All right. Welcome back to the Barclays Tech Conference. I'm Tom O'Malley, Semiconductor and Semiconductor Equipment Analyst here at Barclays. Happy to have John Forsyth and Ulf Habermann.
Yeah.
All right. Good. CEO and CFO of Cirrus Logic. Thank you so much for being here.
Thank you, Tom. Yeah. Great to be here.
Appreciate it. And I appreciate you guys coming over the last couple of years. It's been great to kind of watch the content story grow. So why don't we start with kind of that framework where you've talked kind of over the years about adding on the audio side. You've added on the power side. And I think you've always done a very good job of kind of dropping hints along the way about the content roadmap in the future. Let's talk about where you see content expansion and where you're focusing your efforts right now from an engineering perspective.
Yeah, absolutely. So maybe just to tell that story of how we've expanded content over time, in particular in the phone space, that went from being a codec to a codec plus amplifiers, all in the audio space, to then haptics on top of that, which is really a derivative of the audio amplifier, to then adding camera-related content and then, more recently, power-related content. So we've consistently grown content over the long term and deepened our relationship with our largest customer, but also really expanded the number of domains and different technology areas that we serve and where we develop innovative IP. So as we look forward from here, and speaking specifically about the phone space right now, we still believe there's lots of opportunity ahead for us to continue to deliver more value through both iterating on content areas that we serve well today and pursuing new sockets.
So from an engineering perspective, a lot of what we're doing kind of falls into those two categories. One would be where we take an area like the camera functionality, for example, where we play an important role in helping to enable some of the advanced camera features that tend to get kind of talked about a lot in the launch keynote. And we still believe, and when we look ahead, that there is a huge roadmap of further enhancements and additional features that we can deliver there. I would say that our customer has no shortage of ambition for what they want to do in that space. And the level of engineering and innovation within that customer is phenomenal around the camera. And so we want to be a very close partner in that. And today, we have multiple developments focused on delivering new camera chips in the future. Beyond that, we really believe that power is a space where we can bring innovation and continue to add value to our customers' products.
So something that comes up very frequently these days, and you guys have dealt with it more than anyone else over the last decade, is just the view that over time, Apple will integrate more and more and more. So there's conversations about the assets that they bought from a modem perspective. There's conversations about potentially them integrating something in terms of Wi-Fi this coming year. So the natural follow-up is always from someone who hasn't looked at your name before, what is the value that audio brings and why wouldn't you integrate that? Could you just remind us where that value creation is and kind of how you work with Apple and why that situation maybe is a little bit different than those others that I mentioned before?
Yeah, yeah, yeah, absolutely. I mean, it all has to be caveated by saying that they're one of the most powerful companies in the world. They can basically do whatever they want. But I think there are two threads to what you're highlighting there. One is integrating from a technical, from an architectural perspective. The other is bringing something in-house, bringing a development in-house. I think there's something important to say about both of them from the point of view of understanding Cirrus and how we think about our relationship with the customer. So in the first case, you're really talking about functions. There are functions which over time naturally gravitate towards one or other of the big chips on the board or a chip that, for example, our largest customer is making today. And a lot of the time, that's because of some compelling architectural reason.
For example, if the function is purely digital and there isn't a very particular reason why it needs to live in a certain part of the system, it will ultimately make a lot of sense for that to be integrated into the main SoC because that is where transistors are most abundant and cheapest. By contrast, though, if a function is something that needs to live close to the analog-digital boundary or where there's a very, very tight latency constraint, then it will tend towards sitting in the edge of the system, and a really good example of that is the camera content that we have, where if we're doing autofocus and stabilization, stabilization is a process of taking inputs from sensors which tell you how the device is shaking.
As rapidly as you can, providing corrections to move lens elements and the sensor to compensate for that, you absolutely do not want to send that information on a round trip to the main processor. You'll lose time. The more times you can do that per second, the better your stabilization. So that's a function that has to live at the edge. So that is a very good kind of example of the way we look at the things we do. We ask ourselves constantly, does this need to live close to that analog-digital boundary? If it doesn't, then you can't really build a plan around it for the long term. You may have a function there for a while, but we build our product strategy around things that need to live at the edge of the system. So that kind of accounts for the first part of it.
We specialize in things that don't naturally integrate elsewhere. And the mixed signal aspect of that is very important as well because when you have analog circuits as part of your chip, there's a very different calculus about what the optimal node to be on is. It's just not true that you would be better off on 3nm than 22nm, for example. The decision about what the optimal node is is determined by the ratio of analog to digital. Digital shrinks really nicely. Analog doesn't. Analog gets more expensive as you go down the nodes. So that ratio determines what kind of process you're going to use for the chip. The reason that matters is if you think about that, it means there are a whole load of functions within the system which will not gravitate towards the finest geometry.
And in fact, there's a strong case for there being other chips at different nodes in the mixed signal space where functions get integrated. And that's actually so rather than being integrated into other stuff, we integrate functions into our chips. The codec is a great example of this, which the one we just launched is at 22nm. And it used to be the case we talk about the codec as an audio chip. We bucketize it as audio revenue because it's always been an audio chip. But with the current codec, more than 50% of the silicon is not actually doing audio stuff. It's integrated a whole range of other mixed signal functions because that's a great node to integrate mixed signal stuff on, just in the way that 3nm is a great node to integrate pure digital functions.
So I feel like I've been speaking for a while about just the first part of your question there. But the bottom line is we look for compelling logic as to why functions we deliver need to live outside of other parts of the system. They need to live at the edge. And if we can't identify that, then we're very unlikely to invest in it. And the stuff we invest in, we feel, is defensibly at the edge of the system. Then you have the question of, OK, but could your customer or any of our customers just do that chip? And yes, again, I mean, they could. Theoretically, they could. But that's where our execution record really shines. And that's really what that's about. We have, over very, very many years, executed at an extraordinarily high level with that customer.
And I think the greatest compliment I've been paid by not just our largest customer, actually by any customer, is an executive in that company said to me, "the thing we really like about Cirrus is the deterministic outcome with hard problems." And you don't get told that they value the deterministic outcome because you did it well once. It's because you've solved hard problems again and again and again. And when you give them a plan, they feel they can take it to the bank. And so if you have a supplier who can do that for you and executes so well, I think your motivation for disrupting that and saying, "hey, why don't we bring that in-house" is very low.
You said you were talking for a long time, but you've done that for me before. That was the best time you've ever done it. Just for the record, that was a good explanation. So in terms of the content roadmap, so you talked about where those two worlds meet. You need to be very close to the actual processing. So you mentioned cameras first. So I understand camera proliferation right now, where if you look at different models of phones, you have as little as one, up to, I think, four is the most now. And so I understand that over time, you'll see more content there. You've got power conversion and control. You have the amplifiers. You have haptics. And then you've got the codec, right?
So when I think about different applications in the phone that would be the next spot for something like this to converge and for you guys to have more content, I run into a little bit of a wall where I'm like, again, this is not my field either, but trying to understand exactly where in the mobile device that may happen. Maybe you could walk us through. You obviously may not even know the next area, but some of the areas that would have problems like that in a smartphone. Is it something driven by AI? Is it more power-related things? Anything to help us with where we should be focusing on for potential next solutions?
Yeah, I'll give you a couple of things. One is in the power area where I think I've made no secret of the fact that we're investing in that area in developing power IP. And it's typically not in more conventional power management stuff. It's, in particular, stuff around the battery. So the power conversion control chip we have today sits between the battery and the rest of the system downstream from the battery. But we really believe that applying higher precision sensing to the battery improves the knowledge about what's going on inside the battery because the state of the art there is not that great. And it could stand to be improved a lot. And that matters when you want to extract maximum efficiency from batteries and minimize how you age them, how rapidly you age them.
So we believe that sensing more precisely what's going on in the battery and using that to inform other parts of the system, including how we charge the battery and so on, is a great set of opportunities for us. They sit in this sweet spot where they require really high precision analog and a bunch of logic and processing alongside them. And so that's a great example of stuff where we believe, and you can see on any teardown, there are other chips that sit around the battery. We believe we can offer something innovative in that space in the long run. And then you mentioned AI. I think there's an area where we don't know exactly how this is going to land yet in terms of features or evolutions of products we have today. But I think AI is really going to expand the use of the voice interface. And every time you speak to your phone or listen to your phone speaking back to you, those signals, that information is passing through our silicon. And the better job we can do of that over time, I believe the more we can enhance the AI user experience.
Super interesting. I remember years ago, the voice biometrics, maybe a different take on the voice biometrics kind of angle in terms of your interaction with that assistant is probably going to increase drastically.
Yeah, yeah. I think a lot, and we see that. I mean, that's something that we see not just in the phone space, but with talking with our customers in the PC space now as well, and really there have been very, very limited examples of people talking much to their PCs. I mean, of course, you had things like dictation and so on, but things like Cortana never really took off. There wasn't that much you could do that was super useful with it, I would maybe say, but I think that's about to change quite radically. With Gen AI, it seems obvious to me that the conversational interface is a really natural way of interacting with AI assistants in general.
If you use something like ChatGPT and you ask it to create the first draft of a thing for you, whatever it is, and it maybe gets 70% of the way there or something, it was a good start. The bit that really grates is typing into the prompt to try to refine it again and again, and it seems like that is exactly when you want to have a very natural conversation, as if you were talking to a coworker about which bits are great, which bits work, which bits need more attention, and so on, so I think, again, I think we can be real enablers for that kind of voice-driven interface in the future.
Super helpful. So other areas, just moving away from mobile for a second, we'll circle back. But other areas of content growth, you've talked about PCs. You've obviously found yourself on a reference design this year for an AI PC, talking about, I think, tens of millions of dollars kind of exiting the year in aggregate. And then also, you're hearing more like at a Qualcomm Analyst Day I was just at just a few weeks ago, they talked about an XR, AR, VR market that's going to be on the order of $2 billion for them by 2030. So one, maybe spend a little time updating us on how things are going in the AI PC world for you guys. And then two, where would you find yourself in some of those opportunities? And have you already found yourself in some of those opportunities on the XR, VR, AR?
Right. OK, well, and the PC, our kind of entry into the PC market, we're continuing to track to that. I say, yeah, low tens of millions of dollars in this fiscal year, which obviously ends at the end of March. And beyond that, we're anticipating a significant step up from there. And then even as we look out further, we anticipate a really healthy growth rate for us in the PC part of our business. Being on the Lunar Lake reference design certainly looks like being a big catalyst for us. We expect to see a lot of devices coming to market during calendar 2025. I mentioned in the fall that we had over 100 PC laptop designs in development with customers. And the vast majority of those would be coming to market in the coming calendar year. Our content there kind of, again, spans at least audio, power, and haptics. We feel very, very excited about that opportunity and continue to kind of gather momentum there. On the XR, VR, AR stuff, yes, I heard those comments. I'm excited about it. I was excited about it the last time and the time before that as well.
That's a great story.
So you know it's definitely the case that AR and VR exacerbate some of the drivers that have helped us be successful in phones. The smaller and the more power efficient you can make your chip, the better. That becomes a kind of extreme constraint when you have to wear the thing on your head, right? And so that's a great example of something where I think is very favorable to Cirrus. We tend to be invested at the leading edge of technology nodes for the kind of chips we make. So there were times when our competitors can win on price with a bigger package and maybe less power efficient. But if you care about power efficiency, performance, and the physical size, then we're going to have a big advantage.
AR and VR are absolutely things where we can deliver a lot of functionality and a lot of benefits to our users. We have, to date, been in most of the AR and VR headsets that have been launched on the market. Mostly, to date, not entirely, but mostly, to date, that's really been kind of off-the-shelf parts, which is often how it starts with new product categories. More recently, over the past year, we saw a product come to market with some custom silicon from us, which is highly optimized around some of the specific VR features there. I think, I mean, we still, I guess, it wouldn't be the biggest needle mover in our kind of five-year revenue plan. I would love it to be. I think we're incredibly well placed for it. But to me, when I look at something that weighs quite a lot, we're still some way from that being quite mature enough to be super mass market.
Yes. And we'll get to the highest volume area of your business in a second. But there was one more thing I wanted to touch on in terms of growth drivers. I was laughing with Chelsea. I saw your clocking press release. I was like, wait, you guys have a clocking business? It had evaded me for years, apparently, and I had forgotten about it. Why now? Why do you put it out now? I think that the two applications that you mentioned were automotive audio and pro audio systems. So what are you bringing to the market there? Who are you competing with? And why the timing?
OK, so yeah, one of the things that's less well understood about Cirrus, probably because we don't talk about it a great deal, is we have a chunk of our business which looks much, much more like a conventional analog business. And it follows the kind of pattern that those businesses have. It's got many customers. It's very broad. And typically trades at very high gross margins. And the products live for a long time. It happens to be the case. So when we look at that, and to be honest, there were many years when we were just kind of sprinting to try to keep up with our largest customer. We didn't pay a ton of attention to that part of the business.
But when we did pay some attention to it, the interesting thing was that there were some things within that segment or that part of our business which really, without any attention from us, were growing. And one of them, in fact, was timing related. We have a very good, but honestly pretty old timing product that was growing. And we wanted to understand why. And one of the main drivers was to do with Automotive Ethernet. Now, one of the things that has happened in the kind of decade and a half since we produced that product is we've jumped many technology nodes and now have a whole bunch of IP on significantly more advanced processes, which can improve the performance and the cost of those typical kind of products. So from the point of view of continuing to build our long-term opportunities in other markets beyond smartphones, we are consistently looking at these segments that we serve today and looking for what's growing and making some investments and gauging the customer reaction and building from there. And in the case of the timing chip, we've had extremely good response from customers. We will see that mandated in some automotive platforms going forward.
Is that 1 GbE , 10 GbE ? I would assume that it sits with the digital chip, right? So is it a clock? Or is it an oscillator? Or is it a combination?
It's a clock. Yeah, clock recovery. Yeah.
OK. Very helpful. All right, Ulf, switching over to you and some of the moving pieces on the model, so at last earnings, I think most of the ecosystem sounded a bit more sanguine on both the December and the March quarters. There were various descriptions of why that was the case, but you guys had a dynamic with last year versus this year and the timing of the ramp, and then some commentary on just the back half of this year as a result of that. Could you maybe walk through, is your view of kind of the December and March time frame still similar to kind of what you laid out? And are there any other moving factors other than that timing shift that are leading to kind of the trajectory that you laid out?
No, I mean, no change at this point in time. A bit early to tell, too, right? It's driven by smartphone unit volumes for both the December quarter and the March quarter. But yeah, as a reminder, right, we had an additional week last year. So that makes year-over-year comps for the December quarter a little bit harder. We had some Android content ramp early last year in the December quarter. That was one of the factors. And with that extra week, we actually shifted some of the demand more into the September quarter. So that was kind of the drivers for the shape of this year. And yeah, too early, a little bit too early to tell. Typically, if we see volume adjustments, we see them a little bit later in the quarter.
Just remind me, in terms of timing of shipping into that ecosystem, you guys now offer a wide variety of parts to that specific customer. Some are module related. Some are just going directly on a PCB. Could you remind me of the timing differences there between products?
Sure. Yeah, yeah, yeah. So the majority of our content ships for new devices in the quarter when they ramp. So that for us is in the September quarter is when that starts. Camera controllers, since they do go on an individual module, they will ship a little bit earlier than that. So you could see some impact in the June quarter. But still, the majority would be in the September quarter as well.
You guys were calling out the difference in timing from the year-over-year perspective as the major difference, not anything related to the timing of the camera controllers this year versus last year.
Yeah, that's right.
OK, super helpful, and then on the gross margin side, kind of exiting 52% this calendar year, you guys have been pretty consistent with gross margins and really helpful. Like, this is how you should think about it from a longer-term perspective. It actually got bumped up a bit recently, so what are the big swing factors there? Is there really anything, and you have a high concentration with the customer, so you can't comment, obviously, on them specifically. But what are the moving factors outside of that customer that could move things up or down there?
Yeah, I guess, especially when you're introducing new products, there can be a fair amount of volatility around our gross margin, as we have been this past quarter. So obviously, mix plays a part. But in the long term, what drives gross margin for us is work on the supply chain side. I mean, with high customer concentration, you have to be very realistic, very pragmatic about your expectations of ASP trends and gross margin from that perspective. But that gives us a very, very strong incentive to work really hard with our supply chain team. And it's not all about managing down wafer prices. It's to do with yield. It's to do with having a super competitive back end amongst the OSATs and so on. And I think we really have a world-class team working at that. And over the past couple of quarters, that's driven an improvement.
That's contributed to an improvement in gross margin as well. So as we go forward, like I say, I don't think the gross margin profile for us differs in the long run in relation to our business with a high concentration with our largest customer. I do think as we grow in other areas of our business, in other markets, that in the long run can give us potential for gross margin accretion. Because most of those markets would typically have a broader spread of customer exposure for us. And some of them, like the kind of, like, for example, the timing chip and so on, just traditionally trade at significantly higher gross margins.
And then the other area from a financial perspective, I just want to drill in a little bit more, was you had a foundry customer, well, a foundry supplier that was introduced during the pandemic. There's obviously some inventory that's associated with that. Forgive me if this is a stupid question. But is there any way to redirect that? Or is that going to just be wound down slowly?
We've worked with GlobalFoundries for a long time, which I think is the part that you.
Or at least was during the [crosstalk].
Yes. Well, yeah, we reserved. We did this agreement with Global, which reserved a bunch of capacity. I would say that I think that that was beneficial to Global. I think actually in the circumstances, we were in a situation where nobody had enough wafers, and coming out of that, having absolute confidence about our ability to serve our customers was actually hugely beneficial. As we look forward, I mean, you can see that we've got inventory level that's risen, and we've given some comments around that. It's not something concerning to us, given that when we ship a product into our largest customer's product, they typically ship for multiple years, and that's one of the things that I think, again, about Cirrus is not always perfectly understood, because we have this exposure to consumer products.
But it's far less volatile than your classic kind of consumer market, where you've got to, the race starts again every year. And you've got to rev the chip and so on. It's really not like that. With our last codec and last boosted amplifiers, they've been shipping already for five and six years, respectively. And will still continue to ship for some time. So there are situations where it's to our benefit to take on additional inventory. There may be some incentives to do that. There may be some benefit to the foundry as a partner to us doing that. And we have confidence that we know there's going to be demand for those products for many years out to the future. So if that makes sense to us, then we do it.
Smaller line item in the model. But on the mobile Android side, we had Qorvo on stage today talk about this bifurcation of the market towards the high end and the low end. And they're actually taking down investment in the sector. You obviously are very different than an RF guy who's broadly exposed. I think you have a very significant customer there. What is your feeling for mobile Android? Is that something that you think about growth over the long term? Or kind of continuing to serve the existing customer?
No, I think about it from the opposite perspective. I think it shrinks over the long term. For us, we don't make a lot of investment in Android. We still have good business there. We still have fantastic products. But it's really opportunistic. We'll win what we can win with those products. Investing in new product development focused on other markets is far more attractive to us. And it's just a better return on investment. But there's also, I mean, there's a kind of bifurcation that is not so much to do with the high end and the rest of the market. It's to do with the fact that you have a collection of OEMs in China who, I mean, honestly just want to make great products. And many of them use our silicon. But they live in fear of some kind of export control change.
They obviously have a huge amount of local incentive explicitly encouraging them to use domestic semiconductor vendors. And that just makes that a market where if we can win business, it's great. But it's not something to plan on for the long term. So we're fortunate in that, for example, the PC market presents us with, I think, a bigger overall opportunity in terms of SAM. And it's a real greenfield with dynamics which are very, very favorable to us just in terms of the architectural evolution and what's happening to the products there. So that's really where we steer our R&D spend.
Looking forward to seeing the PC success translate into some numbers and also the future of the mobile opportunity as well. Can't wait to figure out where it pops up next. Appreciate it, guys. Thank you for being here.
Thank you very much to you. .
You're welcome.