All right. Welcome back. I'm Tom O'Malley, Semi and Semi Cap Analyst here at Barclays. Very pleased to have Michael Hurlston and Ken from Synaptics here. Thank you for joining us.
Cool. Thanks for having us.
Morning.
I like to start with a broader thematic because I think that most of what we hear today is around AI. Maybe start with we've seen $3 trillion-plus of announcements today. It's a big pie. People are attacking this in different ways. Some are calling this a bubble. Maybe your view on where we are in the spending trajectory? You've been around AI for a long time in various fashions, but where are we in the spending trajectory? Do you think that this is maybe an over-accelerated beginning of this ramp, and how are you addressing that as a company?
Well, I think first and foremost, thanks for having us.
Yeah. Thanks for being here.
And most importantly, I think this whole $3 trillion is all data center driven, right? We are at the edge. It's not my place to comment where the $3 trillion are in terms of being a hype or not a hype, or circular or not circular, and all kinds of discussions. We'll stay away from that. For us, the opportunity is at the edge. Even if you talk to data center companies or hyperscalers specifically, the world is going to be needing hybrid compute for AI. That journey is about to begin, basically, in terms of deployment. We are seeing the first incarnations of products in the realm of leveraging hybrid compute through some personal solutions like wearables that you've seen from Meta last week, from Google, and a few others in China. It's coming about.
And I think more and more as time progresses, what I feel is the world of IoT systems is going to graduate to being agentic AI IoT systems. Ultimately, some form of hybrid compute from data center coming in. And there's a good reason for that. The world of our personal interface solutions and personal solutions is going to evolve to being agentic AI as well. And so another place for hybrid compute to come in. And this for a few good reasons. Number one, the latency of the experience of AI. If AI is going to be the key element for the next set of human-aware, cognitive-aware interface solutions, then AI is the core element in that. And so for that, you have to have that ability at the edge. Number one, you have to have a latency that's tolerable.
And doing everything in the data center and relying on wireless networks sometimes can be challenging. And so putting everything in the data center is prohibitive to the experience. And the third big thing is privacy, privacy of data and how data is managed. And so if you look at all these equations, there is going to be a need for moving from data center compute only to data center plus edge compute. And that is what the opportunity for Synaptics is, given where we are in our product cycle, given what we have in our product cycle to offer to our customer base versus our peer set in the marketplace. And the combination of wireless connectivity with the processing architecture that we have in our Astra line of products, and then the ability to provide the physical AI interfaces through our human interface products like touch, voice, and vision.
I think that's where our opportunity is. Just earlier today, I was sharing this example. The edge AI train has arrived at the station, and momentarily, it's going to make motion in the right direction, and I think we are one of those trains, basically.
So when we've seen investment dollars so far, you accurately mentioned that this is in the data center. We've seen hundreds of billions of dollars of CapEx on a yearly basis go towards building out large systems. And you have seen this push towards the edge. You've seen glasses introduced. You've seen smartwatches in particular, two of the verticals that we see most of. But there's been a limiting factor in that you haven't really seen a critical application that's really pushed everyone to move to those devices. Generally still speaking, it's your smartphone, right? Everyone has a smartphone. And increasingly, you'll see more apps that drive more AI use cases. Now, is your play one that can address the smartphone, or are you more focused on some of these other edge devices? And does that transition need to move away from the smartphone for you guys to have success?
Well, I think we have all come to live with and love our smartphones, or in some situations, not so much love, but long story short, I think the user is going to demand a lot more cognitive deployment of the experiences, basically, right, and for that, you need AI. Today, at every turn, we look at our screen, basically, on the smartphone. I really think we would benefit if we don't have to do that. If we don't have to open up an app, if we don't have to kind of go down a path of punching certain things in the app to get to know what exactly is happening around us, and so that's the cognitive dimension. That is human contextually aware dimension. All of those aspects are going to be addressed in the new generation of devices.
For us, the benefit that we bring in accelerating that dimension is along the lines of the belief that the developer community is key to making that progress at a speed that is needed in the marketplace. We are, on our processing platforms, I can think of the only company at the edge that is open source and developer-friendly. The world of AI models is developing at a rapid pace. And you can't survive that pace if you have a walled garden approach in your software dimension, basically. And I think from that angle, we bring to that marketplace an advantage for the developer community to engage very easily, very quickly with our platforms and build out their end product and experience and go to the market at the application level. And so that's where I think we need to go. And I believe we are well equipped to go there.
Along with that, we have the right products. We have a product that we have already sampled to the marketplace, which is our first generation of highly integrated AI-capable multi-application processor core monolithic implementation, what we call as Astra family of products, SL2600 series products. It's got a vision processor. It's got an audio processor. It's got GPU. It's got our own NPU. It's got NPU that we have integrated that is Google's NPU just to help integrate the transformers that ultimately will facilitate easy integration of various applications and AI algorithms that come from Google's ecosystem, and we've got a general purpose CPU, and so I think it is a platform at the far end of the edge, which effectively enables exactly what I said, and then we are going to build out a SKU map.
We are going to bring out a connected Astra, which is going to be an MCU-class product imminently that's going to have the best-in-class wireless connectivity, Wi-Fi capability, best-in-class Bluetooth capability, a small NPU with some GOPS, 50-70 GOPS, and a low-power MCU. And I think along the lines, we'll also have a semi-custom MCU for a hyperscaler that comes about in the next few months. And so the software strategy supported by play on the processor side that is exactly what I believe the doctor ordered for the IoT systems to evolve to agentic AI IoT systems.
So it sounds like if you look at the portfolio, you have an MCU, then you have a processor, then you have this connected processor, and then you have a semi-custom chip. Can you talk about where you're seeing the most traction amongst customers? Are you introducing the low-end because you felt like you needed the low-end to get to the high-end, or are you maintaining the low-end because there's a lot of volume there today?
Yeah, I think excellent question. I think, so the SL2600 family of products is capable of going into broad market, basically, from consumer to industrial applications, right, because it's a performance-centric product. The MCU class of products are going to go into very battery-sensitive applications, right? It can go in your door lock. It can go into your wearable devices. It can go into your smart appliances where a lot of processing is not needed, but it's good enough for delivering a voice interface, a human interface that otherwise would be a punch button thing, right, and so I think that is the scale of the products that we are seeing, and so power-centric implementations, and then there is this MCU class product that goes into wearables, basically, where battery life experiences along with AI are very important, right?
And so not only that, ability to display, ability to capture what you see and post-process, all those aspects in a form factor, and also a battery life envelope that's tolerable by the human experience is very important. So that's where we're going as well.
Helpful. So let's dig deeper into your relationship with the large hyperscaler. It's known to be Google. How did that originate? What has it been like to work with them throughout this process here? You mentioned there's some Google IP that's in your solution as well. How does that help you go to market? Are you able to bring that IP to others? And then as you move to a more semi-custom solution, how do you see that relationship developing over the longer term?
Yeah, I think it's an excellent question. I think there's a long history with Google. Our processor team has engaged with Google for a long time. I think it dates back to the days when Google was building the Chromecast stick, right? And so the first set of products that came out were based on the same processor team's product delivery. And so recently, at the beginning of this year, we announced this deeper relationship with Google where Google Research and Synaptics R&D partnered to build out this whole strategy around the software ecosystem, the compilers, the multi-level interface representative compilers where you have one interface, but you have multiple compilers underneath it, again, making things very user-friendly, building out the architecture on our products, integrating the Google Coral NPU.
It's a RISC-V-based NPU into our silicon along with our NPU, putting together the entire fabric of multi-processor engines, the one that I referred to earlier, in a way that it's less about number of TOPS. It's more about processing multiple things that you would see in the pipeline from voice, vision, GPU in a power envelope that makes sense, in a form factor that makes sense, and so I think the entire architecture at the silicon level also was contemplated in combination of Google R&D team's interests as well as Synaptics capabilities on R&D side, and so it's a very deep relationship that started off as one from, I would say, the Chromecast days to something recently as building out the right processor software platforms for broader engagement with the marketplace. They believe in open source.
We are one of the few, or probably, I think, the only company at the processor front or the edge that believe in open source, and it was a perfect marriage for two like-minded companies, obviously, they being the powerhouse, we being the nimble, one very anxious and driving a roadmap that aligns with their goals.
So you have the relationship with Google, but you also just did an acquisition with Broadcom. And I believe that it gets you kind of industry-leading Wi-Fi technology. Can you talk about how protected you are in that technology for the next couple of years? When can others step in? And why was it critical for you to kind of future-proof yourself one and two generations?
Yeah, I think it's an excellent question. I think so, wireless connectivity is very essential, in my opinion, if you want to deliver a processor experience. There is no two ways about it. Our world, be it industrial, be it consumer, is going to be completely untethered, right? And if you're at the far end of the edge without wireless connectivity, you are going to make it very difficult for your customers to deliver a solution in a timely fashion. You're going to make it very difficult for your customers to build a solution that is going to be not R&D intensive, right? And so as a company, we saw tremendous value in bringing a solution approach to our customers by bringing the best-in-class wireless connectivity along with our processor assets and obviously the sensing and human interface assets.
So those were the drivers for moving in the direction of licensing the IP capabilities from Broadcom, which you know is one of the leading Wi-Fi, Bluetooth, UWB, GPS, GNSS technology providers in the marketplace. And if you look at our processor portfolio in the marketplace, the big ones, right, very formidable, have been in this business for some time, big incumbents, not so much leaning forward in the dimension of AI, have very little to no wireless connectivity. And it's a complete differentiation that Synaptics brings to the table. What helps us feel confident is our agreement takes us all the way to 2030. Not only does it help us on Wi-Fi 7 generation, but it also helps us in the Wi-Fi 8 generations. And meanwhile one or two of our peers that are doing some wireless connectivity are still stuck in Wi-Fi 6 generation.
We are already on Wi-Fi 7, right, so along with that, I think we made a strategic decision because I think licensing IP is one thing, but building derivative products that are targeted for the IoT market is very important because Broadcom, if you look at their assets, they're focused largely on access point and mobile phones, and so for us to get into the realm of building products that are targeted to IoT systems was very important, and from that point of view, we also struck the agreement to bring in about 130 engineers, wireless engineers from Broadcom into our team, basically, and so not only did we get the IP, but we got the capabilities to kind of build out derivative products and products that are very much required for the IoT marketplace.
And just one clarification. We own that IP for that IoT market. So we can continue to prosecute that well beyond 2030 timeframe because we own the design, the design databases, and the ability to prosecute it.
I think not to make some of our peers get nervous, but I come from the world of wireless, right? We are thinking about Wi-Fi 9 as well, right? I think this world of wireless connectivity for the reasons of the bandwidth in congested environment, if you think and believe, and I do, that the world is going to be completely untethered from an experience point of view, we are going to be needing the best-in-class wireless connectivity. To keep that going in the right direction, along with the need for transacting tremendous amount of AI data across wireless networks, you need not only to be able to operate in congested environments, but also deliver a bandwidth capacity that otherwise would not be easily possible, right?
Wire-like wireless performance in congested wireless medium is what I envision we need to kind of continue to strive to deliver. For that, pushing the R&D going all the way into Wi-Fi 9 beyond 2030 is very important. That's where we are. That's how important wireless connectivity is for us.
So put that all together. You've got the MCU stretching all the way to the connected processors. What does that look like from a growth profile for your business? And then who are you increasingly running into from a competitive dynamic? You singled out a couple of things that make you very unique at the edge. But as you move into the connected IoT realm, there are some larger players there. Who are you running up against? And what do you think kind of the TAM or growth trajectory, however you kind of want to describe it?
Yeah, I think, look, if I look at the IoT market today, and I go back to one of the statements I made earlier, the IoT market today from a silicon processor point of view is about $15-$20 billion a year, right? And if nothing new happens other than the fact that that IoT market graduates itself to becoming agentic AI IoT market, that silicon market, $15-$20 billion remains, right? It's brand new market, and that's the opportunity for us. I believe that's not going to be the case. I believe there's more that's going to come into that marketplace. As more and more IoT systems become untethered, you get the wireless TAM coming in, right? And so you add to that. Some of these newer devices, IoT devices, are personal solutions like you referred to the glasses, basically.
And there are many others, right? There's going to be a pendant. There's going to be another form of wearables. There's going to be another form of earbud. There's going to be another form of human interface-assisted solutions like a humanoid or a robot, basically. And all of these are additive that I'm not adding to. So that is the opportunity for us, right? $15-$20 billion is the baseline, but there's a lot more that we haven't contemplated. And because the timelines continue to kind of change on when they will show up. Based on our looking glass, there is a tremendous amount of activity in this area, and that's what we are excited about.
And then, competition as you move into the higher?
Yeah, I think if I look at competition today, people mistake us for being competitive or seeing Qualcomm as our competitor because Qualcomm definitely has got really good assets at the edge. However, they are at the higher end of the processing dimension, and we are at the furthest end of the edge, basically. What Qualcomm would bring in is multiples of tens of TOPS. Ours is a lot more lighter architecture, a lot more suited for applications as we referred to. That's how we see. I'm sure some of our peers in the MCU world and microprocessor world will come in with AI. We haven't seen much from them just yet, and we'll wait. But I personally feel today we have a minimum of 18 months to 24 months advantage in terms of time.
Plus this partnership that we have with this hyperscaler excites us, basically, because it continues to kind of propel our roadmap.
Yeah, that's what I was going to ask next is in terms of timing, when do you see the elbow in the curve? You sound like you have a very good technology profile. But as that converts to dollars is when you get more interest from the market in general. Is it the idea that you can do a semi-custom chip? Is it the idea that Wi-Fi 7 is more broad-based and in certain consumer applications? When do you see the revenue trajectory change more dramatically?
Yeah, I think so we come from a very small base on IoT. When I say small, it's $400 million on an annual run rate. And in the last six quarters plus the one that we are in that we guided, we have grown and we are going to grow, including this quarter, on average, 50% year over year on a quarterly basis, right? And we've guided that we are very comfortable with 25%-30% growth rate in IoT over the next three to four years, basically. And so our growth currently is propelled by Wi-Fi, wireless connectivity, and our leadership with Wi-Fi 7. And we are in the early phase of our processors with edge AI seeing traction. Our growth in calendar 2027, by the way, we are very comfortable with our stated growth projections for 2026.
The growth in 2027 will be supported by growth of our processor business with edge AI, right? The three products that I just talked about, I'm not even talking about other things on the roadmap, but these three will hit prime time or production scale in different part times of 2027. The one that we are sampling right now, the one that is going to be a connected Astra with Wi-Fi 7 and BLE, and the third one, which is a semi-custom product. So I think that's how I would say. We're not going beyond 2027 right now, but comfortable with 25%-30% growth rates.
Let's go to the other side of the business in enterprise and auto. So enterprise has PC inside of it. You've talked about Windows 10 end of life, PC upgrade cycles. Something that we're hearing more, and we haven't got a good soundbite on it yet in this conference, but it's memory shortages. So generally, there's a concern that consumer applications will not receive priority in terms of memory, just given that everything is going towards the data center. Those guys tend to be able to pay a little bit more. Are you concerned around any kind of memory issue into 2026 from a PC perspective? Kind of what's your base case on that market there?
I think Ken can chime in because I'm hogging the microphone, right?
No, no, no.
I have another one.
Go ahead, go ahead.
So I think the word that you use is exactly where I want to go. Everything that is going to come as headwinds because of memory is going to be in the consumer space. Our play in PC is largely on the enterprise. And I think that provides a certain level of insulation, right? At some point, that market can also experience headwinds, but today we believe the ability to absorb the pricing jaws are much higher in enterprise than in consumer space. And so the majority of our share, in fact, all of our share rather, is in the enterprise space and the commercial space. And so we are not seeing as much of memory pricing related or supply related headwind just yet in the enterprise marketplace as others may be seeing in the consumer marketplace.
And the other good thing in that business is recently we gained market share. So a little bit of the ability to absorb the current pricing dynamic on the memory side, plus the benefit of market share gains, I think is playing out in our favor.
And then the other side of that business is auto. You guys have already pointed out sluggish auto trends. Any kind of sign of an inflection or an improvement there or still kind of early days?
Well, I think auto has been sluggish for us, and we don't play in China. And so the majority of the upbeat commentary is coming out of China. And so I think that's going to remain the way it is right now for us. Number one. Number two, I think more importantly, increasingly auto is getting smaller in terms of our total revenue contributor, right? And so today it's about 10%. And if our IoT business continues to grow, which we, like I stated earlier, and very comfortable, it will continue to grow 25%-30%. auto becomes even a lesser percentage, less than 10% share. And so at this point, I'm really encouraged by the 90% of the revenue that we have and the one that's heading in the right direction. And so auto is definitely going to be a smaller portion of our business on a going forward basis.
Last one to check the box on the business that's not the 90%. Mobile, does foldable change the conversation for you guys in any way? Just to check up on the Android ecosystem, I get that question a lot. Does that change the dynamic for consumers?
Oh, absolutely. That's an excellent question. Again, really proud of what Synaptics is all about. 40-year history. The majority of it is on the backs of Touch, Mobile Touch. And the reason why I kind of bring that up is the company has the reason why I'm a big fan of the capability in the company is it has reinvented itself on multiple turns to take and maintain leadership position in this area, right? And so foldables is another category where Synaptics continues to take leadership position in the marketplace, will gain market share as a result of foldables getting market share within the smartphone market, basically. And every foldable that ships effectively has an opportunity for two instances of Touch and greater than 2x of revenue that you would otherwise have from a phone, largely because the larger screen dictates a capability that is premium quality and premium value.
And so as the foldable markets increase, which has shown growth in the last year, and we anticipate it continuing to grow, us gaining market share in the foldables in the 2026 cycle that kicks off in the fall, not only in Korea but also in China, in the Android ecosystem is very encouraging for us on that front. Ken, if you'd like to add anything.
No, I think that's right. We talked about a win there in our last earnings call with a Korean customer, and we have great traction, all on the high-end Android market, and so very excited about that as we head into the second half of calendar year 2026.
Let's switch it to capital allocation. So it sounds like over the next couple of years, you've got a pretty good line of sight to not only the piece of the business that you are most excited about growing, but also around profitability improving as well. Maybe talk about how the operating model is going to change and potentially where your priorities are in terms of capital returns.
Yeah. So if you look at the capital allocation, I would say first and foremost, continue to invest in the business. And even there, one of the things that we've been doing behind the scenes is reallocating capital to these higher growth segments over the last 18 months or so. And so investing in terms of AI at the edge and core IoT, we'll continue to do that and make adjustments within the spending envelope that we have today. Number two is if we look on the M&A side, we've done a few of these tuck-in acquisitions, all on the Broadcom purchase in terms of the IP and some headcount there, as well as a small BLE type of acquisition that we did last year as well.
So we'll look at these ways to enhance the existing portfolio, but primarily focused on these smaller tuck-in opportunities that can enhance the scale. And for the most part, focused on that core IoT segment and AI at the edge. And then what you've seen over the last really year and a half or so is returning capital back to shareholders. So if you look at fiscal 2025, we returned about $128 million vis-à-vis share repurchases. In July, we announced another $150 million authorization. We did some of that in Q1, and we talked about doing more purchases here in Q2. So that's another avenue in terms of returning capital back to our shareholder base.
Well, we're out of time. I very much appreciate you both being here. Things sound very exciting over the next couple of years. So good luck as it goes along.
Appreciate it.
Thank you all.
Thank you.
Thanks, everybody.
Thank you, everybody.