CES 2025 Day 1: Video Interview with Synaptics' Satish Ganesan

Hi Nathan, I'm good to be here. Good to be with you.

There's a lot of exciting things going on here at CES for us. We are, as usual, showcasing our sensing, processing, and connectivity technologies—those are the three pillars of Synaptics. Our focus has been on figuring out how to push AI to the edge of the sensing technologies. That's the focus here, other than showing our AI-native solutions and connectivity solutions. There's a lot going on. If you look at PCs in general and the rage about AI PCs, most of it is about introducing a Copilot PC. Microsoft is driving that ecosystem, and the whole piece is how many TOPS can I drive in my main processor and how do I run Copilot seamlessly integrated with all the applications? That's what the evolution of AI PCs is all about. You have Arm and everybody else participating in this ecosystem, and we are expecting MediaTek, more processors—Nvidia, Jensen was talking yesterday—so all of them participate. Nvidia and MediaTek have done big announcements here as well. What we are trying to focus on is not necessarily the main processor—that's already been tackled—but the main thing is we do a lot of solutions: the touchpad solutions, the fingerprint solutions. That's our history. Our history of fingerprint and biometrics starts from an AI neural network—I said neural network at that point, not AI—so it started from there. But the question is, can we push AI technology, machine learning technology, to the edge with very little embedded processors, very little memory, and very little processing power? We're not talking about teraops, but gigaops or just ops. How much can we do to enhance the user experience? That's what we're trying to talk about: distributed AI within a PC, within a mobile phone, within automotive, and so on. That's our focus for our sensing solutions—push AI on those particular edges, so more distributed computing.

That's right. It's very interesting, and we are showing some demos here as well. If you think about what we do in automotive, we are traditionally known for displays in automotive. We are introducing more new projects like the smart bridge and things like that. But in terms of the display, we can now do interesting things like driver detection. If a driver presses the display, it won't activate the display, but if a passenger presses it, it will activate, so you get different views. That's adding intelligence to our touch sensor capability using small machine learning algorithms to ensure safety and security. You can also do things like glove detection and moisture detection. Simple algorithms that are typically procedural but prone to a lot of errors—with machine learning and AI, those errors are eliminated, so now you have very high-accuracy solutions for simple use cases that enhance the consumer experience tremendously.

Exactly. What we're talking about is low memory and low compute, so how can you run algorithms in that manner? We're doing a lot of innovations, picking up from a lot of this AI-native stuff we're doing, but trying to run it on a more constrained ecosystem. For example, in the touchpad and things like that, we're trying to introduce solutions where you can detect your palm versus the finger. Typically, you'll find your cursor moving when you don't want it to—a simple issue—but you can resolve those with AI algorithms. The experience for the user is like, 'Oh, it's a better PC,' but the AI algorithms are running in the background to enhance it.

Right. I think our consumers—when you talk about PCs and everybody else—our customers are asking us, 'Hey, everybody is going to get the processors from Qualcomm, MediaTek, Nvidia, Intel, or AMD. What can we do to differentiate ourselves from our competitors?' So we have to give them new technologies in the AI PC, automotive, and so on, where they can do this differentiation. That's very similar to what Arm's SLMs are going towards, but we are incorporating it into existing products. Things that are usually boring are becoming more interesting now.

That's right. That's how we see AI as being more useful to the end consumer, and we have to enable our customers to bring those features by putting these algorithms in.

Oh, that's right. We have that here as well. It is actually in some of the laptops already today, shipping in PCs, in notebooks. That involves putting a small processor next to the camera sensor and detecting a user. You can use it as a secondary format detection—when you gaze away, you can turn off the monitor, and that'll enhance battery life. We're looking at those kinds of applications, again more centered on improving the customer use case, improving the customer scenario.

You could upgrade. We can set you up with something.

Good question. To zoom out a little, one of the things we are going to focus on—as you can see everywhere, it's a buzzword, but we're trying to incorporate it in the way we operate as well—you're going to find AI infused in all of the ecosystems. You'll find it in our sensing solutions to push AI at the edge. Then, of course, our processing solutions—you saw Astra and the AI-native solutions we have today with the MCU—that's going to be a focus where we become the primary processor. And then our connectivity solutions—the goal is to improve the user experience for connectivity as well. Providing AI across all of these, infusing AI into everything we do, is something we're going to focus on. That's what Synaptics is showcasing here, and you're going to see that as a continuous trend. The good part is the market trends are in the same direction, so all our customers—whether in the PC world, automotive world, processor set-top box world, operator world, even basic—everybody we're doing traditionally wants all of these features. So I think you'll see Synaptics move along with everybody else in that direction, infusing it into almost everything we do.

All right, thanks a lot, Nathan. Great talking to you. Thank you.