Synaptics isn’t a household name, but it makes a lot of the underlying technology behind things like smartphone touchscreens, automotive displays, voice-enabled smart speakers, and virtual reality headsets.
As we make our internet of things (IoT) devices smarter and smarter so that we can get dumber and dumber, or at least more relaxed, we’re creating a tsunami of data. Rather than sending all that data over the internet to a datacenter, we can process, analyze, and act upon much of it right where it’s collected.
The result is that devices at the edge of the network — things like tech-enabled cars, Alexa smart speakers, and security cameras — are getting smarter. And Synaptics has to adapt to this changing world by using technologies like artificial intelligence to improve the chips and other technologies that do the sensing and processing in these devices.
I talked with Rick Bergman, CEO of Synaptics, which makes touch display controllers and a variety of other smart underlying technologies, about what’s coming down the road.
Here’s an edited transcript of our interview.
VentureBeat: What do you see coming into the market soon, this year?
Rick Bergman: You had an article on our new SOC solutions, with neural network and voice enablement. We just began sampling at the tail end of last year, beginning of this year. We’ll see that mass production within our fiscal year, by the end of June. You’ll see some interesting new solutions come to market with that capability.
VentureBeat: Is that a lot more voice-based, smart hubs and things like that?
Bergman: Voice-enabled anything, pretty much. The initial ones, as you can imagine — the smart speaker category is the fastest-moving. But it applies to mesh routers, mirrors, toilets, TVs, set-top boxes, all utilizing the far field voice capabilities that we offer to the market. More specifically now, with the new ones we announced, the neural network capability can improve the voice capability.
VentureBeat: We’re used to having that pause when we talk to Alexa, or having to say things twice. Do you think those days are going away? Or might we still have that kind of problem?
Bergman: It’s getting better and better. I saw a presentation a couple of days ago where the smart speakers may be better than humans at interpreting speech now. It’s pretty close as is. The latency certainly is potentially moving the recognition, or at least the limited vocabulary, to the edge. You could see a bit of improvement in the latency.
VentureBeat: Is that a trend you also see, that a lot of the intelligence is moving to the edge?
Bergman: Yes. We’re just sampling these devices, so there’s going to be some lag time or latency from when we get the hardware capabilities to when the software fully takes advantage of it. But in the second half of this year, when you see some of these systems announced, that intelligence is clearly moving to the edge. The chip we announced has a billion operations per second.
I know you’ve tracked gaming for a long time. I remember distinctly, in 2008, shortly after AMD acquired ATi, they introduced the first billion operations per second GPU. That was a $499 solution. The chips we announced will be $2 or $3, with that same level of capability, at least measured by the number of operations. It’s the same idea. It’s matrix math. One is intended for graphics and the other is intended for AI.
VentureBeat: When you put some of these things together, like 5G and this intelligence at the edge and new interfaces, what do you think comes out? What are some things you’re looking forward to?
Bergman: To break it down a little bit, on smartphones certainly 5G is going to drive gaming — as we move to foldable and bendable and that sort of thing, as well. We’re being encouraged by our customers to increase display refresh rate because of gaming. For a long time it’s been 60Hz on a phone, but now we’re getting requests for 90Hz or 120Hz. We’re moving in that direction. Some of it — not all, of course, because 5G exists today — but some of it is being delivered by 5G enabling higher performance gaming. People want that response rate associated with it.
VentureBeat: I saw Razer was really touting the 120Hz on its phones. Apple had still not quite gotten there. They had 120Hz touch, but people want 120Hz refresh on the screen.
Bergman: It also helps with videos, for the same reason TVs have a higher refresh rate. Some people in the industry are still a little bit dubious. It’s one thing to see it on a 70-inch TV. On a 7-inch smartphone, can you actually see a football move in a way that makes that meaningful? I’m not sure. But in any case, people are very discerning. That’s what they want. We certainly will offer these capabilities. Asus is another one that also has a high-refresh phone.
VentureBeat: On the VR side, do you guys have things that are seeing use there?
Bergman: We have display drivers that are actually dedicated just for VR, to drive 2K x 2K screens. Of course, that doesn’t sound like a lot, but the screens are one inch on a side, so it’s a very high DPI. You don’t see the screen door effect with the glasses an inch from your eyeballs. You’ll see that again, actually, with those coming out in systems in the second half of the year, as well. Very cool stuff.
VentureBeat: I started seeing some new models at CES that were talking about 4K per eye. It sounds like VR systems are moving ahead of where this first generation was, with the Oculus and Vive, the kind of visual quality you could get with them. For those, are you confident that we’ll get a whole new generation of these pretty soon?
Bergman: Oh, yes. I think everybody’s waiting. There are many opportunities to improve the experience with VR and the visual quality is one of them. Obviously, the size of the headset is another one, and tethered versus non-tethered, and being able to see the environment around you so you’re not stumbling into the coffee table. They’re all fundamental challenges, along with the long-term issue of a small percentage of the population just having a fundamental issue with latency and so forth. We’re working with our customers on a few of those challenges, as well. Foveated rendering is another one where we can help out.
VentureBeat: With 5G in particular, do you also have some solutions there? Or is that going to drive a different kind of use of what you already have there?
Bergman: It’s more the latter. We don’t do any 5G components. We don’t have any connectivity capability at all. It’s more about how 5G impacts our devices. I already mentioned the refresh rates for displays and touch controllers. Also, because it ends up requiring a little more RF and a little more power, we’re getting pushed hard on our display drivers to reduce power and help create thinner screens, so more battery can get squeezed in there. Ultimately, of course, we hope it causes a new wave of smartphone replacement.
VentureBeat: What do you see taking hold in cars?
Bergman: From a Synaptics perspective, we’re primarily in the display area. In there, the trend is clear. I don’t need to tell you that you’re going to see much larger displays, and curved displays — some of those with more of a minimal curve still being LCD. It’ll take a while before you see OLED displays in vehicles. We’re seeing that type of approach. But you’d be amazed by how big of displays people are contemplating putting in their vehicles in the coming years. They’re almost like TVs.
VentureBeat: I saw the Byton display. That’s a pretty big one, from one edge of your windshield to the other. The demo that Intel and Warner showed pretty much meant that the entire side windows and back seat side of your car would be one gigantic display for the passengers. That was pretty amazing.
Bergman: You get to see more demos than I do, but I do visit customers. I’m talking about mainstream stuff you’ll be seeing in two to three years, where you’ll have a display across the combined instrument panel and infotainment area that’s two or three feet across, with one continuous piece of glass, all touch-enabled. A Tesla on steroids type of thing.
VentureBeat: What do you find a lot of your competition coming from these days?
Bergman: It depends. Our classic stuff we’ve been talking about, the display and touch, we just don’t have U.S. competitors anymore. It’s typically Taiwanese competitors. Sometimes vertical competitors, like a Samsung. That’s one group of folks. There are probably names even you don’t know, like Novatek, EyeMax, those type of guys. When we talk about the voice, now we’re starting to compete with the big boys in the industry, like Qualcomm, Broadcom, MediaTek. We’re offering full SOC solutions, dual quad-core ARM type of stuff.
VentureBeat: In the foldable phones, are you guys in there as well?
Bergman: We certainly are working with many of the OEMs. I think at this point — I’m not sure there’s an officially launched phone.
VentureBeat: Is there enough activity there that you believe foldable phones are going to be plentiful?
Bergman: Not in the next two to three years. The whole idea of foldable is — at least today, in effect it’s two displays. They may connect, of course, but you’re doubling the cost of the display at a minimum. At least to me, in 2019 and 2020, any foldable phone will be above a $1,000 price point. I’m not sure — for someone like Apple that means very low volumes.
VentureBeat: For all these windows into what’s coming, I guess we still don’t know exactly what’s going to take off among all this stuff.
Bergman: We know it’s going to take off. It’s just a question of what year. I have zero doubt foldables are going to take off. I don’t know about you, but I still carry three devices. At least in the next couple of years, I want to eliminate at least one. They’re each very suited to their particular task. I think foldable will allow me to not carry a tablet anymore, to watch movies and so on.
I’ve been a huge fan of VR. I’m frustrated that it hasn’t moved faster. Maybe for someone like you that’s been in gaming forever it’s different, but looking at something — even a large display is artificially constraining. I hope we solve some of the bigger issues in VR pretty quickly, because I’d be really excited about what VR means, and not only for gaming. You could do incredible movies with a VR headset. That I’m starting to see. I don’t understand why there isn’t more focus there. But you have to start off with it in mind that this is going to be a VR movie. Eventually, filmmakers will catch on and it will fundamentally change the cinema experience.
5G, of course, that’s going to happen. That’s already happening. You don’t need a crystal ball for that one. Or maybe you do, because that one involves government officials. [laughs] That’s the only thing that will slow the rollout. There are a lot of exciting things happening in the consumer space. We see all the home stuff, too. The home is a fragmented market, so it tends to march a bit slower than some of the other consumer devices, like smartphones. But again, I think everyone’s moving to a highly connected home, with all these different devices.
VentureBeat: As far as the consolidation in the semiconductor industry, what do you think about that? Is it finally slowing down a bit?
Bergman: One of the banks came in and showed us the consolidation trends. Their line didn’t show it slowing down. In fact, they mapped it against other industries. I think one was pharmaceuticals. I can’t remember the other one. Anyway, it was interesting, because it was different decades, but the actual trend line was very similar. I don’t see it slowing down, at least over the next few years. Scale certainly helps in the semiconductor industry.
Now, I can’t recall an announcement over the last three or four months. I’m sure there have been some. Some of it again goes to government regulation. When you’re either acquiring or being acquired, deal certainty is a key factor. If you have to get the Chinese or U.S. government to approve, the risk on this is going up 10 times. The truly international type of thing may end up getting hung up for a while. That’s slowed things. But I think the broader trend line, once we get through this rough road, will get back to consolidation.
VentureBeat: How much do you see Synaptics needing to dive right into all of the AI technology? Is it going to become a fundamental part of all of the components that you’re making? Do you become similar to a lot of the other companies that are chasing AI right now?
Bergman: Interestingly, AI is still fashionable. [laughs] We’ve been using AI in the last few years on our fingerprint recognition. We would train our devices to better understand your fingerprint over time, as an example. On our touch controllers, we also used it to understand an individual’s palm versus finger. We do offline training on that. On the voice-enabled stuff, as I mentioned, we’re putting in the hardware capabilities.
We clearly are believers in moving that intelligence to the edge. From a company level, we have zero plans to jump into the training side of things. Plenty of guys going after that are much bigger and have a huge lead over us. We’ll stay on the edge side. Every product we do now, we look at it and — you can add a neural network capability in a few square millimeters of silicon, so it’s not a huge cost. There’s emerging standards there, so you don’t have to do an onerous software support role either.
VentureBeat: Do you find you have to chase after some of the AI experts, the engineers?
Bergman: We’re fortunate. We’ve worked on some applications, so we have our AI team. Because we started early, three or four years ago, developing neural networks for fingerprint, we’ve been able to retarget that team to focus on IOT now. It was a good young team previously, and now it’s a very good, more experienced team.
VentureBeat: What’s your own perspective on how we’re doing with Moore’s Law?
Bergman: It’s definitely slowed down. Intel is a great example there, where they’re somewhat stuck on 14nm, and of course TSMC has split ahead of them. The cost return has slowed down on a global level, which has slowed the investment intensity to get to the smaller geometries. We don’t play in that space like AMD or Intel does, down in five or seven. But our 22nm chips are finishing. Now we’re working on 12nm. That’s a small minority of the products we do.
VentureBeat: You just don’t need it, generally?
Bergman: No. If you have a high analog content, it’s way too expensive to even contemplate doing it. Our display driver devices tend to use older geometries because they’re basically large analog chips with a little bit of digital in there. Not a lot of benefit.
VentureBeat: That sounds like a good place to be right now.
Bergman: Avoiding the $10 million mass charges is nice. You can’t place too many bets, so you’d better be right. It’s hard in this dynamic market.
VentureBeat: How many people does Synaptics have now?
Bergman: We’re about right at 2,000. We did two acquisitions. We bought the old Conexant, and then we bought a division of Marvell about a year and a half ago. That popped us up from the mid-1,000s to 2,000. That’s the ballpark we’ll stay in for a while.
VentureBeat: Any other big picture things you want to share?
Bergman: We’re obviously shifting to be more of an IoT type of company, from our traditional notebook and smartphone products. Automobiles are a big growth area. We’re starting from a small base, so in the next few years our revenues will accelerate quite a bit as you see all these touch screens popping up in automobiles.
VentureBeat: I’m starting to review some of these new cars. I’ve got really old cars in my family, so I’m just not used to driving new ones. It’s fun to see all the new gadgetry, like heads-up displays.
Bergman: They’re trying to close the gap with consumer technology. I have a Tesla, which was certainly an eye-opener in comparison to other vehicles. I had a 12-year-old Acura before that, so I wasn’t exactly pushing the state of the art.