Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript & Summary

All right. Welcome back, everybody. I'm Joe Moore, Morgan Stanley semiconductor analyst. Happy to have with us today the management team of Lattice Semiconductor. Lorenzo Flores, CFO; and Esam Elashmawi, Chief Strategy and Marketing Officer. So guys, welcome. Thanks for coming.

Good to have us here. Thank you.

I don't cover the stock, so if I don't ask you the questions I should ask, just answer what I should have asked. I understand. But like the main thing is for you guys to tell what you want to talk about. I guess it's been 6 months since Ford joined Lattice, so what has surprised you guys? What are you most excited about with the opportunity to have in front of you?

You have the best contract.

Yes, Ford joining 6 months, I'll talk about what excites us is the -- really the opportunity ahead for Lattice. Lattice is -- has a leadership in small FPGAs, and we've just recently introduced our midrange FPGA, which doubles our market opportunity. We positioned the company in our core strategic markets of comms compute, industrial auto, which now accounts for 90% of our revenue. Those end markets have long-term circular growth drivers associated with them. We've strengthened our product road map. Our [ Proto Fam ] is the strongest in the customer -- in the company's history. This is the strongest roadmap that we've had, and we're pretty excited about what's to come as well. And the road map is very differentiated from the competitive devices that are out there. If you look at Lattice FPGAs what we bring to the market is differentiation around low power, and our small FPGAs could be 4x lower power, midrange 2.5x lower power. Our small form factor, which is really important to a lot of our customers. We have our form factor being anywhere from 6x to 10x smaller than competitive devices. So that's significant as far as miniaturizing the systems. And with the power differentiation small form factor, we don't sacrifice performance. You get double the bandwidth in our midrange FPGAs, twice the performance as well in some of our small FPGAs. And our customer intimacy has never been stronger than what it is today. We've got over 11,000 customers. Customer intimacy is very strong. We've invested quite a bit on our software and our software solution stacks. This makes it easier for our customers to get to market as well. And so when we put all of that together, I think we've demonstrated over the past several years, good growth and market share gains in the FPGA market and forward looking, we just see better times ahead of us.

Yes. Okay. Great. A bit of a tricky environment that we're in now. You've taken quick action on the cost structure. Is that due to the cyclical slump, more structural reasons? And you can just talk about generally the strategy behind that?

[Indiscernible]

[Indiscernible] I think, yes, the entire industry has gone through a cyclical slump the way you phrased it. Lattice has taken action. The action we took, we announced in Q4. We took a Q4 action, which was a 14% reduction in headcount and OpEx. And a lot of the headcount reduction was really strategically around repurposing where our headcount is. We've repositioned some of our key RECs in lower-cost geos so that we can leverage the efficiencies we can get out of that as well. So we did it in such a way that did not impact our product road map. We were very conscientious. We've got strong conviction of our organic growth in what we're doing. So we did in such a way that doesn't impact our road map. And so that's the actions that we took. And as far as the company and going into 2025, we've shared at our last earnings call that we expect this year that -- although we're in this cyclical slump, we're still shipping under true demand and that, that will continue through the first half of this year, and it will get back to normalized demand revenue by the end of this year.

Because -- I'm not really avoiding the question, I just think Esam has better historical context. But what he described as how we repositioned the resources will allow us to continue to invest but get more leverage out of the business model as the revenue returns back to more historical expectations of growth.

Okay. And you mentioned the strategy to move in low-cost regions. How much leverage can you came from that?

So it's really around adding additional resources because you can imagine the cost difference between a high cost region and low cost region, that allows us to actually accelerate some of what we're trying to do as well.

Okay. So in addition to Ford, you've announced a number of new executive hires, including Lorenzo, who I've known from FPGA industries in the past, among other things. Lorenzo, maybe you could talk a little bit about what attracted you to Lattice?

Yes. I had this question a couple of times since I started. And I will start maybe in an unexpected place, which is when it first came by me, my reaction was, I'm not interested. And for a couple of reasons. One is that if I looked at Lattice from the outside in the years that Jim and Sherry were here, I looked at it and said, there's not much left to do. They've really done a great job of positioning the company and repairing it, if you will, from my history in FPGAs, they were really not very well run and not executing really well. And so they fixed a lot. And it came down to -- I met with Ford a couple of times. I got to know him. I got his view on how Lattice and its focus on execution and its current road map put it in a great position to where we can add a different thought process, a different way of looking at things to really accelerate Lattice. As you said, I've been in FPGA business before. I know the benefit of focus. I know the benefit of investing around in our ecosystem and how to use the assets we have to drive growth. It takes a little bit longer than some other semiconductors, but think the platform is there. And then the last step is I met the team, Esam included, and I said, this is a really fantastic group of people to work with. And I've so far founded a really enjoyable place to work. So a great business platform, exposure to all the trends in the industries from the FPGA perspective, great leadership team. So I got very excited about coming aboard.

That's great. And then Esam, maybe you could talk more broadly about the new hires and kind of what makes them a good fit and what's maybe shifted. Because as Lorenzo said, this company was really well run before Ford, who I've known a long time. It probably has a different style bias and a different end-target bias. But like just what is the general trend of kind of the executive team being [indiscernible]. . .

Yes. So let me start off with Ford. Ford has been with us now for 6 months, really good industry veteran. A lot of success that he's had in the past, and he brings to Lattice just a wealth of industry knowledge as well as energy and he's really got us to really focus on what matters. I mean he's all about let's make sure we focus, let's make sure we execute. And in any company, the more focus, the better you are in execution, the faster you run, and I think Ford brings that to the organization. And on top of that, a real good human as well, good individual. So it's a pleasure to have him on board leading Lattice. And with the other changes we've done, clearly, Lorenzo joined us as CFO. We have Nicole Singer, who joined us. She was -- as our Chief People Officer, and she had prior experience with Sifi, Synaptics and Xilinx as well. And then from internal, we've had two key promotions, Erhan Sheikh, who is now our Vice President of Worldwide Sales and Tonya Stevens, who wass Interim CFO, is now our Chief Accounting Officer. And I think that's a reflection also of the deep talent we've had within the organization. I think when you combine the outside skill sets that we brought in with Lorenzo and Ford plus the promotions we've done internally and Nicole is well from outside. I think we're just set up to be a very strong team to lead Lattice into the next era.

Great. So maybe shifting to products a little bit. A lot of focus on AI these days. Can you talk about the specific AI use cases where Lattice FPGAs can make sense?

Yes. Actually, good question. So when we think about AI-related revenue and use cases, it actually falls into three buckets for us. The one that's most obvious to everyone is general -- specific servers for AI. So there's general purpose servers, but then there's AI servers, AI-specific servers. So we look at that as one area. And we enable a lot of that. In fact, if you take a look at the entire rack, and this was mentioned at OCP, and there were some of our customers there, Super Micro had a rack. Lattice who's had a rack and somebody actually did a write-up and there was no less than at least 50-plus FPGAs in a rack. So what do we do in these AI servers? Well, we do board management, we do security opportunities. We've got interface, power management. So we enable that. And we're not just on the CPU or motherboard cards or GPUs, but we have stuff on in [indiscernible] card, storage cards, et cetera. So we're proliferated across in entire rack. So that's one area where we see growth, and we continue to drive a higher attached rate. So as that market grows, Lattice also opportunities grow and revenue grows. The second area for AI-related revenue for us is any time somebody wants to do AI on the edge, well, that's a good opportunity for Lattice as well. And we are in that data path. So for -- as an example, at Lattice's latest developers conference that we had in the past, we had NVIDIA on stage talking about how we engage with their Holloscan or Bajorin project. Where we actually do the sensor aggregation, we do the co-processing, we do some NVIDIA secret sauce that they've shared with us in our FPGA and then we send that to their GPU for edge-type applications. And why are we there? Why are we integral, and why do we work with NVIDIA? It's because these GPUs were never architected and designed to work with all the billions of sensors that are out there. How do you interface? Well, you need something to aggregate that data and Lattice does that very well with our small form factor, low power. And we also do the same on ADAS type applications. In fact, Mazda actually sent out a press release talking about, in their crossover models, how Lattice enables our ADAS systems, it's the same methodology. We're doing aggregation, preprocessing and we send it to a compute. So we do a lot of that. So we're in the data path of these AI applications. That's the second bucket. Now the third bucket is we're actually doing the AI on the Lattice device itself. The inferencing actually done on the Lattice device. And examples of that could be client devices. If you buy a Dell XPS or Latitude model or a Lenovo ThinkPad. You may have a Lattice device there doing AI, artificial intelligence. And the type of artificial intelligence we do is it will detect the user. Is it the right user, controls the screen will turn on. We do shoulder surfing, if somebody is standing behind you. Big deal that they're standing behind your looking at the screen, we use our AI technology to tell you they're looking at you. Then we add stuff on wellness, so your posture. We can do things around collaboration, make sure you're in the middle of the screen. There's examples that we've demonstrated on industrial panels. So let's say you're an operator in the factory and you want to control a panel, we actually run on the Lattice FPGA artificial intelligence. Number one is that an authorized trained user to control that piece of equipment. And then if they're selecting a button on the equipment, we actually use Gaze technology that we have, are they looking at where they -- what they're selecting. So if an operator is distracted and you're just hitting a button, talking to someone, you won't take it. They actually have to be looking at the button. And we've demonstrated with our Gaze technology that we can move across the screen, and you can actually see what the person is looking at. And then there's applications that Lattice is doing actual AI on food sorting and object detection. You have the defective pepper in China that's going through a factory. We'll tell you that's a defective piece and help clean itself. So three areas, again, that Lattice has AI-related revenue. Any AI servers that are going out there. We're in the data path of Edge AI, high-performance applications like the NVIDIA one. And we're also doing inferencing and AI on Lattice FPGAs.

And then there's some irony, I guess that the other two FPGA companies were acquired for compute, the eastern part, probably unfair to [ MDNA ] a little bit. I think it was oriented around compute and was acquired because it's a good business. But in the case of Voltero, but specifically for compute, and you guys are doing so well in compute. And I guess why is that? What is the focus that you have that has enabled you to win those sockets? When, if anything, your two competitors are too focused on those same sockets?

Yes. I think we're addressing two different types of applications into compute. When you think about Lattice, we have these small, power-efficient and mid-sized power-efficient FPGAs, we're doing control functions. We're doing board management, power management and our compute -- if you look at a client, for example, some of the artificial intelligence there, the target markets that we're not focused on is the high-performance compute acceleration in a data center. We're not doing that. And that's not been a focus of ours. We're really after what I described earlier. . .

Because nobody's doing that.

Good point.

It probably matters to different accounts more or less. But we don't -- we're not competitive with NVIDIA.

Right.

Right, right.

We don't have any -- we are a perfect companion to them in the platform. So we don't have anything that we're trying to take away from them.

Any idea of doing like a board management controller on an FPGA in a general-purpose server, that's a standard part, generally. Is that just because it's moving so quickly that you need the programmability to adapt to the circumstances changing that quickly? And is there a risk on the road that it does turn into a standard part?

Yes. So if you go back, 2019, we had shared publicly that our attach rate on general purpose servers is 25%. And you fast forward to now, our attach rate is well over 1%. And it's not just board management and control, it's also security applications that we're doing in the servers. And what we're also seeing is that there's a disaggregation in the server market, where they're taking what used to be a single motherboard that had everything and now they're breaking it up into sub boards. So you'll have a board specifically on security and control. You have a board that has the CPU. You may have now a separate storage with memory. And so it's being disaggregated. Every time they disaggregate, that's another opportunity for Lattice. And what we're doing in these -- like in the board management, is we're actually innovating with the industry, and we're developing with them new interface standards, such as LTPI is the most recent that we talked about at the OCP conference, partnering with other hyperscalers and OEMs. And that requires -- there's not enough time to go spin in ASIC for that. These are new standards coming in, you want to get them out as fast as possible. The other advantage, also, why programmability is that we are CPU agnostic. It doesn't matter what CPU they want to design in. They can pick an Intel CPU, an AMD, another arm CPU. It doesn't matter. We're on the board doing the same function, and we can communicate an interface with whatever CPU that they're deciding on. And that helps from a [indiscernible] materials. So we're CPU agnostic, and that helps from the programmability aspect of it as well.

Great. So AI, both in compute and at the edge, is obviously a huge topic. But can you talk about any other kind of end markets? Any trends that we should be particularly enthused about for Lattice?

Yes. I would say one other one that I'm enthused is about, and I know the team is as well, is around not just AGI but beyond AGI, is security. I think security is an area where there's just a lot of focus today. And there are more bad actors out there and there's a need to cure our systems. And we talked about securing our servers. That's really where it starts. Think about it. If you're going to secure a server, it's critical because that's our data centers, it's our financial systems. If anybody wants to hack into an enterprise, you're typically going through a server to get in there. So we work with the hyperscalers, the OEMs, who work with the community, and we develop next-generation security, which is very important. And what we're working on now, which excites me even more, is really -- and we've demonstrated this at our last developers conference with partners. If you go back to Developers Conference, we had Meta on stage talking about it. This last developer's conference, we had Microsoft talking about it as well. And that has to do with the post-quantum era that's coming. You may not realize this, and I don't want to be a downer on your day, but the adversaries are actually stealing your data today. Everybody is stealing your data right now. They're collecting the data. And what they're waiting for is when quantum computing comes, they're going to break all those keys. And what we used to say, everything is protected with AES, encryption, and it takes years to break. With quantum computing, it's just a matter of hours or days where they can break those keys. So the industry is getting ready for that next era. And to do that, what NIST has done, working with the industry and Lattice participates in this with our partners and ecosystem, what are the right cryptal algorithms we're going to need to secure our future. And it's not just about servers, think about all critical infrastructure that's out there. Think about the communication networks that are out there and how are you going to secure that. Industrial factories, all that has to be secured. So we're actually leading right now in providing those algorithms on our FPGAs. We had Ericsson on stage at our last developers conference talking about post-quantum cryptosecurity, what's needed and why Lattice was selected just to secure the future wireless communication systems that Ericsson is looking at. But it's not just Ericsson. I said Microsoft was on stage as well. And so that's an area that excites me because this is a critical problem for us. I don't know how people realize it. This is going to be really important. We're going to have to update our systems and put in higher levels of security when quantum computing is available. And people [indiscernible], when is it available? Well, some will say 2030, some may say it's beyond 2030, maybe it's 2033. I've heard, 2035. But the key message is it's coming. Nobody is debating if it's going to come or not, you may debate at when it's going to come. So the NIST standards out there, the government requirements out there that people have to be ready for that. And you want to be ready for that, not when it comes, you want to be ready for it before it comes. So there's a big push right now in next-generation servers, and we're working very closely with hyperscalers and OEMs and the industry on putting that in. But it's not just servers, it's communication networking and infrastructure. It's industrial equipment as well. And I think that's a really good opportunity. As we've demonstrated a technical and industry leadership as a trusted adviser for our companies -- for our customers on security. I think this is another area that we see growth potential in the future.

Great. I wonder if you could address the competitive environment of Altera and Intel both -- Altera and Xilinx both being part of bigger companies would probably seem to give you guys more focus. But how do you stand on building blocks, software and go-to-market versus those two competitors?

Yes. So let's talk about what Lattice is doing that's unique. Number one, it always starts off with a really strong road map. And our road map is focused on small- and mid-range FPGAs. We build an architecture that's tuned purpose built for small and midrange. And we do that with something we call a 4 LUT architecture. So it's more efficient, it's power efficient, it's smaller size, gives you all the characteristics that I talked about earlier. If you look at other FPGA companies that you referred to, they tend to build 6 LUT architectures. It's public information, 6 LUT architectures, and they're able to build much larger FPGAs with that architecture because they're targeting much higher densities, much higher performance. So think of the FPGA market. We have small FPGAs, midsize and large. We're building purpose-built small and mid, and they'll tell you publicly, they're building these large FPGAs. They have the large FPGA architecture. How do we address small and mid? Purpose-built four LUT. How do they address small and mid? It's a watered-down philosophy where they take these large architectures and shop them down to small and mid. It's not the same efficiency. There's a difference there. So a product road map that's very differentiated, that brings compelling value to your customers. It starts off with that. Second, it's the customer intimacy. It means that we go to our customers, we talk with our customers, and they know they have a road map tuned for their needs. We don't go in EOL devices. We're very moderate on what we do on our pricing agreements with them and how we manage pricing. We make sure they're supported. We invest a lot in our software tools, and we invest a lot in our software tools to make it easier for customers to migrate the Lattice or to get their products to market faster. We're building prebuilt applications in what we call our software solution stacks and we've got six of these now that are out in the market. And these contain prebuilt applications for different types of vertical or horizontal type applications. So we have one called SenseAI, which is about AI type applications, prebuilt, you want facial recognition, object detection, key phrase. We put a [indiscernible] in it and it makes it easier to get to market with that. You want it for security. We have Century. You want something around automation in factories. We've got Lattice Automate as an example. So investing in our software tools has also been very important for us as we compete with the competition as well. So it's road map, customer intimacy and making it easier to get to market are things that we focus on.

Very helpful. So maybe we could talk about numbers. For Q1, your outlook is better than the larger peers. I think where are we in the cycle? Do you feel like we passed the low point? And what gives you confidence going forward?

So first, I think a lot of the groundwork Esam said -- points out, we're a FPGA company is not exactly the same as all the other device companies out there, right? That's one. Two, this precedes me, of course, and it relates to the customer intimacy point Esam made, which is we were very focused on being supply-chain friendly during the supply-chain crisis and distribution friendly during the crisis. And so we, in short, have less of the inventory overhang at both the end customers and in our supply chain partners and have a lot better understanding with them on where things are going. So we don't have the hangover effect as severe as some of the other industry players. So what we're seeing right now is a draw down in the channel inventory as we are trying to get that back to more historic normal levels. And I think we will get there faster than some of the other people who are talking about challenges in the industry. So by midyear, we should get closer to where consumption is equal to sales in and with that, we'll see some turnaround towards the end of the year. We're not being very specific right now, Joe, because there are still uncertainties out there, as you know, from macro on down. But within the next quarter or so, we should have much better visibility in the second half. As that dynamic of consumption to inventory plays out, we should set those platforms for growth with that natural rebound as well as the share gains that we will see and -- from our new products and our existing products.

Okay. Great. and then gross margin, you've done a pretty good job of managing that in a tougher period. Can you talk about the path to your target levels of margin in it?

Yes. So we are close to the target. And I think in the past, we've published 70%, low 70%'s. There's a lot of work, as I said, done by Jim and Sherry and the team that was at Lattice. So I include the managing of costs very, very rigorously but also being very smart about pricing. There are very high-value applications that we sell into and very good about capturing the value. So we're going to continue to manage the prices appropriately, give customers great value, but actually return that to us and be extremely aggressive on managing costs from design for cost all the way through supply chain costs. That's just -- so it's just -- it's a blocking and tackling game and we'll do this all day every day.

And there's three things that drive it, as Lorenzo said, one is our pricing optimization methodology that we've had in place for quite a while. The other is the cost reduction that Lorenzo alluded to as well, we're focused on getting cost reduction into our systems. And then the third one is the mix. And if you look at the mix of our products, we used to have, at one point, 60% of our revenue was industrial. Everybody knows that, that's been soft. So industrial is now 40% -- mid-40% of our revenue, right now is industrial. As that comes back, that tends to be at a higher margin as well. So that will actually help us from a margin perspective. And closed the last -- I think our forecast for this quarter is 69%, I think is what we said? 69%.

Yeah, that's right. Close, yeah.

So we're close to the low 70%s. And as the mix -- as the cost benefits come in as we can do pricing optimization, that will take us to our target.

And I guess back to the gross margin in a competitive context, when you had two public companies that were the significant majority of revenue, they both report gross margins every quarter mark-to-market enforces a lot of discipline. And like when you're part of a bigger company, it has to be at least tempting that when something gets converted to ASIC to chase it down and go down market and get lower margin. Do you think you still have the same pricing discipline across the industry? And obviously, you guys are able to do very well in that context. But would it help to have more public companies that are doing the stand-alone?

I think so. I think the industry as a whole, if you look at the trends of the margin of the industry, I think the industry has had some good appreciation on gross margins over the last decade. And I think that the [indiscernible] semiconductor industry has been able to capitalize and get more value from what we sell as an industry as a whole. And I know we went through the supply chain, and that went up even more because of demand. And I think some companies as revenue drops, there's pressure on gross margin and people are realizing how resilient their pricing was or their manufacturing was and some companies are impacted more than others. But I think if you come back now to a Lattice perspective, our gross margins have been resilient through this process. And that's an indication of the value we have in the market. We closed '23 at 69% -- or 70% was our 2023 gross margin from an annual basis, and we've been hovering around 69%. So it hasn't really been impacted that much, and that's a reflection of the company's value in the market and how companies really -- our customers really appreciate what we do for them.

In the industry, there's a lot of impact on margins based on end market mix. And one of the things that Lattice has done historically is move away from consumer, which is relatively lower it's an opportunistic business, I think everybody would say, it's a relatively lower margin. And going into a focus on industrial be generally beneficial to margins.

Yes. Okay. I have one more question and then I'll open to the audience. The role of M&A going forward, can you just talk to the importance of that? Obviously, Ford and Inphi was very focused on building scale. Is that still a sense here that you want to do that? Or just what's the role of inorganic growth for you guys?

Why don't you start?

Yes. So I think -- so I love what Ford is doing as far as the questions and getting us engaged and how important we should be looking at inorganic, but the fundamental that we start off with is always our conviction on our organic business. We have a lot -- very strong conviction of what we're doing organically. We know the market share of our gains we're growing. We know the revenue profile that we could achieve. We know the operating margin cash flow that we can achieve. So we have really strong conviction over our organic business. So that is always going to be number one. So when you think about it from an inorganic perspective, what we look at is things that could help accelerate what we're doing inorganically or things -- or we would avoid things that would interrupt what we're doing organically. We don't want to do something and it diverts away from all the goodness that we're doing organically. But Ford has a good history of successfully doing M&A at Inphi. He did that when he was at Broadcom. And he's bringing that culture and challenging us of how do we look at M&A in ways that we did not think about to help accelerate that organic strategy that we have and #1 priority always is to ensure that whatever we do brings in good shareholder value as well.

There's more in the spectrum of inorganic activity than just M&A, right? You wouldn't spend -- you can get a good return by investing in the ecosystem and accelerating your organic growth through that type of activity. So that's an area we're looking at much more rigorously now than we've had in the past. I know for -- my own personal experience, that's a great way to help the business grow faster. But we -- as Esam said, we look at pretty much everything. It's what we're paid to do. A very sharp lens that filters things that will distract our business, things that won't create value where we could add exceptional value to whatever we do to drive the company to the next level.

Right. Appreciate that. Any questions from the audience?

There's one over there.

[indiscernible]

[indiscernible] 80% of your volume is through distribution channels. And maybe if you can tie that back to your comment about [indiscernible] being advantage in also question number 2 is about Nexus and Avant. I think that was 15% of revenue last year, if I recall it correctly. So is that going to be the majority of revenue in a few years? Or help us understand how that may unfold?

Yes. So [indiscernible] good question. On the channel strategy, you mentioned the data point. 80% of our revenue going through channel. That does not mean that 80% of our design opportunity and design win and activity is going through channel. That is a fulfillment number. So the way we operate with channel-- number one channel is very important for us, and we've got really great channel partners, and we've been strengthening that channel partner relationship over the past several years. And they're starting to invest more and more every year on dedicated resources for Lattice, which is really good. A very good relationship with them. But the way we look at the market and the customer base, we have over 11,000 customers. So we can't touch them all. So we focus on the market makers. So there's about 200-plus companies that really account for the majority of the revenue. We focus on those, which means a badged Lattice employee will be knocking on those doors. They're measured on that. We have direct contact with them. The rest of it, the broad base, that's really the channel to help us out or if there's a select customer or something that has a very unique location that's hard to get to, we may partner with the channel partner for that. So we work with the channel partners. They're very important strategic to what we do, very good relationship. Badged employees are really -- think of that with the market makers. The broad base is the channel. The 80% represents fulfillment revenue plus some of the demand creation from distributors. It's not all the demand creation. So that's from a channel strategy perspective. And then the second part of your question was around Nexus and Avant revenue, 15% of our revenue this year. When does that become more material revenue? I think I'm re-wording it differently. But we're really excited about the new products. Nexus and Avant are very differentiated products in the market. And if you look at it, what are the proof points that indicate that? Number one, we've been gaining share because of that. And a proof point of that, if you look at our new product revenue last year -- last year was a down year for the industry. It was a down year for FPGA companies. We weathered it much better than others. But our new products, which include the Nexus and Avant, that's predominantly Nexus devices, actually grew double digits in 2024. And that says something. That says something about new products driving new revenue streams being adopted by customers. And so what we'd expect. We'd expect those revenues as we're introducing more Nexus devices. We're introducing more Avant devices that, that revenue will continue to become more material as time passes. So whatever revenue we had in '24, it grew double digits from '23. We expect that to continue '25, '26, '27 and over time, it becomes more material to Lattice revenue. And that's the model of the FPGA business. We still sell quite a bit of devices that you would call pre-Nexus before we introduced Nexus. And those will be in industrial equipment, in automotive and applications that could grow for 15, 20 years, and we still sell those devices. And so what you're doing is you introduce new products that are very differentiated and that layers on top of that revenue then you get another layer, then you get another layer of revenue on top of that as well.

Great. Well, we are out of time. So maybe we can follow up after. Thank you so much. Appreciate it.

Thank you, everyone.