Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript & Summary

Thanks, everybody. Welcome back. It's kind of loud -- this room [indiscernible]. Very happy to see you, guys. I'm Joe Moore from Morgan Stanley research. Very happy to have the executive team of Lattice Semiconductor with us today, Jim and Sherri.

So maybe we'll just go straight into Q&A. But if you could start with just a little bit of an overview, what are the big corporate priorities? What are you -- you've talked about this as being one of the biggest kind of development periods in the company's history. Can you talk us through some of that?

Yes. When I look over the -- Lattice has been around for 40 years, over 40 years. And what we're really excited about is we are going through the biggest product portfolio expansion we've ever done in the company's history. And that really started -- it's been a little over 5 years. You and I are into our sixth year now. And beginning 5, 6 years ago, we really started double down -- doubling down on the build-out of the product portfolio for the company. And step one in that was Lattice had always been really strong in small FPGAs. That's where the history of the company is. Lattice is the highest-volume FPGA maker in the world that we always focused on very power-efficient, very size-efficient FPGAs. But what we needed to do is rebuild our product portfolio. And so this is back in 2018, 2019. So over the course of the last few years, we've tripled the rate and pace that we're bringing new products to market. And today, in small FPGA, we have by far the strongest product portfolio that we've ever had in the company's history. And then in parallel, what we've been doing is expanding the portfolio as well. So we also developed a new mid-range FPGA portfolio called Avant, which doubles the addressable market of the company, expands our ability to drive faster share of wallet gain with our customers. And if you look at the target customers for Avant, 90% of those target customers are the customers of Lattice today. The Avant product line uses the same software as our existing products. So we're not just sitting today with strongest portfolio we ever had in small FPGA, but we've now doubled it to include mid-range devices as well. And we're just at the very beginning of the revenue ramp for Avant. Avant just started ramping at the very end of last year. So we're sitting in front of a multiyear revenue ramp with Avant.

Okay. Great. It's good to hear. So you've been investing for a number of years in the software portfolio. Can you talk about the software strategy? Give us an update on your investments there, and maybe contrast it with what some of your competition is doing.

Yes, this was -- what I was talking about was mostly chip-focused, hardware-focused. But the second big part of our strategy, and this goes back to 2019 when Sherri and I did our first Investor Day actually for the company. And what we said is, "Hey, we're also going to ramp up investment in software." And if you look over the last 5, 6 years, we've certainly been growing investment in the hardware portfolio, but we've actually been growing investment in software at an even faster pace. And one particular area where we were focused on was building out a portfolio of what we call application-specific solution stacks. And so think about these as, Joe, these are sort of prebuilt software tools, libraries, a prebuilt set of software for our customers that makes it, number one, really easy for customers to adopt Lattice products or to switch from a competitor's device to our device but also really helps them speed up their own time to market, time to revenue. And so that was a big goal of the company. We've now built out a portfolio of 6 different software solution stacks ranging from AI to security to factory automation, a number of different software stacks but focused on common usage models and applications. And actually, the very first one that we invested in was around artificial intelligence inference technology, which was SenseAI, and that's actually been our most broadly adopted software stack.

Okay. Probably a little bit of interest in that these days.

Probably.

Yes. I remember talking to someone in the ADAS space in driver assistance probably 3 or 4 years ago where they -- it was a nontraditional -- it was more of a software system designer. And you said if the FPGA companies could do more of what you're describing that it would really open things up because they don't have the capability of doing the traditional FPGA design from the beginning.

Yes. So what that software stack, and like I said, the first one we introduced was SenseAI. It was around inference technology. And by the way, when we talked about this in 2019 at our first Investor Day, nobody cared. So nobody...

I cared a little bit.

Yes maybe...

Not like today.

Maybe not as much as...

Like today, you get like 5 multiple points just for mentioning AI.

But we were serious. We saw this as a growth area for the company but what -- a lot of those customers -- so a lot of our customers are industrial customers, automotive customers, customers that may not have a lot of expertise in computing workloads like AI. And so what we were trying to do is abstract out the complexity for them and make it easy for them to adopt this to design the functionality of AI, inference in particular, into their systems without having to understand the details of how the FPGA functions. And so a lot of the purpose of the software is to abstract out and to make it easy for the customers. But also as the customers use the software, they get -- they use it in their systems, they integrate it into their system-level software, that actually creates long-term stickiness for our devices over time as well because that creates a multigenerational need for Lattice solutions and technology. So that's kind of the dual benefit of it. But it's, yes, really focused on making it easy to adopt inference -- Lattice devices for inference application.

And what are the -- maybe you could give us a couple of examples of applications. And it seems these are like vision in an industrial sector, automation environment, that type of thing?

Yes. There's kind of 3 categories in general that -- and we just recently broke out what we call our AI-related revenue. And we said about -- if we look at last year's revenue, about $100 million of last year's revenue, which is about 14% of our revenue last year, is AI-related revenue and then -- and broken into kind of 3 general categories. One is inference done in a wide range of applications like some of those that I was just mentioning, industrial automation, robotics, automotive electronics. That's kind of a broad category. Second category is AI-enabled PCs. So Lattice devices used in things like Dell laptops, Lenovo laptops to do AI inference tasks in a PC. And then the third category is where Lattice chips are being used in control management and security of AI servers. So AI enabled their AI-optimized servers in the data center. So those are kind of the 3 broad categories. And as I said, we -- organically, we started investing in this in 2019, but we also made inorganic investment, too. We acquired an AI software company back a little over 2 years ago in 2021, which is -- was the world's leader in computer vision software for AI. And computer vision has application across all sorts of different markets, industrial, automotive, client devices. And that was a company called Mirametrix. They were already monetizing their AI software. They had been deployed in across tens of millions of devices. They had a really solid revenue stream. They were more profitable than Lattice when we acquired them. And we've been driving synergy from that acquisition over the past years as well, both getting that Mirametrix software established across broad range of our thousands of customers but also using that software to help get Lattice devices designed into new AI applications. So that's another example of we've been investing both from an organic perspective but also an inorganic perspective.

Yes, someone in the automotive domain was talking about this, another ADAS conversation. That sort of irony of that market is that we started off doing things in GPUs as a prototype and then kind of migrated it to FPGAs. Like the opposite of what you think about, yes.

Yes.

Okay. And how do you think about the sort of -- is it monetization of that software simply that you enable the use of FPGAs in more applications? Or is there a software monetization strategy?

Yes, there's both today. So that Mirametrix, the company that we acquired, we continue to monetize that software separately as a separate revenue stream. But then also that software that we're building out in terms of that solution stack, so 6 solution stacks I talked about, that's really monetized as part of the overall hardware/software solution that we sell to customers. So that's monetized along with the silicon.

Okay. And any of the other solution stacks that you're seeing particular traction?

Yes. So we built out a security software stack. This is around platform hardware security. So what it does is, for instance, in a server application, it works with our FPGA that has special security technology embedded in the FPGA, but the software works in conjunction with that for platform root of trust. So it makes sure that as the server boots up that, that server boots up in a secure state. So the combination of chips and software goes and make sure that nobody's corrupted the hardware or the firmware in the system. If the firmware is corrupted, it can repair that to make sure that the server boots up in a safe state. So that's an example of another one of our solution stacks. We recently introduced one for automotive electronics, industrial automation. So it's a pretty good, wide portfolio today, and we expect to continue to expand that portfolio moving forward.

And would you engage at a sort of single customer level for the right opportunities around that if a customer asked you to help them do...

Yes, if it was a big enough strategic customer, but the job or the focus of those solution stacks were we were trying to take where you may see 10 or 100 customers all trying to do a similar task like trying to do inference at the edge of the network. What we're trying to do is rather than having all those 100 customers each implement it separately themselves, do a lot of that work for them, kind of standardize that work, provide that as a built-in software library package that they can just take off the shelf or they can customize themselves. So we're trying to make it easier for our customers to try to take some of the engineering workload on ourselves, right, and provide that software across the wide customer base to, again, all about making it easy for customers to adopt the solution and get to market quickly.

Yes. Okay. That's an interesting stuff. So maybe a couple of questions for Sherri. Could you talk about the gross margins? You've seen significant gross margin expansion over the last 5 years. I remember we used to talk about can you get to the levels that some of the bigger peers have gotten to, and you've certainly executed towards that. Can you talk about the gross margin puts and takes from here?

Yes, sure. Thanks, Joe. So if you go back to the end of 2018 when we put our gross margin expansion strategy, put it out there, we've improved our gross margin by just about 1,400 basis points, and that's through the end of Q4. So the elements of that strategy have really been a multiple. Pricing certainly has been part of it, pricing our products for the value that they provide. We've seen a trend from our customers towards higher capacity, higher capability that's driven improvements in gross margin. Certainly, software attach, we talked about our software solution stack. That has added value to our gross margin. So that's been one of the factors. Mix has also been a factor to our gross margin expansion strategy. Mix towards industrial and automotive and [ quantum computer ], about 90% of our business now. And so that's been a shift away from consumer, which used to be about 1/3 of our business several years back. So that mix shift has contributed to our gross margin. And then also product cost efficiencies have also helped drive gross margin expansion. So all those factors together are areas that we continue to focus on. And when you look at Q4, our gross margin, down a little bit sequentially. That was due to mix coming from our industrial and automotive market segment softening a bit. And when you look at our Q1 guide, 69% at the midpoint, that again, it's down a little bit, but that's also due to mix as we continue -- we talked about softness that we continue to expect into Q1 in the industrial and automotive market segment. And that market segment tends to be a higher gross margin market segment, and so that's really what's happening there. But in general, you can certainly see impacts from mix. That usually would be the driver. You can see some fluctuations on a quarterly basis. But we continue to focus on gross margin.

Great. I mean I wonder the role of pricing in that. I mean I remember when Xilinx and Altera were both public, pretty strong discipline on margin because if margins fluctuated 50 basis points, the stocks really felt it. And then they got absorbed by bigger companies. You worry that particularly the ones that are struggling might price a little bit more aggressively. It doesn't seem like that's really played out.

Yes. And pricing from what we've seen is pretty stable. If we look over the last 5 years, we've had a program around doing pricing optimization across our product lines. And there's been a lot of different market conditions over the last 5 years. And when we look at our pricing on a per unit or per SKU basis, been very stable through lots of different market conditions. So I think that's partially just the structure of the FPGA market. I don't think there's dramatic fluctuations in pricing. So we view pricing as pretty durable and stable. The one factor in our pricing or maybe more ASPs is we have seen ASPs over the years steadily go up each year. Now that's more of a product mix, a function of product mix because as we've introduced higher capability, higher capacity, higher devices with more software, et cetera, that is -- and we've basically been introducing higher ASP products over time as those have become a greater portion of the mix. We've seen ASPs steadily go up over time along with the product mix and the higher capability device. So -- and so that's one fact that -- or one factor we've seen in our pricing overall ASPs. We would expect that to continue as well because the new product line of Avant for the mid-range, the ASPs of Avant are about 10 to 20x higher than the company's pre-Avant ASPs or our ASPs for small FPGAs. And so as Avant becomes a bigger portion of the revenue, that should be a significant uplift to the ASPs.

Great. Yes. Okay. And then another question for Sherri. Free cash flow has been really strong. Can you talk about that, what you see for free cash flow going forward and maybe uses of cash?

Sure, sure. Thanks for noticing free cash flow. We actually put out a target at our Investor Day last year for free cash flow at greater than 30%. And so we've been executing on that quite well. For 2023, we had another record free cash flow of 34% for the full year. Q4 alone was 40%. So really pleased with those strong results. What we did during the year, a couple of things to note on capital allocation is we did pay down our debt, so we are completely debt-free. I'm really pleased with that. And we've been continuing on executing on our share repurchase program now for 13 consecutive quarters. And so now that we paid down our debt that -- we don't have to use money for debt pay down. We can earmark that, if you will, more toward our share repurchase program. And this year, our -- or at the end of last year, rather, our Board authorized a $250 million share repurchase program. So that was a greater amount than what it was in prior years, and that expires at the end of this year. So a greater envelope of share repurchases that we're authorized to do because we have more cash that we can earmark towards that. So we're really pleased with that. [indiscernible], since we've been executing on share repurchases, we bought back just under about 5 million shares, and that's reduced our dilution by about 3.4%. So really pleased with that. We hear our shareholders are really pleased with that as well. And we also indicated in our last earnings call that we intend to continue executing on share buybacks. We want to keep doing that on a quarterly basis to return capital to shareholders. But the first priority from a capital allocation perspective, it's really organic investments in the business. We want to continue making investments for the long-term growth of the company. If you look at our R&D spend, year-over-year up significantly. We want to continue making those investments for the long-term road map. So that continues to be the top priority.

Okay. And the role of M&A in that, I mean, you talked about buying an AI software company.

Yes.

Are there still sort of tuck-in type assets like that, that you're looking at?

Yes, good question. We're -- I mean we're always scanning the environment if there's any inorganic opportunities. If there's things that we can add to the portfolio that accelerates our strategy, we always look at through the -- look at it through the lens of our organic strategy. So we have very, very high conviction in our organic strategy. So for us, inorganic, the way we look at it is what are the things we can do inorganically that would help us accelerate that strategy over time. Those could be hardware or software acquisitions. It could be smaller tuck-ins or more meaningful, bigger acquisitions. Yes, and the most recent acquisition was that software acquisition. And that was a great example of -- that software company, Mirametrix, we are already working with them together at a number of customers. We saw the ability to, if we brought them in-house, to drive synergy in both directions, synergy of being able to establish -- basically bring their software across our wide range, so we have over 10,000 customers. So being able to accelerate their introduction to a lot of customers but also their ability from a software perspective to help us proliferate Lattice silicon across a broader range of applications as well. And so yes, if there's software or hardware acquisitions that would help accelerate that organic strategy, we're definitely open to that, yes.

Okay. Great. So in terms of growth, you've had 3 years of double-digit growth. We've hit a bit of a cyclical correction here for really pretty much everybody in our coverage. Oddly out of phase, different sectors seeing it at different times, but everyone's seeing an inventory correction. What's your visibility as to when that bottoms out and when you can get back to resuming growth?

Yes. I think in terms of the correction, first of all, I spent my entire career in semiconductors, there -- yes, it's part of the industry. There are cyclic corrections. And no matter what anybody says, in the future, there will be another cycle, right? So this particular cycle aggravated a little bit more by the pandemic experience. We saw -- as an industry, we saw supply chain lead times get very, very long. And when supply chain lead times get long, customers start to hold more inventory. And now that supply chains are back to normal, lead times are back to normal for the industry, we're seeing now customers draw down those inventory reserves that they had built up. So this inventory correction, we think, certainly affects demand through the first half of this year. The view across kind of our customer base, I think a consensus view which we agree with is that, that inventory correction or drawdown starts to dissipate in the second half. And it's not like a light switch, it turns off. But it starts to -- that drawdown starts to dissipate and end customer orders start to return back to natural consumption rates. That's kind of where we think we're at in the cycle overall. But for us, the short-term cyclic air pocket, inventory correction, whatever you want to call it, that the industry is going through, that doesn't change our fundamental strategy. It doesn't change the product portfolio build-out, the software build-out that we're doing. It really doesn't fundamentally change where we're headed long term as a company. And I think our growth algorithm that we -- that we've talked about in the past, absolutely, we remain committed to that growth algorithm. And that growth algorithm had 2 parts to it. Part one of the growth algorithm which was -- which we talked about all the way back in 2019 at our first Investor Day was we said, "Hey, in the small FPGA space where Lattice has always been historically strong, we believe we could grow consistently over a multiyear long-term period at double-digit rates, low double-digit rates kind of in that 10% to 15%." And if you look at since that Investor Day in 2019, the company has grown at a CAGR of 16%. So okay, so we grew a little bit faster than the range that we gave back in 2019. But we believe, again, over the long term, in small FPGAs, we can continue to grow at that same growth rate of 10% to 15% in small FPGAs as we continue to build out new products, new software, et cetera. But then what layers on top of that is now we're at the beginning of the Avant ramp, right? And so Avant, our new midrange product line, again, same customer base, same software that has just started, it's just at the very beginning of the revenue ramp. And as Avant -- and that's all additive revenue. Avant doesn't cannibalize the existing small FPGAs in any way. It creates an entirely new additive revenue stream. And as that revenue stream grows and becomes a greater percentage of the company's revenue, that starts to accelerate our growth rate over time. And in our last Investor Day, we said, "Hey, Avant growing on top of the small FPGA growth rate, that should accelerate our growth rate into the 15% to 20% range over the next 3 to 4 years as Avant becomes a more meaningful component of the company's revenue." So we remain committed to that long-term growth target or growth algorithm. We think that's the right algorithm for the company.

Great. I want to come back to the product splits. But I just -- do you have any sense for is end demand different? Like you talked about the inventory aspect. It's sort of axiomatic when lead times come down, people need to hold less. But it seems like, broadly speaking, for FPGAs, there's one end market that's softer, which is 5G infrastructure, less of an issue for you than some of your bigger peers. But the other markets seems like demand is okay, and it's just an inventory issue, but do you have visibility on that?

Definitely an inventory issue. But I do think that even in some of the other end markets, there was a -- during the pandemic, there was a spike in demand that our end customers saw that is now -- that spike in demand has started to normalize back to what I would call more normal demand levels. So I do think there's a drop off in demand that's happened as well, right? Back to a more sustainable long-term growth rate, right? So I think we're back to normal -- not just normal supply chain dynamics but also back to a more normal growth rate with our end customers as well.

Yes. Okay. And you kind of walked through the accretive aspect of Avant that there's incremental revenue that doesn't take anything away from your existing growth. Can you help us quantify that at all? Like how much does that contribute kind of per year in the next 2, 3 years?

Yes. In fact, what we shared at the last Investor Day is we said, if you look at -- when we gave that growth algorithm is what we said is when you look 3 to 4 years out, we expect Avant to account for about 15% to 20% of our revenue in 3 to 4 years out, right?

Okay.

So it starts ramping. Well, it started at the very end of last year. It runs through the course of this year and through the next 3 to 4 years to about 15% to 20% of overall company revenue.

Okay. Okay. And your -- the underlying growth of the Nexus products continues to be strong.

Yes, exactly. So that -- well, it's not just Nexus but pre-Nexus products as well. This is why we don't -- we never refer to pre-Nexus as legacy products because those products continue to grow, all right?

Okay.

And we're expecting those over the long term to continue to grow. A lot of times, that new software that we introduced, which is used on Avant and Nexus, also used on pre-Nexus devices, that has rejuvenated the life of some of those older products. And we're seeing through last year, if we look at last year's growth of overall whatever, 12% last year, a good component of that was pre-Nexus products continuing to grow. So over the long term, we expect those to continue to grow as well. So -- but obviously, Nexus and Avant, some of our newer products growing at a much faster rate. But yes, in that small FPGA space where we've demonstrated the ability to consistently grow at this low double-digit rate, we expect that to continue over the long term and then Avant to be additive on top.

Okay. Okay. Great. And then how do you think about managing financially through this kind of period of weakness? You prioritize spending money in R&D. How do you think about OpEx flexibility in a year like this?

Yes. So if you look at our Q4 OpEx versus our Q1 guide, we've guided OpEx to be essentially flat at the midpoint for Q1. And within that, though, we want to make sure that we're continuing to invest for the long term. I talked a little bit earlier about that #1 priority of investing for the long term. And so that continues to be our priority. Certainly, on a quarter-to-quarter basis, you can have just fluctuations in OpEx and R&D spend due to the timing of certain programs and things like that. So that, of course, is part of the business. But the #1 priority will continue to be investing for the long term. But having said that, certainly, Q4, revenue a little bit lower. Q1, lower at the midpoint of the guide. We want to make sure that we're disciplined about the way that we spend. So it's really a balanced approach, making sure we're disciplined in how we look at our G&A spend, in particular, but investing for the long term.

Yes. Great. So I just have one more question, and we can turn it to the audience. From my perspective, I've always covered Altera, Xilinx. I've seen them be acquired. And my perception of a lot of the benefit you guys have is being standalone, being focused on these businesses while your competitors are part of a bigger company, distracted by different things. Do you perceive that to be the case? And I guess with Altera at least, we're seeing now moving towards an IPO, moving -- having more of a visible CEO making decisions that they haven't had for a while. So it seems like you're getting some [ emergence ]. And AMD has always kind of run it a little bit distinctly from the rest of the business. But how do you perceive that competitive dynamic as a small company competing with much bigger companies but as one that has more focus and probably more potential for innovation?

Yes. I just don't believe it fundamentally alters the strategic landscape within the 3 FPGA companies, right? So Lattice, Altera, Xilinx were all founded about 40 years ago, right? And for most of Lattice's history, actually over 3 decades, Lattice competed against Altera and Xilinx as independent companies, right? And so okay, Altera temporarily part of Intel, now will be part or will be separate again. I just don't see that fundamentally altering the strategic landscape that the company has been operating under for 40 years. I think for Lattice, I think a lot of the improvement you've seen over the last years or maybe the underperformance of years ago, I really think that had little to do with our competition and really much more to do with Lattice. A lot of our issues were self-inflicted, right, historically. And I think a lot of what we've seen over the last 5 to 6 years is Lattice driving great products to market and speeding up our product development, doing the right things for our business. And so that -- the benefits that you've seen are really, I think, more Lattice-driven than some result of who owns our competitors, right? And yes, I don't -- I just don't see that as fundamentally changing the landscape that the company's -- the competitive landscape the company has been operating under for 40 years.

Okay. Great. Let's see if we have any questions from the audience.

Okay. Could you talk a little bit about competitive landscape, especially like in China, considering China is putting more effort in localization, and there are 2 rising FPGA players in China, PANGO and Anlogic?

Yes. So definitely, let me start by saying we take all competitors -- regardless of whatever geography they're in, we take all competitors seriously. And actually, when I joined the company in 2018, one of the things that I really stressed with the team is our competitive philosophy is that we assume that there will always be robust competition across every product category, across every market, across every geography. That's our default assumption. And we build -- in our product planning, we create phantom competitive parts to measure ourselves against to make sure that we're driving a strong competitive road map. And then look, if competition materializes, we're ready. And if competition doesn't materialize, that's upside for us. So that is the default philosophical operational framework that the company operates under is to always assume strong competition. Now within the China market specifically, there's definitely -- obviously, the Chinese government is trying to cultivate domestic competitors for not just every FPGA, every single type of silicon that exists, right? Now across all the different types of silicon, I would argue having been part of many different types of silicon over the course of my career, I would say FPGAs are one of the hardest types of products to replicate. The FPGAs have very high barriers to entry. The silicon itself is difficult to design. The software is difficult to design. And the hardware/software combination together creates a very high barrier to entry. That's why you've seen over the last 40 years really just 3 general purpose FPGA companies: Xilinx, Altera and Lattice. Now in the China market, yes, there are definitely some Chinese start-up companies in FPGAs. We take those competitors just like all of our competitors very, very seriously. We have a very good team in China. Actually, Lattice has been in China since I think it's early '90s. We have a very good team with -- Lattice team within China that focuses on our Chinese customers, doing software and application engineers that's actually -- application engineer that's actually unique to the China market and to our China customers. So I think we've done a really, really good job of competing in that China market, certainly our intention to continue that. And -- but globally, I think we'll always take the -- not just in China, but we always take the view that we assume that there will always be robust competition, and we'll be ready for that.

Right. Any other questions? Okay. If not, we'll wrap it up there. Jim, Sherri, thank you so much.

Thanks, Joe.

Thank you, Joe.

Appreciate it.

Thank you. Thank you.

Thank you.