Good morning, everyone. We're gonna get started. Welcome to the Cirrus Logic session here at Stifel 2024 Cross-Sector Conference. My name is Tore Svanberg. I'm a senior semiconductor analyst. I cover analog connectivity and processor companies, and it's my pleasure to introduce Cirrus Logic.
With us from the company to my immediate right here is John Forsyth, who's the company's Chief Executive Officer. And here in the front, we also have Chelsea Heffernan, who runs Investor Relations, and also Ulf Habermann, who is also on the finance team. The particular format for this session is basically a fireside chat.
As some of you know, if you're familiar with Cirrus Logic, you know, they have a very large customer, and, John sometimes, is not comfortable talking about that particular customer. But, I'll, I'll try and dance around the questions as, as well as I can so we can get some information out of him. Well, then we get started. So thank you, John, again, for, for coming. Maybe we could just start with a very general introduction, for those investors that may not be as familiar with Cirrus Logic.
Yeah, absolutely, and thank you. Thanks, Tore, and thanks for everybody attending today. I want to begin by just maybe correcting you on one thing. I may be a little uncomfortable talking about our largest customer. You should see how uncomfortable I get if I talk about our largest customer in detail. That's not a good place to be.
So, I will, of course, you know, give as much color as I can on the business and our outlook. But yeah, just to get everybody on the same page regarding Cirrus Logic, we're a fabless mixed-signal semiconductor company based out of Austin, Texas.
We're about 1,600 and change employees, $1.79 billion revenue in FY 2024, which ended in March. The main sectors that we sell into, the smartphone sector, and then we have a growing business in PCs, and then we have a kinda long-tail business with which looks much more like a kinda traditional analog semiconductor business.
The main areas where we develop and sell products have traditionally been based around audio. So audio processing for smartphones has been the main revenue generator for the company over the last decade and a half.
We expanded beyond audio into a broad category that I refer to as high-performance mixed-signal. It includes a range of applications where we get to use some of the same, some similar IP and technologies to audio, but for other things such as haptics, camera control, and increasingly, more recently, in power as well.
Great. Thank you for that. So maybe I can just start talking a little bit about the industry, John, 'cause it's been a crazy few years in the semiconductor industry. I know there's always a degree of volatility in your business, just given your customer concentration, but how did these last three years impact Cirrus Logic from a strategic direction perspective?
You know, what are some of the things that the company did right when it came to trying to fill up all that demand, and then how did the company sort of react when things started falling off the cliff a little bit? Obviously, you know, Cirrus Logic did not have that same element of cyclicality as the industry, but it was still a very unique time. I wanna hear from you what the company did right and perhaps what could have done better.
Yeah, that's... Okay, very, very, very broad, very interesting question. I guess, okay, to hit on a few points, and we can just drill into any particular areas of interest. If you wind back the clock over the past kinda couple of years, we were in this situation, as many were, of demand vastly outstripping supply.
Part of our response to that was to invest more in our foundry partnerships. So we're fabless. We believe that's the right way to go for a company like us. But we put in place and invested in capacity reservation with foundries, where we were partnering on expanding the capacity there as well.
Put us in a position where we could give a lot of confidence to our end customers about what capacity we had, both for existing sockets and for new sockets, which was very, very important to us. There was a period when we couldn't win new business without being able to show that we had committed capacity coming online to service that business.
More recently, we've actually continued some investment in foundry partnerships with the goal of driving competitive environment there and also pushing the technology forward for us in the longer term, given some of the things we want to do.
So that was the period we were in. I think, you know, we... I feel that that put us in a good position. You know, I think we have to recognize the fact that we were, you know, we were short to the overall demand. We would've had, you know, bigger numbers on the scoreboard if we'd had more capacity.
But I think we probably, in comparison with many of our peers, benefited on the downside in that, our particular customer concentration is with a customer that actually has been, has ridden this cycle very well and has had comparatively resilient demand.
S o even as a lot of the semi peers that we look at have seen demand kind of really evaporate, and then large inventory problems, we have, you know, knock on wood, seen comparatively little of that. We've seen pretty sustained demand throughout.
It's maybe not as, you know, as kind of white-hot as it was during that period where nobody seemed to know where the ceiling was. But at the same time, we're also servicing a customer and a supply chain there that is not known for wanting to sit on a lot of inventory.
Right.
So I feel we actually managed to, and that's more good fortune than anything else. I mean, but we were fortunate enough to navigate and avoid some of the downside risks there. So if you then look at our, like, our last fiscal year, and then particularly the last couple of quarters, what we've seen actually recently is steady and sustained demand a nd that's really been the story as we reported the last couple of quarters.
To an extent, that was actually surprising to us. I think you know, we were outside our guidance boundary in both cases, on the upside. We don't aim to do that. We aim to provide a guide that embraces you know... When we guide, it's based on our most accurate view of what's likely to happen. But we came in higher than that, and that was really a reflection of just steady and sustained demand throughout the period.
So I feel like, you know, in more volatile, less certain times, actually a lot of benefits accrue from the kind of customer concentration that we have where it was very, very strong, resilient, baked-in demand for many of those core products.
Right. No, I mean, I know you have some exposure to China smartphones, but I mean, if you would have had higher exposure, you, you'd have had a much rougher time, that's for sure.
Yes.
So, that's pretty clear. So John, you've been CEO now for several years, and, you know, when you came in, I think there's always been kind of pressure, at least from shareholders, to diversify the business, especially when it comes to customer concentration. But when you came in, you basically said: "Well, you know, yeah, I wanna do diversification," but you did it more by technology, right?
Yeah.
So I mean, when you came in, it was predominantly an audio company. You've invested heavily in mixed-signal the last few years, and of course, you know, you've seen great results from that. What I wanted to ask you is, when you came in, what kind of foresight did you have? Because when you think about things like Camera Controller or mixed-signal technology, you think about power.
There's a lot of other companies out there that have those capabilities. So, I mean, did you get sort of signals from your largest customer to start investing in that IP, or did you say, "You know what? Here is a great opportunity for us because we have the underlying, or we can build the underlying IP and technology to be very competitive in some of those technologies?"
Yeah, that's, that's a great question. I think, I think it's a mixture of those things. I think, a lot of it, a lot of it really comes from taking, taking an idea to a customer and finding that you were maybe half right.
Right.
T hen they reveal to you: "Well, actually, there is an interesting problem in this space, but it's not quite that," and you struggle for a while to get your head around that. But eventually, if you're fortunate and you have smart enough engineers, you will identify areas where you have the opportunity to bring something different.
We can get more into that, but I yeah, I maybe just to back up a little, the path that we've been on and since I came into the role is really based around three things. So yes, there's been, there's always been a lot of talk and pressure externally for about diversification. Actually, mission number one is look after the golden goose.
I mean, we have an amazing customer where we've successfully continued to grow revenue and expand the number of sockets over time. At a certain point, I think. I think there has been more recognition of this. I understand that to some people, that's risk. There's a certain amount of risk associated with it. I think there's also a credible view that, you know, that's really one of the things that we're very good at, and kinda uniquely good at.
So there is some process power there that we have in terms of how we manage the relationships, how we uncover new opportunities, how we execute to service that customer, which is not for the faint of heart, to ramp a new product, to go into hundreds of millions of devices, in a single year without missing a beat again and again and again.
That takes an extraordinary infrastructure, a lot of which is not that visible from the outside. So mission number one was really about, okay, look after this, invest in it, grow it, grow it. We have an opportunity to continue to serve that customer very well and bring new innovations to them, and I think that's been very rewarding.
T hen the other side of it was, okay, yes, grow our SAM. T hen beyond that, I was very keen that we looked more at how we could drive profitability and free cash flow over time. Growing the SAM really came from, i t was certainly an overarching understanding that I and many others have had in Cirrus for a while, that there's probably not enough audio out there for the long-term aspirations of the company.
There's plenty of good audio out there, but we're, we're at the high-performance end, the low-power end. That means investing in leading-edge technology nodes. That means nothing we do is cheap. It's not cheap to do, it's not cheap to buy, buy from us.
So if you're buying it from us, it's because you really care about performance, size, power, and, you know, if you wanna grow your—if we want to grow our SAM, we're not gonna grow that by, you know, attacking audio for sound bars or something where, you know, a lot of them are built to a price. They're plugged into a main socket. They don't care about cost, power, sorry, power and size, enormously care about cost.
T hat was never gonna be enough for us. Now, we identified a number of areas of audio outside of smartphones where we did think there was very exciting growth opportunities, but it was very, very obvious we needed to expand into technology adjacencies.
So, to really long run expand our TAM meant building beyond audio. O ur approach to that has been to focus initially. So this is the high-performance mixed-signal kind of general category that I talked about a nd our approach has been to focus initially on developing those technologies in the smartphone space.
T hat's not for tremendously sophisticated reasons. It's simply if you think about that matrix of, like, existing products and technologies, existing markets, new markets, new products and technologies, we kind of thought we'd mostly try to avoid the suicide quadrant of, you know, totally new technologies in a market we know nothing about.
So we set out with this mission to say, "Okay, well, let's expand and win and develop IP in as many new high-performance mixed-signal areas in the smartphone as we can, and we'll develop a very strong base of IP and expertise that goes well beyond audio, so into haptics, camera, power, and so on, and then look to see how we can leverage those technologies into other applications and markets."
So we're kind of, y ou know, we're somewhere along that journey right now. We're still expanding HPMS in the smartphone space, but we're also seeing, you know, we're shipping products in power and haptics and audio into other categories, you know, with good growth opportunities around those, I think, for the future.
Great. So let's shift gears and maybe make you a little bit uncomfortable, not very uncomfortable, a little bit. So, and related to what you just talked about, you know, there's years when you have new product ramps with your largest customer, you know, you have some content expansion.
You know, remind us for this particular year, I mean, there's still, obviously, there's the new smart codec, new boosted amps, but remind us where we are for sort of content growth this year, and, you know, can you maybe give us a bit of a preview for what you're expecting next year?
No, on the latter bit, b ut I will talk about one. There's specific new content this year, and then there's also a kind of tailwind dynamic that we have going on around the camera stuff. So, maybe starting with the camera stuff, we call the category of product that we sell into camera modules camera controllers. In practice, what they do is drive autofocus and stabilization systems, other lens elements, and so on.
The reason we're kind of exceptionally good at that stuff is those are applications where you need a loop from the analog world (just to get slightly geeky for a moment) from the analog world into the digital world and then back out b ecause what's happening is you're taking some analog input, like the shake of the device, so you're getting information from inertial sensors about that.
You're calculating how to move some lens elements or the sensor to compensate for that many thousands of times per second, and then you're sending a signal back out to some motor, some kind of motor, which will move that element in space. You may also be sensing where that element is in space so that you know deterministically where it is. S o for the longest time before we could talk about what this new product did, you will remember we referred to it as the closed-loop controller.
Right.
I think that was actually maybe in our patent application as well. But that's what it means to talk about a closed-loop controller. So we're very, very good at that because we do very good analog to digital, we do very good digital to analog. We are at an advanced process node, which means we can pack a lot of digital processing in, so more corrections per second, and we can pack that into a very, very small, compact package.
A s you can imagine, the camera modules in smartphones are extremely space sensitive. So, okay, the reason I was explaining all that was, we have this kind of long-run, favorable dynamic around camera controllers. We introduced our first one, I think, in calendar 2020.
In the fall of 2023, we introduced the third generation of our camera controller. So there have been, and each of those camera controllers has had a kind of modest AI-ASP increase, but the total value of our camera content has continued to rise because the attach rate grows as well.
T hen, as we see features cascade from the kind of top SKUs down through the range, there's been a very kind of healthy tailwind for us, where we just see this kind of. It's not perfectly linear, but it's been a pretty steady growth every year, almost regardless of whether we've got a new device coming in that year or not.
But that will be reflected in this fall's content for us as well, in that we had a new Camera Controller that came out in the fall last year. We'll see that when that comes in this year, that will now be spanning two generations, and so on.
So we get some accretion from that. W hen I look forward, we have a very healthy roadmap around the camera stuff. Kind of expect that to continue in the future along those lines. We don't really see, as yet, a kind of limit to our ability to improve performance, add features, and grow value there.
Then we have all the audio and haptic content that we have coming this fall. So we have two major pieces of content coming. It's a smart codec and a boosted amplifier. The boosted amplifier, typically, there's three of those in a device, in each of the devices. So with the smart codec, there's one.
The smart codec is responsible for a lot of A to D, D to A, routing, mixing. Kind of as an interesting aside that I'll just bookmark for now, increasingly over time, it's doing stuff beyond audio as well. So it's very interesting seeing that going on with one of our chips, where when we look at that chip now, honestly, more than 50% of the gates of it are for non-audio, kind of other mixed-signal stuff, which we regard as great.
But that is—that's like the central. That was really the first socket we won, way back, in the early stages of us being in this device, before anybody was even using boosted amplifiers. That's the kind of central audio and mixed-signal chip in our phone. They've run for a long time now.
So as we get into replacing it, it will have been running for 5+ years, the one that we're replacing. T his year, we've got a particularly kind of interesting upgrade in that it takes us from 55-nanometer to 22-nanometer as well.
So it's a very significant investment, but I think it's always worth kind of emphasizing this point because there are aspects of our business which, although we're selling into a kind of consumer end market, they don't look awfully like a lot of other consumer end markets. They happen to not even look very much like the rest of the smartphone world in that, you know, we knew we'd won that socket a long time ago.
We developed the chip. It will go into devices. It will ship for multiple years. We know that a nd we can kinda put some, you know, some pretty decent... Because of the nature of the products that it goes into, some pretty decent estimates on volume over that time as well. So actually, we have kind of reasonably high confidence in the demand that we see against it over, you know, half a decade plus.
T hat also enables us to drive great leverage from our R&D teams because they can go on to work on other new innovations. Boosted amplifiers, there's a lot of innovation there as well, although it's not a node transition. It's still on 55-nanometer, which is probably more appropriate for a boosted amplifier.
But they're responsible for driving the loudspeakers. There'll be two of those for stereo playback, then it's also used to drive haptics. So there's a third one, which drives the linear resonant actuator, which provides the haptic feedback. So it used to be the case, way back when, I'm sure you remember this story, that there was a reluctance to change the codec and the amplifier in the same cycle.
It so happens this time, after many, many years of planning and test silicon and so on, everybody is comfortable enough, with our execution and, and the technologies and, and the silicon that we've got, that, that we'll see both of those, chips, get a, a meaningful upgrade this fall.
As we get into the next cycle, you know, we will again see a kind of good tailwind effect of that. So there's a tailwind effect of the camera content that I talked about, but also when we introduce a new codec and amplifier, you would expect that to have some ASP accretion. As we go into the fall of next year, that then becomes, we'll see that ASP accretion across two generations. So we get kind of a second wind from that on the next cycle.
Great. N ot to sort of overhype things, but, I mean, you know, everything is about AI these days, and, you know, everybody thinks about NVIDIA as the poster child for AI, for obvious reasons, you know, Microsoft. Very few people have actually talked about your largest customer in the context of AI.
I personally think that they are gonna be a leader in AI for edge devices. I think they intersect really well with your IP, right? Because the ultimate interface for AI for a device would be voice, which obviously means a lot of audio and voice technology and IP. But there's also clearly an intersection with AR/VR, and, you know, that's where your camera controller is an essential IP block.
So again, without pre-announcing, you know, what your largest customer is working on, at least just from a big picture perspective, can you talk about how, you know, AI and AR/VR in devices, whether it's laptops or smartphones, could potentially impact your business going forward?
Yeah, that's a great question. I mean, I share a lot of your thoughts and expectations there. I think one of the great things - One of the things that our largest customer is really uniquely good at is distinguishing, whereas, is identifying what a thing might actually be for, that's useful.
Right.
Once you get past the kind of excitement over the kind of flashy new technology, I think they, I think they put an extraordinary amount of effort into only delivering things, or really trying only to deliver things with genuine use, to them.
I don't, I don't see that approach changing, and I think that will be their, their, their approach in, in AI as well. S o, yeah, our thoughts about that and how, how that benefits us, firstly in the smartphone or out in any kind of device, I do believe, like you, I think, I think voice is, is gonna be a central part of the experience of, of having, of using AI.
I'm personally—like, I would love to have a genuinely kind of intelligent, useful, smart assistant that I could use via voice. Kitchen time is a great, you know, there is, there is so much more that can be done there a nd I think voice is just such a natural basis for that interaction.
If you wind back, you know, about a decade now, to the era when the voice interface was, and voice assistants were, were just kind of coming over the horizon or in their infancy, say, we did see that was a period of quite a significant amount of innovation and increase in requirements around the voice interface. Typically, those landed, many of those landed in the codec. For example, that every signal that is coming in through the microphones is passing through our codec.
So it's a great opportunity to enhance value. I think that can apply across a range of devices. So, yeah, I agree, in AR and VR, that's one of the areas, you know, when we talk about... I talked a lot about technology here, in the smartphone space, but when we look at the other markets where we have content growth in both audio and high-performance mixed-signal, you know, laptops is a key one, but AR and VR is another one where we've been delivering both audio and high-performance mixed-signal.
Yeah, again, like I see that as a great, great, great opportunity for us. I think our AI can drive a lot of use cases there. Where we fit into it is not necessarily providing the AI experience, but, everything, everything that I know about AI at the edge suggests to me that it's going to consume a lot of power and greatly increase the pressure on battery life, battery size. In a way, it's kind of been sort of more in this part of the S-curve lately, the pressure on batteries.
I think that that's gonna increase again, which means that if you can deliver any given function at significantly lower power in a significantly smaller package, you know, that's a double win because you're creating more space for a battery, and you're consuming less juice from the battery, and that's really exactly what we do. Everything we do is based around those principles, and you only get to do that by investing at the leading edge, you know, for analog mixed signal.
Right. We only have three more minutes. I wanted to make sure I address any questions from the audience before we wrap things up. Any questions? So, just a few more, John. So, interestingly enough the company before you here at the stage was Texas Instruments, and they were talking about how, you know, the reason why they do manufacturing themselves is because they worry about the cost curve of the foundry industry.
Y ou know, TSMC is your largest foundry. I know you also have a North American foundry that you work a lot with. How do you see the supply side of the economics for fabless companies going forward? Because, you know, there is sort of a thesis that, you know, wafer pricing is gonna remain inflationary. It was very inflationary, obviously, the last few years.
Yeah. Yeah.
It's gonna remain inflationary. Then, obviously, you know, you have one large customer, and I'm sure they obviously would expect some sort of price declines over time. So how, how should we think about those dynamics?
I think the idea that it's long-run inflationary, like, that's a huge bet to me a nd if that were true, I'd be worried about everything, actually.
Right.
Like, the entire kind of foundations of the products that our products go into, I think, you know, the markets have been built around expectations that certainly they're not long-run inflationary, that you can get improved performance over time for a given amount of expense. I think we've been through a disruption.
Part of the reason why, like I say, that's a big bet, I also doubt it. Part of the reason I say that is there are some basic economic principles there, which, like, I don't see how they go away which is that it costs money to build a fab, but that gets depreciated, and, you know, there's a cost curve associated with the output from those fabs.
There is a slight difference in that, perhaps the demand for mature node chips exceeded. It did exceed for a while the capacity of mature node fab infrastructure. I don't know that that's a kind of— Like, it'd be very weird to me, and defy everything we know about markets, if that's a long-run truth.
So, you know, there's probably a little bit more activation energy required, as we have seen, in order to build out some increased capacity, but that capacity will ultimately gonna respond to the same, the same kind of laws of economics as we've seen in the past. They will get depreciated, and the cost per unit will decline over time for a given facility.
So, you know, I think we're very thoughtful about this. One of the things that we think is essential, that's a component of that, is having a highly competitive environment in the foundry supplier market, and that's where we make strategic investments, to make sure that for a given product, given set of technologies, we've got multiple partners who are hungry to work with us, to win the next socket.
Right. O n that note, I would also remind everyone that Cirrus Logic was the first-ever fabless company. So they clearly have a lot of history being a fabless company-
Right. Thanks for that.
... and k now their stuff here. So, we've run out of time, everyone. Thank you so much for joining the Cirrus Logic session. Thank you, John-
Thank you
... and the rest of the team for coming to our conference. Really appreciate it. Enjoy the rest of your day. Thank you.