Lattice Semiconductor Corporation (LSCC) Earnings Call Transcript & Summary

Okay. We're going to get started. I see the room is empty. So it's great to see all these people here, and hopefully, we have enough chairs. So I'm Tristan Gerra. I'm the semiconductor analyst at Baird and we're delighted to have with us today Lattice Semiconductor, who I probably don't need to introduce, but it's a leading FPGA and related software platform supplier. And we're pleased to have with us today Esam Elashmawi, Chief Strategy and Marketing Officer; and Sherri Luther, who's the Chief Financial Officer. And with that, let's get started.

Let's get started. Thank you for having us.

Thanks, Tristan.

So maybe first question directly on -- for Sherri on gross margin. Obviously, you've had a very, very significant gain over the past several years. And if you could talk a little bit about the catalyst going forward, given the resiliency in pricing and your mix continuing to improve. And I know that you provided financial targets, the recent analyst there, but if you could maybe elaborate a little bit on what you see medium term for gross margin?

Sure. Sure, Tristan. So the gross margin expansion that we've generated over the past -- going into our fifth year now is 1,360 basis points improvement. And it's really the result of the strategy that we laid out in 2019, and had 3 main elements to it, pricing optimization, which is really -- what that is all about is really how do we get value for what our products provide to our customers. And so really getting that value. We've had undergone the most rapid expansion in our product portfolio in the company's history. So we have a lot of products out there on our Nexus platform and our Avant platform that we've launched at the end of last year. And so our products provide a lot of functionality. And so that's an area where we can get value for our products. We also have a significant -- we see ASP -- higher ASPs with our software attach. And so that adds value. And the other thing that we've seen that's helped driven our gross margin expansion is our product mix. And so product mix is an area where we see that trends for customers wanting higher capacity, greater capability products. And so that has helped driven gross margin improvement. The other areas are product cost reductions where we continue to work with our suppliers. We have long-term relationships with those suppliers. And through that, we've worked to generate operational improvements in areas such as yield and cycle times. And those strong relationships with our suppliers also helps add value. So all of those areas -- items combined together are really the drivers for our gross margin expansion. And again, entering our fifth year, we see opportunity there to continue to drive gross margin. And that's why at our Investor Day a few weeks back, we increased our gross margin target to the low 70s as we continue to execute on that. So we believe our margins are very durable. And so we'll continue that focus on gross margin expansion.

And then in terms of end market and looking at data centers. So I think your attach rate is already above 1. Are we getting to a phase now where your growth in data center is going to be more unit related? Or do you have opportunity for increased content or ASPs, if you could talk about the opportunities that you see to continue wiping that business?

Yes, good question. So data center, what you're referring to specifically is our server end market. And what we had articulated, if you go back a few years at our first Investor Day in 2019, we talked about the fact that we are adding more attach rate for every server and we're driving more functionality with our FPGAs that drive an average higher ASP. And at the time, we had an attach rate of 25%. And then we showed in 2021 Investor Day that we took that to 85% and now we're driving attach rate well above 1. So if you think about it, what does that mean? That means for every new generation of server that's being produced, we're increasing our attach rate, and we're driving more capabilities and functionality that drives a higher FPGA with a higher ASP. And we see that continuing. We don't see that slowing down. In fact, we're working with our customers on the latest generation of servers and we also see a higher attach rate, and we're driving more functionality, which is driving higher ASP. And the reason why we see that and what's happening in the server market. Number one, it's very heterogeneous. We see that there's more modularity in the servers today. So what used to be a simple, hey, here's my motherboard. I've got a CPU, I've got my storage. I got my memory. Today, if you look at server architectures, it's becoming more modular. What that means is that they're breaking the motherboard up into not just the CPU, but you have a control and security module. But we're also seeing more and more plug-in cards in servers today. You have network interface cards, you have GPU cards, you have storage cards. And each one of those are driving more opportunities for Lattice, which is driving our attach rate. And as we bring more functionality with our customers that's driving the higher ASP as well. So we see that trend continuing.

How should I look at the competitive landscape? Is there any ASIC or MCU that can compete? And what's the value add of what you offer relative to other suppliers for that particular functionality?

Yes. So when we look at -- if you are referring to the server market specifically -- so in server market, what we're doing there is; number one, we're bringing functionality around control and board management. And if you go back in time, when I talked about 25% attach rate, there's 2 things that are happening in the server market from a competitive perspective. Number one is we are taking sockets, which used to be traditionally like MCU type sockets, the technical needs of our end customers are increasing and so that migrating more and more towards FPGA. So we're actually replacing what used to be traditional MCU sockets now into FPGAs as we bring in more function out. That's one aspect. The other aspect of it is we sit with our end customers. And we're also identifying new ways to leverage our FPGA. Within the servers, in new control functions, new security functions that prior generations did not have before. So that's what I refer to as really market expansion opportunities for FPGAs. Those are the 2 areas which we're getting more traction with our FPGA.

Great. Obviously, AI, very topical these days. In the server market, what opportunities do you see going forward? Does AI require more vote of trust security or other content enabled by FPGA? You mentioned briefly the increase in GPU-based architecture. There's also smart neck ramping. How is all of that changing the need for FPGA content in service?

Yes, it's actually increasing the need for FPGAs. And I want to be clear, when we talk about AI in the data center, there's AI from a machine learning perspective, where they use large FPGAs. That's not the focus of Lattice. That's not what we're trying to be, and that's not what we're building. We do, however, have AI solutions, and that's really around inferencing around the edge. And there's lots of examples where Lattice is being used on AI and artificial intelligence applications, more on the edge -- not in the data center itself. But as AI in the data center continues to accelerate, people are looking at AI into the data center with GPUs as an example. That drives more opportunities for Lattice. I talked about the fact that servers are becoming more modular, that there's more plug-in cards. If you take a typical AI box that has GPUs in it. You're going to find multiple GPU cards, multiple net cards. There's going to be storage cards, there's going to be a CPU card. These are all opportunities for Lattice. And what we're doing in those, again, is control management as well as security. And yes, you need security even on artificial intelligence. You want to make sure that, that system is protected, that the bad actors aren't coming in and changing the data or taking control over your server. And so security is something that's really, really important, not just for the acceleration of adoption of AI in the data center, but just in general, data center security, that's really important.

So if we were to summarize that in server with this kind of your CAGR opportunity and what's your expectation in terms of outgrowing the whole market, let's say, 5 years out?

Yes. I think we've demonstrated in the past that if you look at the general server market, and I want to start off saying that, okay, so the server market can go up, it can go down. And Lattice isn't immune to the fact that if there's fewer servers being shipped, then there's fewer servers being shipped. But I think the Lattice specific growth drivers go back to the higher attach rate and the higher ASP devices, which drives an increase of dollar content for every server generation. And it's important to note that when we're in the servers, we're also CPU agnostic. So it doesn't matter if servers are being built with AMD, with Intel processors, with ARM processors, et cetera, we're CPU agnostic. What we do is independent of the CPU architecture. In fact, that's a big plus for our customers that we're CPU agnostic. It gives them the choice of what they want to use. So when you think about the server market, it can fluctuate, but it's really the Lattice growth drivers of the higher attach rate, the higher dollar content per server which allows us to weather it better than the overall market.

Great. So you touched on AI at the edge as being an opportunity for you. Maybe if you could expand on that, on the opportunities that you see both near term and medium term? And maybe kind of a review of where this could go in terms of other end markets, you've been traditionally strong in communication, for example, what are the potential applications there and -- as well as in the other end market.

Yes. AI seems to be the hot topic, but we've been talking about AI since 2019. We actually introduced our first software solution stack called SenseAI that was introduced in 2018, and we continue to invest in that software solutions stack. And what Lattice is really good at is producing these very small, power-efficient FPGAs. And FPGAs in general, are parallel processing units. So we do artificial intelligence very well because we can do everything in parallel. And you talked also what are the opportunities. So we're really around inferencing. And I'll give you an example, you take a robot in a factory. These robots now are becoming more intelligent. They're just not moving in a particular manner, but they're putting intelligence in these robots. Why are they putting intelligence in robots. It could be for predictive maintenance. It could be for object detection, package detection, it could be for worker safety, human presence detection. It could be for where is the individual intent? What's his intent? Is it gesture? People don't want people touching stuff as much as they did before. But they're adding these intelligence to these robots. That's an example where we're leveraging opportunities for Lattice FPGAs at the edge. The other examples of that are smart home devices or even in client devices where I think it's public information that there are new client platforms that are actually ramping up, leveraging Lattice FPGAs for artificial intelligence, which brings a better user experience but as well brings improved battery life of the system. So we're seeing more and more of our customers across different end markets, leveraging, putting more intelligence to their systems and we can do that really, really well with these small, power-efficient FPGAs.

Great. And then a question from the audience. Lattice is supposed to be in its largest new product cycle. What's most exciting for customers and for Lattice? And how does it break down by end market? And kind of going back to my first question, how does that catalyze gross margin going forward?

Yes, I think we're all excited about our product portfolio. So the company actually this year becomes 40 years old. Lattice becomes 40 years old this year. And if you look at our product portfolio, this is the largest expansion of our product portfolio in our history. And it's not just in small FPGAs where we're bringing out differentiated innovative products, but also in mid range FPGAs, but also in our software solutions as well. And when you combine those 3 of small, mid and our software solutions, this is a time that the company has never experienced in the past. We've just got many, many products coming out. If you look at the Nexus product line, we have 6 device families that have been introduced into the market since December of 2019. That's a very accelerated rate of new device families being introduced into the market. We just launched our first mid-range FPGA in December of last year, with 2 more that we talked about are going to be launched with our developers conference at the second half of this year and expect more of those to come out as well. But also, it's not just the hardware, the software solution stacks that we're building, these are driving for us, not just higher ASPs on average, as Sherri talked about, but it really helps our customers get to market faster. It helps -- we've got good examples of how customers can get to market faster. The attach rate is higher and it creates stickiness with our customers as well. As they leverage our software solutions into their systems, it creates stickiness which makes it easier for us to continue to win multi-generational designs from our customers. So we're excited about the product portfolio that we have today, what's existed that's already in the market, but what's to come as well.

And then another question from the audience is what is the target inventories? And I think it's referring to your internal inventories. When is the equilibrium? And as a second question, multiple companies citing demand moderation in industrial and automotive and wanted to get your take on it.

Yes. Sure, maybe I'll take the inventory, and I'll let you take the industrial automotive question. So from an inventory perspective, Lattice owned inventory. You would have seen over the past year and maybe even a little bit longer than that, we've been increasing our Lattice owned inventory sequentially quarter-on-quarter. And we've been indicating that, that was our intention to do that. And why are we increasing that inventory? It's because we've got design win ramps and we've got new product ramps. And we want to make sure we've got the right amount of inventory on hand for our customers. The other thing about why I'm in favor of increasing the inventory the way we have is because our products are not perishable. They have very long life cycles. They can last for 10, 15, 20 years. And so that inventory that we're building is a very strategic, very intentional to support that growing demand by our customers for all of these new product ramps that we have. The other thing I'll note is that in Q1, we did indicate that we expect going forward into Q2 that our days of inventory would come down. So that's the directional guidance that we've given. We haven't provided a target or anything. But the way we look at it is we just -- we want to make sure that our customers have the product they need. And again, it's not perishable inventory. So we feel good about the inventory levels that we have.

And the second part of the question was industrial market and automotive and what we're seeing there. This has been a really good growth driver for us, industrial and automotive. In fact, we've had 3 years rate of double-digit growth in industrial and automotive. And what we see there is just a lot of adoption of FPGA, specifically Lattice FPGAs, which are really tuned for those types of applications. Industrial Automotive grew 41% I think in 2022 last quarter. When we look at it year-over-year, it was 50%, and it's just been accelerating. And so what's driving that adoption? If you look at the industrial market and what's happening there, there's just a lot more emphasis on automation, robotics. How do we get more throughput. There are geo supply chain, things that are driving more factories around the world. Cost of labor is going high. So when you look at just factories around the world, it's really around how do I improve my efficiency, how do I get my cost of ownership lower. And it's all around driving new innovations around factory automation. Take robotics. I talked about robotics earlier. When you look at robotics and automation systems, working with our customers, we're assuming there's more than 100 plus million devices deployed on an annual basis. That's a huge opportunity, not just for Lattice, but it's a reflection of what's happening in the industrial market. And each one of those drives opportunities around motor control, artificial intelligence, networking for these robotics, FPGAs do all of those very well. Then when you look at the automotive market -- so that's industrial. When you look at the automotive market, what's happening, what do we see in the automotive market that's driving the growth. Automotive is our fastest-growing segment. So if you look at industrial automotive, and I gave you some stats. The automotive portion is actually outgrowing that total stats of industrial and automotive. But what's happening there? Well, the car manufacturers want to differentiate with electronics and with user experiences and that's driving more and more electrification of vehicles. And we all hear about it. We all see it, where electronics are becoming more important to the OEMs and driving differentiation. They're adding -- just like the industrial, they're adding more sensors, they're adding more displays, and that's driving more need for technology just by FPGAs that can take that and aggregate it, do something with it, preprocess it. And FPGA is being programmable, gives them additional flexibility where over time, they can be able to add more capabilities while these systems are actually deployed themselves, so the programmability is an advantage. And what Lattice does with our FPGAs is we automotive qualify our FPGAs for these types of applications. So industrial and automotive has been a good segment or end market for us and for the multiple reasons that we had talked about. We're very excited about it.

Great. And then going back to the first part of that question, Sherri, you talked about raising inventories. How do you characterize your lead times relative to MCUs? And how would you characterize the supply demand for your product? Is it in balance? Or are you still experiencing shortages. I know that's not the key reason for the share gain, but I just wanted to compare that with the availability of other chips and what does that mean in terms of customers making decisions because they can get products from other suppliers?

Yes. So in terms of our lead times, I mean, our lead times, I think, across the industry, frankly, but certainly for us, have improved from the supply chain constrained time frame. And in terms of the availability of supply, that has also improved in general. And so we've seen that loosen up a bit. But I think just in terms of your comment regarding MCUs and the availability of our supply versus MCUs and my customers kind of come over to Lattice for that, I guess, I mean one thing I'll say, and I certainly have Esam on this as well, is that we haven't seen customers come over to buy FPGAs because they couldn't get supply of MCUs. That has not happened because we would definitely tell them, don't do that because the time that it takes to qualify an FPGA and then to turn around and go back to an MCU would not make any sense. We would tell them, don't do that. What we have seen though is that customers who wanted to come over to convert to FPGAs from MCUs, that we have seen some acceleration of that through COVID. And so they were intending to do that anyway. And so we did see that happened, but not just solely because of supply. Esam, do you want to add anything to that?

Yes. I think that's the good point that a lot of application systems are becoming more and more complex and customers are recognizing that they need more capabilities. And that's what's really driving towards an FPGA. On the lead times, though, what I would say to add on to the comments is, I think overall, we've done a really good job. I'd say better than most in the industry with our supply. And I think that started off if you go back to even at the time of the pandemic, Sherri allowed us to raise the inventories at that time. So it put the company in a much better position when it came to supply. So when it comes to lead times, the way we measure it as well is are we satisfying our customers' demand and needs. And I think if you talk to our customers overall, they'll say we did a really good job in supporting them through the supply constrained environment and especially now that a lot of that constraint is now behind us.

Great. And then back to the second part of the prior question, which was the drivers in automotive. Maybe at a higher level, if we look at the changes in architecture. So for example, there is a shrinkage of ECUs in automotive with much more centralized computing going forward. And then you talked about how FPGA use case are multiplying because of more electronics. So how do I reconcile the 2? And how do I look at the future architecture in automotive and how FPGAs fit that versus what we're seeing today?

Yes. And I think for FPGA, it's similar to the server analogy when we say we're CPU agnostic. If you look at an automobile, doesn't matter if it's a centralized ECU or distributed ECU, we benefit from both. In both cases, they're adding more sensors. They're adding more displays in the vehicle. If you look inside of a car today, you've got a display for the dashboard. You've got infotainment system. You've got a navigation system. Even the mirrors are becoming e-mirrors. You got rearview displays as well. It's just there's more displays in the car. That's independent of centralized ECU or distributed ECU. There's sensors that they're putting in the vehicles today. And it's not just the cameras. It's temperature sensors, pressure monitors, just there's more and more sensors in vigor. That's independent of centralized ECU or distributed ECU. And what an FPGA does really well in these new architectures is it's able to aggregate that stuff, do something with it, and it makes the architecture easier. So let's take an example of a centralized ECU. Well, you're going to have a central large processing unit. Well, that central large processing unit can take so much data in. And when you're adding 10x as many sensors, there's not 10x as many inputs on these. So you have to aggregate the data and you have to do something with it. And FPGAs do that very well. We'll take the sensor data. We'll aggregate it. We'll actually process it or we'll preprocess it. We'll offload that ECU. There's a lot that we can do, and that's independent, again, of centralized or distributed. So we just see more and more opportunities for our FPGAs as we engage with our automotive customers and how they can best leverage the programmability of our devices.

And if you could talk about stickiness in that end market, there's some ASICs also that do data aggregation, but they're not programmable. Do you feel that as the market ramps in volume perhaps there is less of an FPGA attach rate over time? Or do you think that instead, there's absolutely no cannibalization because you're really going to continue to offer upgrades and things that make the FPGA very relevant in that space?

Yes. Really good question, and that's a point that I'm glad you brought that question because you remind me of another trend that we're seeing in the automotive market. And that's -- this is around simplifying the architectures across multiple model vehicles. So when we engage with the customer and they've got different model vehicles, but they all have slightly different specs to them. The beauty of an FPGA is we can build a single system, and they don't have to build multiple ASICs for each one of those different models. They can take like the camera unit or the sensor unit or somehow -- sometimes in the infotainment, you have more features than another vehicle does. But building a central module that's compatible with different model vehicles is something that they're trying to do. They're trying to simplify their lives. They're trying to get to market much faster. You can't do that with multiple ASICs. But with the programmable FPGA, you actually simplify that problem for them. So what we see, again, if I look at this on a very high level. And we talk to the OEMs, we talk to the Tier 1s directly, there's a change in the automotive market. And that's driven by new car companies that are doing things much faster and they're getting to market quicker and they're providing better user experience. That's now being -- you call it the traditional OEMs now are starting to take and learn from them and figure out how they can do things faster, how they can get to market quicker, how can they be more modular from that perspective. And that's driving, again, a need for a programmable FPGA.

Great. And then switching back to a question from the audience, and I think you touched on this a little bit. How big of an opportunity is industrial robotics and for Lattice, is it more Nexus or Avant?

Yes. So again, we estimated over 100 million-plus robots and automation systems deployed for a year. And it's an opportunity for not just our small, power-efficient FPGAs, but also our mid range class of FPGAs as well. We see both opportunities in robotic systems and automation systems today.

Okay. And then another question from the audience, which is the adoption of IoT in factory and manufacturing applications. What are you seeing there? How big of an opportunity is it? And basically, what play do you see on that? NVIDIA has talked about digital twin technology replicating and optimizing the flow virtually. So how do you potentially benefit from that trend?

Yes, good question. We do benefit from it. When you think about a fact, -- in fact, last week, we had an event in Europe where we had imagine top OEM, industrial and automotive C-level suites in GM's meeting with myself and our CEO, Jim Anderson, and we're talking about just overall trends in what's happening. And clearly, IoT and networking in the factories is something that's becoming more important. And if you look at a typical -- take a typical automotive factory in, we're fortunate enough where they toured us in automotive factory last week as well, which is very eye-opening. But there's millions of end points. And there's millions of endpoints that you're collecting data from. And what do you do with that data? How do you improve the throughput of these factories? Well, number one, you've got to be able to connect these systems to some type of a network. And we can do that. We do -- not just do we do factory automation, but we do factory networking as well. And then when you collect that data, what do you do with it? Well, you can do it on-premise with your own data center or you can go to the cloud and leverage the cloud. Well, we just talked about servers and what we do with servers and how we benefit from that as well. But there is a need to do all of that to get that connectivity, but not just connectivity, but you want secure connectivity. That's really, really important as well. And that's an area that we're engaged with some of the top OEMs and how do we do the next-generation connectivity within a factory for not just connecting but also security connecting. So lots of opportunities around IoTs within the factory.

Great. And then quickly on manufacturing, FD-SOI has been kind of a central technology for you with new products in the past. How do you see that road map? Is that going to be a big part of your, I guess, value-add creation in terms of new architectures, and is that going to be across various densities, if you could talk about that as well and the future road map that you see there.

Yes. For those not familiar, our Nexus platform, which is our latest platform for our small FPGAs, that was built on an FD-SOI technology, 28 nanometers FD-SOI. And as I mentioned earlier, we have 6 device families that we introduced to the market and expect us to continue to introduce more there. And there was benefit in us choosing FD-SOI, it gives us lower power, gives us better reliability than a planar technology. When we move to our mid-range product line, which is Avant, our mid-range platform, we actually chose a FinFET technology, 16 nanometers. We're building larger devices, performance is higher, and we chose what we looked at the right technology for that product. So we didn't stick with the FD-SOI for that. We went with 16-nanometer FinFET because that was the right technology to give our customers and for us to develop the products with the benefits and characters that we want. For future products, we will do the same thing. There's always optionality on what process technology node we use and we'll always pick what we believe is the right one to meet our customers' needs and the product needs and we'll choose that at the time.

Okay. I guess I tried to ask about potentially what goes on after Nexus, but...

Stay tuned then.

Great. And then I wanted to touch a little bit about the software because the attach rate has been above 50% now for some time. That seems very differentiated. First, if you could talk about how our competitors handling this. So instead of not having any software platform, and I guess I'm thinking about MCUs. Do you offer something that they absolutely don't have? What is the value add for your customers? What typical feedback do you get about design wins? We really love the software, what are the reasons? And then eventually, what does that mean in terms of ASPs and gross margin relative to just the hardware.

Yes. I mean we started back in 2019. And we said, hey, we've got to invest more in our software solutions. We saw the value of our software. And so as a percent of R&D, you can see that the software portion has been increasing relative to the hardware, but both are very important to us, but we've been investing more in the software. We build these software solution stacks and the software solution stacks are intended to help our customers migrate to Lattice easier. So as customers are using a different solution, whether it be an MCU or another FPGA and hey, I'm really interested in taking advantage of the innovative and differentiated products that Lattice has, but I have never done a Lattice device before. Well, these software solution stacks help them through that process. It simplifies it. There are prebuilt applications, there are software kernels that they can leverage within their system and helps them migrate to Lattice much easier. It helps our existing customers as well that are familiar with Lattice with their applications to be able to get to market faster as well. And we have data that shows that customers can accelerate their time to market 3 to 6 months, leveraging our software solution stacks. So these software solutions are driving value for our customers. And as such, when we analyze the ASPs for the similar device with customers that have leveraging our software solutions act and those that don't, we see that the average ASP is higher. And you would expect that because we always talked about our pricing methodology being around value-based pricing. It's not a catalog. It's about value-based pricing. And so that also drives stickiness for multigenerational design. So it's an area that we're going to continue to invest in. And at the Investor Day, we actually announced that in Q3, we're going to be launching our fifth software solution stack, which is around the automotive market called Lattice Drive, which will be coming out in Q3.

Great. Do you feel that there is more to do in terms of software? I know in a different market, in the high-density FPGAs, there's definitely the issue of not having a software ecosystem in place? Do you feel that there is -- that you're really filling the gap? Or do you think there is much more to do in terms of driving even much higher adoption as your software brings more people to design with FPGAs?

Yes. So one of the good things about the introduction of our mid-range platform is when we look at the target customers for the mid-range devices, 90% of those target customers are already existing customers of Lattice. In fact, when we introduced our first Nexus device, the platform back in December of 2019, those same customers told us, can you extend your power efficient architecture into the mid range. And we said, sure. But are you going to be with us on this journey? So we engaged with over 100 different customers on really defining this Avant platform that we launched at end of last year. So the customers are the same. 90% of the targeted customers are existing customers. They helped us define it. We launched the device. And the other beauty is the software that we have, our software solution stacks as well as design are compatible with both the small and mid range, which makes it easier for them to adopt it. We will continue to invest in our software. If you look at even the software solution stacks that we introduced in 2018, 2019, 2020, there's consistently on an annual basis or twice a year, but we're adding more and more to the exact same solution stack, but we're also introducing new software solution stacks at the same time. So we're going to continue to invest because we and our customers see the value of the software solution set.

Great. And then I've been getting questions over the past quarter about market share. How should we look at your growth relative to FPGAs from other companies? And I know it's not apple-to-apple because we're looking at high-density FPGAs, but it's a different market, and I think they're probably much more lumpy, so they may benefit in some years from some growth and other years might be more weak. And then people are trying to also compare your growth with MCU. So any commentary about how you gauge yourself relative to those other class of semiconductor? And what does that mean for your market share? How do you feel you're growing relative to your...

Yes. And like you said in the beginning, it is hard because other companies aren't public about what the revenues are in small, mid and large. But the way we look at it is we have a $10 billion SAM or market opportunity for Lattice, which means we have lots of headwinds to continue to grow. If you look at our small FPGA, that's had strong double-digit growth for the last couple of years, and we expect that to continue. On top of that growth, we're now -- we've introduced our mid-range platform, which is additive revenue on top of that. It doesn't cannibalize our small FPGA revenue. So we expect our small FEG revenues continue to be double digits but now you've got additional ramping of mid range on top of that. So that's from a revenue perspective. From a share perspective, the way we look at it within Lattice is we're growing our top line based on 3 different areas. The first is we are taking share from traditional FPGAs. We have got a good market opportunity, and we're taking share from other FPGAs, small- and mid-range FPGAs. The second is, I would call it, opportunity conversion or SAM conversion, where they were using an MCU or some other device or an ASIC, as you had mentioned earlier, and we're showing them that now, they can actually convert those into an FPGA because they're adding more capabilities and functionality. And the third area that we're seeing growth is we're just sitting with our cut. Our intimacy has never been as strong as it is now with our customers. And we're identifying new ways to leverage an FPGA in new applications, new functionality that did not exist before. That's the third area. So it's traditional FPGAs, conversion of other SAM into ours or opportunity and an expansion through new functionality that customers are happening to their system for the very first time, and they're realizing that FPGA can do that very well.

What would be the appetite longer term to move even into higher density? I know that if you move into higher densities beyond Avant, you're probably targeting a different class of customers. But at the same time, there's been arguably a tremendous market vote given the acquisition of both Altera and Xilinx and not only in terms -- but also in terms of end market focus. So are you getting interest for new customers or even maybe existing customers about moving even higher in terms of density? And I know your plate is full already with Avant, but longer term, how can we see that in terms of positioning? And how do you see the opportunities there for higher -- even higher density?

Yes. I think customers see our power-efficient architecture and are always asking for more. But when I -- when we look at -- technically, can we -- can Lattice build a large FPGA? There's nothing technically that says we can't. We have FPGA technology, we have FPGA know-how. But if we look today at our opportunities, $10 billion, we've got small FPGAs growing double digit. And by the way, our Avant, our mid range, we just launched it. We said the initial revenue will be the end of this year, it becomes more meaningful in the following years. There's lots of headroom to grow for mid range. There's still plenty of headroom to grow for small. Our focus today is to continue to focus the company and grow the company in our small and mid range. But there's nothing technically that says long term -- how you define long term that says we can't build it if we chose to. But we have plenty of opportunity today with the market opportunity of $10 billion by 2028.

Okay. Great. And then -- on the last earnings call to one of my questions, I think you mentioned that Lattice does not engage in many LTAs, which is really a plus, I think, particularly relative to what semiconductor suppliers with different architectures are doing. How does that benefit you medium term, particularly if we see eventually more excess supply at some of your competition or if the market was to slow down? How do you look at this? And what's the customer response to that?

Yes. So I mean, you're right. I mean, we don't engage in those type of agreements. But I mean, the way we look at it is that our products with the rapid portfolio expansion that we've had on our Nexus platform, the launch of Avant, Esam talked about our software solution stacks. We provide products that helped our customers differentiate and our products provide significant capability and functionality for our customers. And so that's what we're focused on providing from a supply perspective and working with our customers, we just want to make sure that we can provide the products that our customers want. And that's what we're doing. We're developing those products. We're -- we talked a little bit ago about how we're increasing our inventories or have been increasing up to Q1 this past year to make sure that we have the supply that our customers want, and that's what we're focused on.

Yes. You're generating a substantial amount of free cash quarterly. How do you think about capital allocation, balancing reinvestment with return on capital -- and how much reinvestment do you envision the business requires on the ongoing basis.

Sure, sure. So from a -- we did put out a free cash flow. I put that out there target, free cash flow margin target at our Investor Day a few weeks back of greater than 30%. And cash is an area that we've been focusing on for quite some time. And the business generates a lot of cash. I like that. And I think our investors like that, too. And so that's why we put a target out there of greater than 30%. From a capital allocation perspective, our first priority is investing in the long-term product portfolio. I mean, all of our products that we've announced in that rapid portfolio expansion is a testament to that. That's our #1 priority. Of course, also investing in demand creation, customer support, all of that is very important to us, and we'll continue that focus. The other area that we've made a lot of progress in is, we paying down our debt. Our leverage ratio is well below 1. It's almost nonexistent. We do have some debt on our balance sheet, but not very much. And so last quarter, we paid down $25 million just at the end of the quarter. Subsequent to the end of the quarter, we paid down another $60 million. So we just wanted to reduce the interest expense burn on the P&L. And so that's something that we'll continue to look at in terms of determining the best use of our cash. We've also returned capital to our shareholders through our share repurchase program. We've repurchased 3.7 million shares of stocks and that has reduced our dilution by a little over 2.5%. So our investors tell us they like that, too. I like that as well. And so we've made good progress there. I think it's about 11 quarters. I think Rick will tell me if I'm right on that, sequential quarters of executing on that share buyback program. We do have $120 million of authorization remaining by our Board. So that's something that we'll continue to look at on a quarterly basis. And then, of course, M&A, we evaluate that, always making sure that we're -- we want to maximize shareholder return. And so we look at opportunities that are complementary and adjacent to our core business. We did buy Mirametrix in November of 2021. It's part of our software strategy. And so we continue to look at all of those elements from a capital allocation perspective.

Great. And then in the last 35 seconds -- sorry, maybe if you could crystallize some of the key thoughts we should be leaving this room about what to expect and what's the key points of the study.

Yes, I think in the last 45 seconds, I'm going to go back to our product portfolio expansion. The most rapid product portfolio expansion we've had in our history and all the benefits of that. It's in the hardware and software side. We're positioned in really good long-term secular growth markets. And we have on top of that Lattice specific growth drivers that make us unique from the overall market perspective and that our customer intimacy has never been as strong as it is now. The results of all of that is the increased financial targets that we shared at the last Investor Day.

Great. Well, thank you so much for presenting with us. Thank you.

Thank you, Tristan.

Thank you so much.