Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript & Summary

All right. I think we can probably get started if folks can find seats, that would be great. My name is Anand Agarwal. I'm with JPMorgan. I'm joined here today by Lorenzo Flores, CFO of Lattice and Ford Tamer, CEO. Again, thank you for joining us today. We have about 30 minutes for this fireside chat. And maybe at the end, we'll leave a couple of minutes for just some Q&A.

Maybe to kick things off, Ford, would you like to give the audience here a brief overview of the business, market overview, competitive positioning? Any recent business updates that you have, would love to get those.

Yes. That's great. So I'm going to start with an overview of the FPGA market and recent trends and then we'll go into the positioning and competitive nature of stuff. So the way I think about it is a spectrum all the way from, call it, on the left ASICs that are going to be very focused, custom built for the application. It's going to be the highest performance, the best power, the best cost per unit. And you can do that if you've got $100 million, $200 million to spare on designing and producing that ASIC. All the way to on the other side of the spectrum, microcontrollers that are going to be more general purpose, easier to program, you could do it in sea and it's pretty flexible. And we're in between. So we would be higher performance than a microcontroller, we would be more flexible than an ASIC. So if you've got applications like security where the requirements are still evolving, you'd be better doing this in an FPGA because you could change the requirements over time. We're going to be good on power, on size. And so the FPGA is this in between. I describe it sometimes as the Swiss Army Knife. So you can make that Swiss Army Knife look like whatever you'd like for the application, a bottle opener or a knife or a scissor or -- so we serve multiple SKUs in that same silicon. If you look at the applications, pick and you set a new application like a humanoid as an example, that has many motors, many image sensor in a humanoid. And microcontroller typically be good if you've got a one-to-one, like a microcontroller would help you to do a motor, an image sensor. If you're trying to do many motors, would be a better fit. If you do trying many image sensor, it would be a better fit. If you try to do things like a multimodal type of application, image and voice, we could be a better fit. Like a lot of our customers today are using us between the analog feed and the AI processor where we could be a bridge from the analog world, we do sensor fusion on many sensors. So you could bring, for example, in a car, you could bring a LiDAR and a radar and an image sensor and in-cabin type stuff into these FPGA that could help you do the fusion and spare some of these cycles in that AI chip to do what's best for those AI chips. Over the past few years, we see a few tailwinds that are helping this FPGA market grow. Number one, there is an increased focus on trying to get that system design cycle to be shorter. So on the AI side, leaders like NVIDIA are pushing to go from a 2-year cycle between design to closer to 1 year. And that is putting more pressure on some of these ancillary functions as opposed to being an ASIC to now be left better done outside an FPGA. The process node end up continues to be more expensive. So as you go to 2-nanometer, 3-nanometer, it's expensive to put these function in an ASIC. You're better off leaving them outside an FPGA. And as I said, some of these security applications, the requirements keep changing. And again, you're better off doing some of these outside an FPGA. At a high level, the market today, call it, is $7.5 billion to give you a rough view, this is, again, very rough, $4.5 billion, call it, Zynex, $1.5 billion Altera and $0.5 billion each for Microchip ourselves and the 4 major Chinese serving the China market. So you could see this is the market share we've got today is [indiscernible]. We focus on the small and midrange FPGA market, which we believe are going to be faster growing than the large FPGA and SoC. And our goal is to be the leader in the small FPGA and the mid-range FPGA market. In the small FPGA market, we've got a product line called Nexus that is going to be -- that is growing and will continue to grow. We see the inflection point of growth there in 2026. And we've got a new product line called Avant that's focused on the midrange, and we see an inflection point there being in 2027. We serve 4 major markets, the industrial automotive comprising about, call it, 45% of revenue and comms and compute are the other 2 that comprise again, 45% with consumer being about 9%. So under industrial and auto, we put aerospace, defense, medical, some test and measurement in those industrial automotive type of markets. When you think of our product line, we've got silicon, we've got tools, we've got IP, we've got solutions on top to make it easier for customers to program these. And we go to market 85% through China, 15% direct. So a high-level view of the market.

Great. That was great. Maybe double-clicking into one of the verticals that you talked about, the compute and infrastructure vertical. What are -- what is your exposure to general purpose servers and also AI servers? Like how should we think about the attach rates and the opportunity there for Lattice?

Yes. No, great question. So on the comms and compute, we have, in the past, clarified that market for us being, call it, 2/3 compute, 1/3 comm. And in that 2/3 that's compute, we have, in the past, discussed the fact that we got upward of 50 FPGAs per rack. On the AI side, it's higher. On the enterprise side, it is on the lower side. But it's a significant opportunity for us. We characterize the opportunity being between $300 to $500 of content per rack. We have a broad range of functions we supply. We are a companion chip to some of these AI processors. And we work closely with NVIDIA, Habana, all of the cloud accelerator. We also design in with some of the pieces of infrastructure that goes along with some of these like the switches from Broadcom or Marvell or the NIC card, the storage card. We supply a variety of functions such as on security root of trust, boot, post-quantum cryptography, PQC is a recent application, bridging, IO expansion, all kinds of functions that we provide on a variety of cards in the rack.

Great. This next one is for Lorenzo. Given you're fairly new to the firm, I don't want to say new anymore, give me...

I'm still new, but go ahead.

Would love to get your perspective on the business. Would love to get your near-term priorities, your long-term priorities. What attracted you to Lattice? It would be great to kind of get your perspective on the firm.

Yes, sure. So my approach where I've been and at Lattice now is what are we doing as a management team to drive shareholder value? And what are the options we have to accelerate the growth in shareholder value. And when I put that mindset in the frame of Lattice, it'd say, I came here and I have a good feel for the business, right? So my time at Xilinx. And I looked at Lattice and actually from the outside at first, and I said there might not be much left to do, right, because they kind of address their problems, had great revenue growth, all stuff that I know it took years to do from my experience. And so I wasn't all that excited until I met Ford. And then you moved to the -- well, how do you accelerate or what are the options you have to accelerate the growth in shareholder value. And that's where Ford and I spent a lot of time and got very aligned on using this really great franchise, building on it, continuing the execution focus that we have to have every day, every day, every day. So nothing is easy in this business. And then looking creatively around that franchise for ways to grow it in not just incrementally, but within step functions so that we can have a more strategic platform from the view of our customers and have the opportunity to do the classic 1 plus 1 equals 3 in the business. But what I most appreciated about the conversations I had with Ford is we share the same objective in the end, but we also have the same mindset on being extremely disciplined about what we're doing in that frame. And we know the value of doing smaller things that might not be even visible to the world and using our technology platform as a place for tuck-ins to accelerate our say, our tools or our IP offerings and that way, grow the business. So we're looking at the spectrum, which is quite exciting. We have the same disciplined mindset on what we want to do in the end. And the last part of it is, after I met Ford, I met the rest of the team, and it's a really, really good team and really a place where we are on the same mission working together quite well. So it's kind of the whole thing. And I said, I really like the FPGA business because you see all the interesting stuff first in a lot of ways in the industry. So...

That's true. It's always good to hear when a CEO and a CFO see eye to eye on a number of different things, the business is aligned and the different functions aligned. Speaking of growth, you talked a little bit about growth. You've been investing in software for a fairly long time now. Could you maybe talk a little bit about your software strategy, how it's -- what it's comprised of historically and maybe provide a little bit of a future perspective on what you think are the areas that you're going to continue to invest in, things of that nature?

Yes. So look, the biggest barrier to entry and acceleration today is the tools and compared to like for a micro that's easier to program and see, right? So we had one of our investors last night say, look, you could be maybe think of yourself to the micro as to what the GPU was to the CPU. I mean, in a way, CUDA overcame the difficulty because they brought value. And if we can bring value in our offering, could we attract a larger following. But the software is very key because you have to program an FPGA at a much lower level in hardware-like type terms as opposed to a micro that's a higher level. And so at a high -- our focus is going to remain on 3 things: tools, IP and now solutions on top. And I think that's what you're referring to. So on top of tools and IP that we share with NVIDIA with Zynex and Altera, we have an incremental layer of software we call solution that we -- is pretty unique to us. So we've got these 6 solutions, SenseAI for Edge AI, Sentry for security, Automate for industrial automation, drive for automotive, O-RAN for wireless and mVision for vision. So these 6 solution set are intended to speed up a customer time to market. And we're going to remain very focused on those solutions.

Are these -- from a software perspective, are these solutions largely -- what would be -- how should people think about attach rates of the software solutions in conjunction with like the hardware?

Yes. I think the -- obviously, the tools is 100% attach rate. And so that -- from an investment point of view, that continues to be where the bulk of the investment. The second layer on the IP is going to be less than 100%, but this is all the various IP that people need for connectivity and et cetera. We have another layer, which is reference design and boards and payment documentation that, again, is taking investment. And the last layer being this solution is we haven't broken up the percent attach rate. But we've got -- we have to continue to invest in solution. But keep in mind that the tools is what's 100% IP is the next layer. And then eventually, we've got some money left for solutions. So as we scale, we put more into solutions. We would love to put more in solution we're putting today, but we've got to be mindful of the other layers being important as well.

It basically increases ease of adoption essentially, right. I have to bring this up because it's very topical. Given the recent investment into Altera by Silver Lake, we call it an investment. How in your view -- and would love to get both of you to weigh in on this, if that's possible. How do you view -- what are the impacts to the industry, to the broader FPGA industry? And what do you see are the impacts to Lattice?

Yes. I mean for us, it doesn't change our strategy. Our strategy is going to be dictated by customers. I mean I've always built a strategy around customers, around markets that are interesting, not around competition. If we build our strategy around competition, we lose. So we want to be leaders. And if you look at my background at Inphi or Broadcom, I've always led. I've always led by innovative solutions, innovative value that's driven and given to the customer, that's how you build a successful and big company. [indiscernible] and Raghib Hussain are both very capable and are good friends and wish them the best and we'll compete with them in the marketplace. But I don't think it changes our strategy. We remain very focused on the strategy as we have defined it and execute on that.

Yes. What I would add to that is from a day-to-day operational perspective, what we are using the situation as best we can to our advantage, right, and talking to customers about their concerns. And the question you ask is probably more generalized and more acute in the customer base. So what's going to happen next? We don't know what's going to happen next. We'll tell you what we bring to the table, as Ford said, with the customer focus and what's the best value proposition you can get. And that includes not just the products we have, but the commitment to the longevity of the products that we have, which is really, really important in our space and the commitment to supporting the customers through the process. So that's, I think, some of the areas where we think right now, customers are looking at the field and saying, well, Lattice is probably the best solution. So that's -- we want to take advantage of that in the near term.

Great. This is -- my next question is a really hard one to answer. And it's fundamentally -- the reason for that is it's a constantly evolving situation. Tariffs are basically like moving around all over the place, left front and center, depending on the time of the day occasionally. What do you see as the perceived impact of tariffs to your business? And are there any other things that you are doing to mitigate potential impacts from those tariffs, right, from a CEO perspective, also from a CFO perspective, I think it would be helpful to hear.

Yes. So look, every crisis is an opportunity. So I think the good news for us is Lorenzo, myself, the rest of the management team have been through these cycles many times before, whether it's dot-com or the financial meltdown in '08 or the COVID and been there done that, you pay for having an experienced management team that's lived through this and not faced by it. What we tell people is, look, keep your head down, keep executing, focus on, again, what creates value, don't worry about the stock price. And -- now having said this, COVID was a huge opportunity for Inphi, and we took advantage of it. I remember in April when everybody was running for the hills, I published a white paper. I don't know if you remember it, Sanjay, where we detailed all the -- why COVID was going to be a huge boom for the rest of the industry, right? And it ended up being a huge boom, and we took advantage of it. I think in a crisis, strong companies get stronger. And so I think we will take care -- we would -- I think we'll take advantage of the situation. Now having said this, we've got to be mindful of the supply chain is going to be bifurcated for the foreseeable future. I don't think that goes back to one supply chain. So we have a non-China, non-Taiwan supply chain. We have a China-for-China supply chain. We're going to work on the U.S. for U.S. supply chain. So the supply chain diversification and split is going to be something that will be with us. It's great to see the recent progress. We remain focused on understanding sectoral tariff and be mindful of this. On April 1, April 2, I was at the Semiconductor Industry Association in Washington. We left the Hill at 3:00. Liberation day was at 4:00, that April 2. And we aligned as a unified SIA with a unified message to saying, look, extend and expand, so extend the ITC tax credit for the manufacturers and expand it to R&D tax credit. The semiconductor industry is a 70% export and trade type of industry. So we've got to be mindful that we've got to align with partners. So do you want to...

Yes. So from my perspective, I would start by reflecting back on your question, which is it's an uncertain environment. We don't have consistent and clear direction. So we're doing probably what everybody else is doing. I would expect everybody else is doing, which is continue to monitor the environment. We have internal experts who are working with external experts to make sure we understand what the current rules are and what directions they may likely evolve given the different high-level directions that we see coming from the White House. And then in parallel, from an operational perspective, we have a good handle on our supply chain. I think we've talked a lot about the fact that all of our supply chain is outside of the U.S., wafers come from Taiwan, Korea and Japan, and the back end is in Taiwan and Malaysia primarily. We have a way to service China, and we have about 20%-ish of our revenue coming into the United States. Now we do that through distribution partner. And so we think we can figure out ways to mitigate the effect of any tariffs. So it's not an economic disruption to our direct customers. But we are concerned as everybody is like, well, what about the bigger picture indirect impacts in our business. And so we're continuing to watch that. We think we saw some of that, in fact. We see some of that, in fact, in industrial and auto in our business and the slowdown. But we'll see. We'll see how these particular rules get implemented. We're also watching, as you are aware of the potential for sector-based tariffs, and that remains out there, not resolved yet. So we have scenarios we have planned against and we have options to mitigate the impact, but we haven't implemented anything yet because the guidance is not yet definitive.

Also, consumer is a very small piece of your business, which is a great, I guess, in this environment. Ford, you have a very strong record of growing businesses organically and inorganically. Here at Lattice, I would -- what are some of the kinds of inorganic opportunities that you would look at to potentially accelerate growth to potentially accrete margin? What are some of the things that you could potentially do?

Yes. So I think the way we like to answer the question is our shareholders pay us to look at everything and our shareholders pay us to stay very disciplined, right? So like Lorenzo likes to say, we've got a very broad lens, but a very narrow aperture. So we're going to be very picky on what is going to fit and keep the very high bar on what we do. So if you look at the offering, the platform that we have is silicon tools, IP solutions and then reference board and various partnership around that platform. You could imagine all of the above being potential inorganic opportunities. At the developer conference we had in December, that was attended by 6,000 people registered, we had outlined 3 vectors to focus on security, vision and Edge AI. And again, those 3 could be areas for us to grow inorganically. We are into many verticals such as industrial, automotive, comms and compute. And again, those 4 could be areas for us to grow. So you could see that we have a broad lens and we're going to stay, as Lorenzo like to say, focused from an aperture and discipline -- financial discipline on what makes sense to make sure that what we do, the 1 plus 1 will equal to better than what we could do on our own growth.

Sure. This one is for you, Lorenzo. Given the current environment, your gross margins have been extremely resilient. How do you think about gross margin trends in the long run? And can you maybe even talk a little bit about the growth versus margin trade-off, right? And how you think about that?

Sure. That's a really important way to get into the business model, a very easy way to get into the business model question because gross margin is a primary indicator of the value we provide to our customers. And we work with a broad set of customers across a broad set of applications and in different spaces. And that's a portfolio number, if you will. But it's important to look down the P&L and look at the investment we're making in the R&D and in the sales support organization because that's also reflected in the value. It's not just the silicon. It's all of the value we bring to the customer. And so when you think about the gross margin and the gross margin percent, one, it's not easy to be where we are and that takes continual work. Two, to move it up, the math starts to get very hard, right? It's not -- you say you go from 69% to 70%, it's not 1 percentage point of price increase or cost. It's actually 3x that when you do the math. So it's actually real work. And we keep working at it from the cost side very, very rigorously, and we keep doing our best on the pricing side to sell value to the customers. But when you look at the opportunities that come to us, we don't have this, it's that corporate gross margin cutoff. We look at how it fits into the overall strategy, and we look at can we use it in conjunction with the overall business model to grow operating profit growth disproportionately faster, right? Because we're looking for return on the investments we make. That return is primarily in R&D, right? And so that's the way we look at it. And we've managed in the past and obviously, personally in my Xilinx experience, managed to keep at that gross margin level while we grow. So we don't have any concerns about maintaining it as we look forward to our new product ramps and their fit in the portfolio. But I will tell you, it's not -- we don't take it for granted, and we work really hard to grow aggregate profit dollars. So...

Got it. Maybe shifting gears from the growth versus margin discussion to essentially drivers of growth. What do you see as like the short-term, long-term drivers of the business? And also, what are some of the end markets where you see a bunch of your demand coming from in 2025 and maybe 2026?

Yes. So look, the driver of growth is going to be the ramp of our new product. Nexus is the small FPGA. Avant is the medium-sized FPGA. And we've broken that out as a percent of new product revenue in past public statement saying in 2023, 9% of revenue came from new products. '24 was 14%, expected to go in high teens in '25, expected to be in the mid-20s in '26. And so if you do the math, if you do 7% of $500 million in '24, that's about $70 million, 7-0. If you take the mid 25% of the consensus for next year, it'd be more than double that number above $150 million. So you could see that new product is growing, doubling in a matter of less than 2 years, and we'll continue to focus on this. On the segments, the strength now is coming from comms and compute. And we expect sequential growth every quarter for the remainder of the year and into 2026. So comms and compute will continue to be very strong and expect that to continue. But we do expect the industrial and auto to regain its strength and growth as we get out of the inventory situation. And we said that we're now looking to get inventory back to regular business by end of this year.

Got it. And no discussion around end markets would be complete without talking about end market recovery. And given where we've been over the last several years or several quarters at least. Have you seen it? Have you seen the recovery pan out the way you'd expect to? And also, what are your expectations for the rest of the year?

Yes. So let me start and turn it over to Lorenzo. But we said we are at the bottom. We're at the bottom of that. We said it's a U-shaped recovery. So we're at the bottom. And we still -- on the comms and compute, we are shipping to true demand. So there's no more inventory issue in the channel for comms and compute. So comms and compute has recovered, is growing nicely, expected to continue to grow. On the industrial auto, again, we are at the bottom. We are undershipping through demand. So we are continuing to decrease inventory in the channel consciously every quarter and expect to do so for the remainder of the year. Lorenzo?

Actually, I think that's pretty much it.

Okay. I think we have a couple of minutes left. So if there's any questions from the audience, happy to take those, but yes, go ahead.

Ford, one thing that's really piqued my interest with your company is the burgeoning motion control opportunity. So the -- you could do motion control with MCUs or CPUs, but that's serial. You guys offer the ability to do it in parallel, so you can cover a much bigger -- a broader swath of motion control devices. So can you talk about that as you look forward that market? And is there any additional IP you need to support that? Or how should we think about growth in that market?

Yes. No, good question. So motion control has been an area that is a strength of ours. So as you mentioned, parallel processing gives us performance advantages compared to a serial processing. But there are many other advantages. Latency is very important because if there's an obstacle in the way, the robot has to be able to stop pretty quick. And so latency is much better in FPGA. Fast boot time. So we can boot much faster than any other solution on the market today. Compared to other FPGA vendors, we're in order of magnitude, faster boot time. So that's very important. Power is extremely important. And we, again, very focused on power. Small size is very important. We're very focused on small size. We're a deterministic solution. We're much more accurate. So if you're trying to do things like screw driving, accuracy becomes very important. We're multi-sensor. So again, if you want to bring multi-sensor and mix voice and video together, we're going to be a much better solution. We're multi-motors. So if you want to do one motor, then -- and micro is fine. If you want to do control many motors and FPGA, we could put 10 motors on FPGA. If you want to synchronize all these motors together, so if you got many motors and you want to synchronize them together, we've got 1588 that allows you to have a very accurate synchronization of all the motors together, okay? And then we are bringing different connectivity to the fold, so we can do Ethernet, but we currently have -- we've been approved for EtherCAT. So we were at Embedded World in Munich in March. And there, Beckhoff approved us to -- for EtherCAT, which is going to be very big in those type of applications. We also have other variety of proprietary communication protocols that our customers have done for motor control that we're working with them on. So -- and we have a whole array of partners on the sensor side. So we've got many types of motors that we're currently focused on in our partnership ecosystem. And very similar equation to the rest of the ecosystem. So if you look at the ecosystem of partners we have beyond motors, we would have image sensor and LiDARs and radar, infrared, proximity, terminal, et cetera. Look, we have a partnership that today ships tens of billions of different sensor and motors worldwide. So I think you touched on a very good and differentiated point for us.

We've probably got time for one last question. But if there are none, I will thank Lorenzo and Ford for their time. And thank you all for joining us today. Thanks very much.

Thank you, everybody.