% year-to-date; thus, you can see that both server and comms are expanding more rapidly than the underlying CapEx, and that ought to persist into '26. Therefore, you'd anticipate that the comms and compute business will fuel faster growth in '26. On top of that, just as you pointed out, we remain on track to have industrial automotive inventory normalized by the end of the year. Inventory in our comms and compute is already normalized; and by the end of the year, our total channel inventory is expected to reach the 3.x range, as we previously communicated to you on earlier calls. And because of that, we should likewise see industrial and automotive move to natural demand instead of chip to under natural demand, which is what we're doing throughout 2025; so with comms and compute speeding up and industrial and automotive bouncing back and shipping to natural demand, we anticipate revenue growth in 2026 will be extremely strong.
Great. As a quick follow-up regarding the Industrial and Automotive segment, could you describe what you're observing region by region? Are there any areas performing better than anticipated?
Sure. Thanks, Kevin. China's automotive market, we find, keeps going strong. For our company, automotive represents under 5% of total revenue, yet China is likely the area where it's performing well. Aerospace and defense, we observe, are robust across the globe, and we're capturing a growing portion of design wins there. We're likewise witnessing considerable activity around physical AI, with many design win engagements — these wins involve companion chips for physical AI spanning every segment, from industrial robotics to humanoids, automotive, medical, aerospace, defense, test and measurement, plus all the other areas within industrial and automotive. Thus, we hold a fairly optimistic view about where that segment heads as we move into '26 and bends upward even more going into '27.
Congratulations on the genuine change in how you're coming across. And I suppose my opening question relates to how sure you are. Obviously, you laid out all those positive factors. But what in your view has shifted during the past 90 days that might have boosted your confidence? It seems like that inventory overhang is working itself out. You mentioned something along those lines three months ago. Still, I'm curious whether any demand problems or demand-related developments have occurred that are producing this stronger confidence.
Yeah. No, that's a terrific question. David, I appreciate it; our Open Compute Summit went very well. That event drew most of the hyperscalers, server and OEMs and ODMs, plus the wireline communication equipment OEMs and ODMs connected to that space. And during that event, people there observed that it was as if someone had flipped the switch. And suddenly, over the last 90 days, activity and spend have clearly picked up. As the latest hyperscaler earnings show, the outlook for the coming year is higher, and that's what we're observing. We're also beginning to engage more closely with certain neo-cloud and enterprise vendors, and we see them pushing harder to grow their AI CapEx. A single hyperscaler handed us a fairly large upward forecast for '26, '27, and continuing through 2028; and our role as a companion chip for some of this AI is gaining traction in both the data center and physical AI. Take the cloud data center, for instance: our companion chip is being adopted for CPU across x86 and ARM, AI accelerators from NVIDIA, AMD, Intel, hyperscalers, various accelerator ASICs across networking, across Broadcom, Mirvan, Mellanox and Cisco, security, board management, rack management, cooling, power management as a companion chip, the small and midrange FPGA from Lattice are performing well, and the pace at which we're being adopted keeps rising. We're witnessing the same thing as a companion chip on physical AI around sensors such as camera, LiDARs, radars, along with assorted industrial and automotive uses.
And I believe — and once more, to close, let me echo what appeared in my prepared remarks — the book-to-bill we're seeing is the strongest in roughly 1.5 years, and that business is scheduled to land in the opening half of the coming year. So every bit of the sentiment and excitement Ford just walked through, along with the drivers behind it, is now genuinely turning up in our orders. What we're witnessing, then, isn't merely the mood — it's real business arriving.
Fantastic. That's a lot of useful detail. I really appreciate it. And perhaps regarding the Avant platform specifically, you mentioned observing expansion in every area. I'm just a bit curious whether we could receive an update on Avant, how design wins are shaping up there, and perhaps what the growth projections look like heading into 2026.
Sure, Nexus and Avant are both performing very well, along with a few of the platforms that came before Nexus. Previously, we stated our belief that 2026 would be Nexus's year, while '27 would belong to Avant. We view 2026 as the data center's year, with '27 bringing a significant rebound in industrial and automotive. Nexus appears tied more closely to the data center, whereas Avant connects more to certain midrange applications within industrial and automotive. That's the way we anticipate revenue phasing across the coming couple of years.
Yes.
Appreciate the details on demand tied to AI. And since we're seeing AI demand pick up speed, I suppose we shouldn't overinterpret a slight quarter-over-quarter deceleration in Q3 within communication and computing, given that this line rose 20% sequentially in Q2 and 8% in Q3. So first, should we expect growth in communication computing to speed up again in Q4? And is the Q2 strength possibly tied to that being the first quarter when inventory levels returned to normal? And because of that, you essentially caught up to end demand and are now shipping in line with actual end demand again. Is that the right way to interpret this?
Yes. I'll take some of those points, Tristan. I believe the guidance we provided for Q4 in fact reflects meaningful sequential growth in comms and compute. I think a portion of the non-server business within our comms and compute segment affects the growth rates we measure. But I'd like to point you back to remarks we made in earlier quarters and once more this quarter regarding the true driver behind comms and the strength being our server demand; and that's above 80% growth year-over-year. I think that's a very strong signal of what will drive that. Now it might not remain exactly that high, but the key point is it's expanding faster than hyperscaler CapEx by a substantial margin, and it demonstrates the growing footprint in terms of the design wins we hold, the quantity of chips in design, ASSP design, and we anticipate that will keep growing, as David's question noted, as we move from generation to generation from pre-Nexus to Nexus to Avant. So I think all of that points toward a strong acceleration; so I don't think I'd focus too much on a few decimal points in a quarter-to-quarter change. I think if you consider the overall trajectory over time, it's quite strong.
That's quite helpful. So when do you expect gross margin to start improving from where it sits today? And what's going to drive that? It seems, given the growth you're seeing in computing, that gross margin should expand next year. I realize you aren't providing guidance, but can we reasonably say gross margin expansion is coming next year? Or is something else working against it? I'd also like to understand what you said about '26 being the year of Nexus. In the past, historically, as you're aware, high-end FPGA revenue usually peaks roughly 7 years out. With Nexus, it seems the first few years were a fairly quiet ramp, since you used to give a breakdown. Earlier this year, Nexus was around the low teens as a percent of revenue, and considering it launched in late 2019, that was probably somewhat under what I'd have expected, at least back then. But momentum now appears to be building, and I want to grasp what's driving that. Why is Nexus gaining even more momentum after all those years?
Yes. So I will — we're probably going to break this question apart among us here in the room. And I'd like to begin with the gross margin question, which is genuinely one of my favorites, since it lets me look back at where we actually stand, 69.5%, which is quite good, and reaching that takes a great deal of effort. Your question really concerns how comms and compute affect the total gross margin. I think we also hear the flip side of the question, whether an industrial recovery will help it. I think all of you need to factor in that profitability varies across our end markets and, inside those end markets, inside specific design wins as they ramp. So we steer the overall portfolio to reach the margin, and it could be plus or minus a few basis points over time. But I think we're fairly comfortable with this margin right now, expecting a number of different mix scenarios. So I think we aren't anticipating a very large improvement from any single factor as time passes, but rather largely remaining where we are. On the Nexus ramp, I do think the — I'll begin on this and Ford can add. One thing to remember is that we're only now broadening the product line with Nexus, introducing more products. The phenomenon you mentioned earlier applies more where there's a burst of products, but we're ramping new Nexus products throughout this year and next; so Ford, do you want to add anything about the Nexus ramp?
Yeah, definitely. Tristan, that's a great question. We keep rolling out Nexus products; take a look at what's happened — during Q2 of 2025, we launched a new Nexus built for I/O optimization. Across '25, 4 to 5 additional Nexus SKUs have come out. In '26, we're expecting 5 to 6 more. So sure, the very first launch happened back in 2019, yet we keep bringing new Nexus offerings to market. A Nexus 2 has been introduced, and that one will, once again, begin its rollout. Nexus, then, is an entire family covering a good number of devices. When we refer to Nexus, it's essentially our small FPGA line, and that line is set to keep going for another 10 years. A huge wave of fresh applications keeps appearing; we moved from booting, I/O expansion, buffering, and control into Root of Trust now, plus PQC, plus board management, plus rack management, and with the newest OCP, we even had leak detection. The count of use cases we keep uncovering is enormous; you wouldn't see 80% growth — more than 80% year-on-year in server — unless adoption were broad and still speeding up. So for the next several years, our outlook on Nexus is quite optimistic; once more, picture Nexus as the small FPGA family we talk about. We hold the lead there. We'll keep holding it. We'll keep putting investment in so we keep holding it.
Could you put a number on industrial and automotive — what portion of the undershipping is baked into the Q4 guidance while you burn through that stock? And looking at it from the other direction, are longer lead times or shortages showing up in comms and compute?
Yes. Thanks. We don't provide that exact breakdown, Gary, though we've mentioned before that we ship short by roughly 2 weeks each quarter, so approximately $15 million to $20 million per quarter; and by the end of the year, we expect that to be normalized. So this could be a headwind in '25 that turns into a tailwind in '26. Lead times for comms and compute are getting longer, but we're managing it closely and are highly focused on ensuring our customers receive supply, and so far we've performed very well.
Right, so a few months back — August, to be precise — we put out word that lead times would be extended. At the same time, we were collaborating with the suppliers we use as well as the customers we serve, ensuring shipments lined up properly. Because of that, our order book gives us the visibility needed to continue in that vein. My sense is that those buying from us have felt good about how we've performed on delivery. Extending lead time genuinely helps, in fact, when it comes to planning the business and planning how we load up suppliers. And like I said before, we've now got solid visibility into a really robust pipeline ahead of us across no fewer than the coming 3 quarters, Q4 included.
When you consider the opportunity in data centers or evaluate the served available market, do you rely on some kind of yardstick—such as the dollar potential for each gigawatt in the capital expenditure roadmaps of the various hyperscalers, or perhaps something different, like your content for each rack-scale solution, covering GPU, CPU, and networking all together? Whatever you could provide on that front would be useful.
Yeah. Well, we certainly have that, Gary; we track an attach rate for each server, for each rack, and for various companion chips—whether we're a companion chip alongside a CPU, a GPU, a switch, a NIC, storage, or board management—since the dollars per chip differ greatly across those cases; so we do—our models are extremely detailed. Naturally, we don't make those public. What I can say is that our growth will outpace CapEx. CapEx is climbing rapidly, as you're aware. Yet our growth is quicker since our attach rates are rising more quickly. The number of users adopting Lattice FPGAs for these data center use cases keeps widening. We're shifting toward newer offerings such as Nexus and Avant, which means a greater average selling price for those newer products. We've seen more AI server adoption, which once more pushes up content for Lattice. So on the whole, you can expect that we'll keep expanding faster than the market beneath us.
So Ford, I wanted to revisit what you said regarding data center being 60% of the total next year. I believe earlier — 60%, correct? Sorry, CNC, yes. So I believe you've previously mentioned something like 15% to 20% year-over-year growth, and that within I&A, inventories are getting back to normal by the end of the year. I think you both confirmed that again. My expectation would have been that once that normalization happens for I&A, it would drive a significant amount of growth next year. So I would have figured I&A is expanding here. I mean, that 60% remark you made has 2 components to it. CNC might be significantly higher, yet on the other hand, I&A might be significantly lower. And so I'm attempting to weigh that, to get a sense of what I&A growth might look like next year. You also said that next year is truly the year of CNC and I&A is 2027; so trying to read the tea leaves, especially for I&A. Do you think it's a much flatter year than maybe we do?
So let me begin from the start and at a high level, and you remember in my remarks, I said our confidence is building when it comes to our growth in 2026. And as a result, I needed to make an accrual for certain performance-based, revenue-based stock compensation; that threshold was 20% into 2026 from where we anticipate finishing the year today at the midpoint of our guidance. So that's the situation. And Ford, would you like to add the color regarding the end markets? And go ahead, or I can do it.
Right. No. So looking at it broadly, Chris, next year we anticipate I&A returning to growth. Our expectation is somewhere around the mid-single digit up to, let's say, 15% range. Comms and compute, though, might land more in the 20% to 40% range. That's the way to think about those 2 segments.
Okay, so if we believed we were heavily reliant on substantial expansion in industrial and auto, then we wouldn't — we wouldn't feel sure about the growth. That area, as you're aware, is more exposed to what's happening in the broader economy, and we're still dealing with inventory as the year wraps up. So as we look ahead to 2026, we chose not to set our expectations based on a rate higher than mid-single-digit growth, as Ford mentioned regarding industrial and auto.
Great. Taking a step back, Ford, when you think about your history in AI, plus networking, and what Lattice might do next — possibly evolving beyond being only an FPGA company. Previously we've discussed embedding hardened ARM cores within an FPGA; however, what about embedding hardened MAC cores or an NPU within an FPGA, and/or developing some sort of ASIC capability that steers certain customers from FPGA toward a hardened ASIC? Or whatever other ideas you have regarding this company's next moves and what you're investigating or considering, obviously without naming specifics?
Yeah. Look, we're really thrilled about the future destination for Lattice; regarding that hardened processing core, so to speak, our chosen approach has been partnering alongside microcontroller and microprocessor firms. And we've achieved rather solid results so far. At our latest customer gathering in Asia, NXP Semiconductor joined us there. We collaborated at OCP, and honestly the alliance with NXP is growing pretty robust, with customers really favoring the two of us as a pair. We're broadening that across additional companies. So when it comes to the microcontroller, interest is coming from other participants in that arena to follow the same path. And you'll learn about additional partnerships down the road. Thus we genuinely believe that on the processing front, be it microprocessor or microcontroller, you'll witness 1 plus 1 equal 3, since we'll deliver combined offerings to customers that draw the greatest value from FPGA alongside those processors beside us. As for our longer-term direction, there are certainly numerous conversations happening along with exploration, yet we aren't prepared to talk about it right now. Customers are requesting that we handle particular things we're putting money into, and that could form part of the higher investment in 2026 that Lorenzo alluded to during his prepared remarks. So keep watching for that. More specifics will come as we move forward.
Ford, I was hoping you might expand a bit on what was said regarding AI usage. My recollection is that you had mentioned high teens moving toward the mid-20s for the coming year. How are you viewing the mix there when it comes to compute and comms as opposed to industrial and automotive? And some additional clarity on what counts as AI usage would be really helpful.
Thanks, Ruben. This year, 2025, we remain on pace for the mid-teens, then progressing to the mid-20s in 2026; and the split is roughly 60% comms and compute against 40% industrial and automotive. Within comms and compute, we serve as a companion chip alongside the different ASICs and ASSP that I referenced earlier. In industrial and auto, our role can be far edge AI close to sensor intelligence, so to speak; and that is where we're focused, under 1 tops, those highly power-constrained use cases genuinely linked to camera, LiDAR, radar, and other industrial sensors. I trust that addresses your question.
Got it, that helps. Now, to follow up on what you said about Nexus — next year being the year for Nexus, and then Avant in '27; you noted that Nexus is more tied to data center while Avant leans toward industrial and auto. So I'm curious why you don't expect to see some mid-range FPGA use in those higher-end comm and compute applications that are emerging. Am I misreading this, or hearing you incorrectly, or is there anything you'd add there?
No. What I mean is, Avant applications would show up in comms and compute. Nexus applications would show up in industrial and auto. So my intent wasn't to suggest exclusion. What I meant was that a majority, so to speak, of the design win will be Nexus on comms and compute and Avant on industrial and auto. For instance, we had Ericsson at our developer conference, and they were demonstrating an Avant application. So once again, Avant applications in comms and compute will definitely exist.
I'd like to dig into a few of the assumptions underlying the growth rates you just referenced for 2026, or rather the revenue split. Can you share any metrics or specifics that would explain how you might land at the bottom versus the top of that 20% to 40% range, since it's quite wide, and perhaps some color on how much general purpose versus AI servers are expected to contribute? Then, on the industrial and auto front, does that basically assume we're just shipping to normal seasonality at 5% to 15%, with no restocking? I'm also trying to get a handle on what normalized sell-through looks like for auto and industrial right now, given all the volatility.
Josh, I'm going to take this somewhat in reverse order. On industrial and auto, as I mentioned, for us to feel confident — considering what's going on in the world and what might unfold — we figured it was wise, as you pointed out, to simply say we're returning to a normal — I'm not sure seasonal is the correct term — but a normal demand cycle, where we start shipping so that consumption and our revenue line up, instead of undershipping demand the way we've been doing. I don't think that's aggressive in any way. Turning to comms and compute, we're fairly confident about that $20 million to $40 million range Ford mentioned earlier; we have solid backing for it through the fundamental driver, which is industry CapEx. That range matches what we observe for industry CapEx. And as Ford noted earlier, there could be upside as our footprint expands relative to that CapEx, thanks to the design wins we've secured and the opportunities we're starting to ship to now — so hopefully that addresses your question. Regarding AI mix — and feel free to chime in and set me straight if I've got this wrong — AI mix will keep increasing, just as AI's prevalence across the overall footprint in both comms and compute and industrial keeps increasing too. We haven't reported that separately. Does that help?
Yes, that's right. Thanks for all that detail. For my follow-up, you mentioned several applications for companion chips and use cases across both general purpose and AI servers. Could you say which ones are getting the most traction right now, and which do you think will grow further as the Nexus and Avant product portfolio layers in?
Yes. Josh, I don't think we're ready to split it out at that level of granularity by application. As Lorenzo was saying, what makes us confident is that if overall CapEx is expected to rise 20% to 30% next year, we believe we'll grow faster than that underlying CapEx. That's why we're citing those numbers. Breaking this down by application, I think, goes beyond this call. We can talk about it on post calls, if you'd like.
Can you hear me now?
Yes.
Okay, I apologize — I'm dealing with a few telephone problems. Perhaps my connection could use additional FPGAs. I think — apologies if this came up in a prior question, but there's been a lot of discussion from you folks regarding PQC; and I remain somewhat unsure about the potential size of that market. So might you explain, like, what — what scale does that opportunity represent for your company? Is expansion happening? What observations can you share? Any additional details regarding the PQC angle?
Well, PQC serves as a significant catalyst behind the uptake of our security offerings. Our belief is that we hold a lead; mandates now require it across every system; consequently, we're currently being incorporated by design into applications of both the comms and compute variety. And we have not separated it out, Melissa. So once more, my sense is we likely won't be prepared on this call to provide a breakdown of the percentage of applications by use case.
That makes sense. Then when it comes to pricing, I just want to confirm that, with the semi cycle sort of hitting its low point, you're noticing anything noteworthy regarding pricing — whether from your industrial and automotive clients, or perhaps on the content compute side, where there might be somewhat more leverage? Anything at all about pricing?
No; clearly, our approach to pricing is to price according to value, which our gross margins make evident; and like I mentioned before in the commentary regarding gross margin, that shifts considerably across segments and, inside segments, across the various design wins, depending on the way customers will use it. We've also consistently followed a strategy of collaborating with our customers to deliver a long-term value proposition. So we aren't opportunistically exploiting particular demand trends, since we believe we'll be working with these customers for a long time and, over time, we'll capture the complete value of what our product set offers them. So at the same time, frankly, we do see pressure from our suppliers driven by demand, and we're handling that on both sides of the supply chain — with them and with our customers — to reduce the effects of it.
Let me go back to the auto industrial segment; throughout the year that operation has consistently generated roughly $50 million each quarter. You've mentioned as well that each quarter you've been shipping $15 million to $20 million less than demand. So for the past 4 to 5 quarters, consumption appears to have held steady at $65 million to $70 million. Honestly, it's hard for me to understand why that would fall to around $55 million, which is what your 5% to 15% guidance for 2026 suggests — that implies a huge shift in consumption level.
Yes. I'm — I believe the difficulty in matching these up — what you're describing versus what we observe is that a portion of our industrial and auto business is direct, meaning it isn't channel-related; and the undershipping demand phenomenon we're discussing shows up mainly within the channel. Still, I don't want to commit to that magnitude of an increase for 2026 until we truly witness the strength supporting the whole industrial and auto business.
Regarding that backlog coverage — apologies, please continue, Ford.
No, go on. So what figure are you referring to at that point for...
Quinn, he's putting out figures like 52%, 47%, 50% for industrial auto across the year. Did I get that right?
Yeah, regarding industrial and auto — Lorenzo A.
Flores: That figure's exactly right.
Well, in terms of responding to what you asked — regarding your figure for — actually, perhaps it makes more sense for us to work through that reconciliation once the call has ended? Quinn?
Yeah. Yeah, we can discuss that separately. My other question was just about comms and compute, folks — you mentioned servers rising more than 80% since the start of the year, and I believe comms climbing above 60% year-to-date. But when I check total comms and compute for '25, even assuming a solid sequential gain in December, it sort of appears you'll end up in the mid-20s for the year. So it seems some piece of comms and compute must be declining quite a bit to drag the whole bucket down to mid-20s, given that the 2 largest pieces of comms and compute — servers and wired comms — are expanding 3 times quicker. So what's canceling that out?
Yes, yes. Yes, what's canceling that out is the client business; we enjoyed a substantial client business a year ago that has essentially vanished at this point. So we've faced a significant headwind in the client business, Quinn, involving 3 client OEMs — large 1 and 2 medium-sized ones. And that headwind is now gone.
This is Chris Myers, filling in for Kevin Cassidy. To start off, could you all give us your take as of now on what's happening in the automotive market? Are inventories beginning to shrink, or is any backlog starting to accumulate once more? Then, regarding the data center area, what pattern are you observing when it comes to your dollar content for each traditional server as opposed to AI servers? Also, do these devices serve differing functions across the 2 kinds of systems?
Yes; regarding AI servers compared with traditional servers, we certainly carry more content in AI servers, given how aggregated these servers are. So we — so yes; and then as for the...
Regarding autos, let me start by reminding everyone that this segment represents just a tiny slice of what we do. And looking worldwide, we aren't seeing any favorable winds at this point, with the one exception being China's automotive sector, which appears to be accelerating more quickly than anywhere else. Still, that market makes up only a share of our total business, and the whole auto and industrial piece is itself minor within the company overall. That's the picture from where we sit. Outside of China, we see no momentum in autos.
Yes. I appreciate all of you being here today. During this quarter, we will take part in these investor events: the Stifel 2025 Midwest One-on-On Conference, held in Chicago on November 6; The AllianceBernstein 2025 Buy-Side Small and Mid-Cap Summit, held in New York City on November 18; and finally, the UBS Global Technology and AI Conference, held on December 2 in Scottsdale. That brings our call to a close. Many thanks for taking part, and enjoy your evening.
Appreciate it. Everyone, that wraps up the call for Lattice Semiconductor. Thanks for joining us. You're free to hang up now.
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.