Okay, we'll go ahead and get started here. Thanks everyone for joining us. I'm Luke Junk. I cover vehicle technology and mobility for Baird. We're very pleased to have Mobileye with us here, kicking off the afternoon. Mobileye is the ADAS market leader, providing an excellent base for both near autonomous and consumer AV offerings, both of which are currently being booked. Joining us today, we have Dan Galves, Chief Communications Officer. The email address for this session is sessiontwo@rwbaird.com. I'll try to work in any questions that, that you've got. And with that, we'll start with some quick introductory remarks. Dan?
Thanks a lot, Luke. Yeah, I'm Dan Galves, Chief Communications Officer for Mobileye. Mobileye's core business, about 90% of our revenue is essentially supporting driving assist systems, active safety systems. So we support a single front-facing camera system. We, we sell a chip that has software on it that interprets all the data that's coming in from that camera. It's really a great business. You know, if you go back to the time we were acquired by Intel, about 15% of all cars produced globally had some sort of driving assist system. We were, o ur technology was on about half of those cars. We generate. W e call that product Chauffeur. And it's been a really positive year for our business development efforts around these technologies.
As we said on our earnings call, in what month are we in? October. We have high probability advanced discussions with 10 automakers that represent about a third of the industry, and there's a next wave of customers that are earlier in the discussions that represent another 15% or so of the industry. If you go back to the beginning of the year, those numbers would be three OEMs, representing about 9% of the industry. I think that what has happened over the course of this year is that we're seeing kind of a flywheel effect that is moving customers towards outside solutions like Mobileye, as opposed to, you know, the desire to insource this technology, create your own software as an automaker, and deploy it on more of a general purpose chip.
You know, this flywheel, there's several components to it. I think the starting point is really China. If you are not following kind of what's happening in China in this hands-free market, you have no idea the pace of innovation that's happening there. The consumer demand for this type of product, the traffic congestion in China is just crazy, and people want cars that can fight the traffic for them instead of them fighting the traffic. There are several systems on the road. We launched a system with our customer called ZEEKR, you know, back in early 2022. But there's other systems on the road that are pretty good as well, you know, mostly by the EV startups in China, like XPeng and Li Auto and NIO.
So kind of the pace of innovation in China and the consumer demand is putting pressure on global OEMs to move faster to get these types of systems on the road. Clearly, there's some level of, you know, being behind on EV. You know, they don't wanna get behind on AV as well. So that puts more focus on pragmatic concerns. So rather than, you know, if, "If I can develop this software and own the IP, you know, I can create a better system than everywhere else," you focus on things like time to market, cost, and performance, and these are areas where Mobileye has clear advantages, particularly in the context of in-house solutions.
And then I think the final piece of this flywheel effect was really, you know, we had had hardware in these ZEEKR vehicles, but the software to actually enable what they call Navigation ZEEKR Pilot, you know, a system that can take you from point A to point B, you know, without any input from the driver, that rolled out in June to 1,200 people, and then broadly to more than 100,000 cars in late August. And the feedback was great in terms of media, consumer feedback, and so this was kind of the final piece to make it real, that, you know, yeah, Mobileye, I've been in their demos before, I've been in their prototypes, but, you know, is the software actually gonna roll out?
It did in August, and that's created, you know, kind of a really groundswell of converging towards these agreements. So I think that's really my intro. I'll say a couple of words about Israel. Obviously, we're an Israeli company, and it's been a tough, you know, last month. I think that, you know, things in Israel are getting back to normal, so the population centers, the stores are open, schools are open, you know, people are on the street, people are on the beaches. So kind of the normality has come back. We do have, you know, quite a few people that have been called up to the military.
But part of kind of the ethos of the company is kind of, you know, number one, always having this kind of, you know, conflict in the background, but also, you know, the ability to get your work done, to not fall behind on any priorities, is really kind of the effort towards the war that everybody at Mobileye can make. And so we feel like operations are normal. Business development can fly, you know, in and out, do customer meetings. We won tons of new business during COVID, so I don't think that there's any issue there. So, you know, just, you know, feeling fairly normal at this point.
So, yeah, thanks for that, Dan, and of course, our thoughts with the entire Mobileye team in Israel in this current situation. Wanna start the conversation with some foundational elements of the business. You know, it's now been about a year since the company has come back into the public markets, and I'd just be curious how being public again has changed, maybe not the focus areas, but maybe benefited the company culturally, having spent about five years behind the scenes, really [crosstalk] at Intel.
Yeah, I think, I think most people's life hasn't changed very much. You know, we operated as an independent company the whole five years we were inside Intel. I think that that was really Intel's desire, was to kind of let us, you know, flourish on our own. So kinda, I think day-to-day life hasn't changed much. I think, you know, having your variable comp in Mobileye stock instead of Intel stock, you know, not comparing the two, but you just have more of an impact on Mobileye stock if you're a part of the company. So I think that's been a huge benefit. The other thing that's helped is we've really generated a kind of higher profile in the media and among, like, industry analysts and, you know, sell-side research talking about us.
So I think now when you see articles about AV or ADAS, you know, we're mentioned as one of the leaders. You know, we're called a technology company. So I think that that's actually helped with the business development because, you know, not everyone within an OEM, you know, knows so much about Mobileye, so you want your name to kinda be a centerpiece, when, you know, in terms of, kinda, the chatter about the sector. So I think it's helped quite a bit.
We're gonna spend a lot of time talking about SuperVision and Chauffeur, so we'll get to that. But I, I do wanna take a moment to just talk about the 90%. So you've still got a business that's growing nicely in terms of EyeQ chips and ADAS systems, plus some content drivers. We've got cloud-enhanced ADAS, and you just had your first booking for SuperVision Lite in the Q3 as well. Just at a high level, how should investors see that underlying business trending relative to production if we look out over the next couple of years in terms of key drivers?
Yeah. So we still think that there's, you know, if maybe 65% of all cars produced globally have ADAS systems this year, you know, we think that that number goes to 85%-90% over the next five or six years. So you should get kind of a steady five or six points of growth just from adoption rate. Obviously, it's your guess as good as mine what kind of overall industry production does, but, you know, that's an impact to us as well. I think we definitely have some, you know, positive average selling price drivers happening in that single chip business.
One is cloud-enhanced ADAS, so adding this crowdsourced REM map that we have to the single front-facing camera system that creates additional safety benefits, additional convenience benefits, and it starts us on the path to recurring revenue because part of the agreements on this is the map needs to be refreshed, right? And on a consistent basis. So we get paid, you know, a small, maybe $15-$20 per car per year to keep that map fresh. We see that as potentially adding $1 to ASP, overall ASP next year, which is, you know, two points of growth. So not huge, but then, you know, it's, it's, it's kind of a compounding effect.
Those cars that are in market next year that are paying us get added to the cars that are produced next year and create a bigger base for the following years. So yeah, we see a lot of potential. I mean, in general, you know, last year, we booked new design wins last year, representing $6.7 billion of revenue in a year that we did, you know, maybe $1.7 billion on our own. Again, this is 2022. The average selling price was $105. That included some SuperVision but really the bulk of it was ADAS. So I think that's still a very powerful business that's gonna drive growth for the long term.
So now let's talk about leveraging that business as we move to both the eyes-on, hands-off, and then ultimately, Chauffeur. And I guess where I wanna start here is, given your base position in ADAS, in terms of customer relationships, data coming off the cars that are on the road today, you know, you mentioned mapping in terms of REM and all that it entails. What would you say are the company's most sort of unreplicable assets leveraging, you know, coming from that ADAS background?
Yeah.
versus, say, a pure L4 player?
Yeah. So I think, you know, the ability to interpret camera data in an efficient and accurate way. I mean, this is the reason why we have such high market share in the front-facing camera ADAS business, right? It's not like other players wouldn't like to get into that. So I think the computer vision, the 20 years of development there, the ability to, you know, detect vehicles, detect pedestrians, lane markings, traffic lights, traffic signs, like, what you need to do with an ADAS system today is much more sophisticated than what you had to do five or six years ago, and that's kinda helped us stay ahead of the game.
I mean, even some of the systems that are trying to kind of develop in-house, you know, by certain automakers that are on the road today, they almost all of them still use our front-facing camera system. So it, it's just really, I think, a big advantage. The integration of the hardware and the software, right? So if you have a chip design team that understands exactly what software workloads need to be run on that processor, they're in a much better position to create specific accelerators, you know, to run that software most efficiently.
They know exactly how much processing power the systems are gonna need. If you're, you know, in a kind of just a chip company, you're trying to guess at what companies are gonna need in terms of processing power, obviously, you're gonna over-engineer that chip, which means more cost and more power consumption. So I think that the tight integration of the hardware and the software is a huge advantage. And then finally, data, right? So we have, you know, we have full video that we get from pre-production OEM testing. We're not pulling full, you know, video data off normal cars running around the streets, but we've been doing this for 20 years, so we have, you know, I think it's, like, 250 PB of data.
Tesla typically talks about a huge data advantage they have. You know, six months ago, they said they had 30 PB. So the idea that we're at a huge, that you know, others have a big advantage in terms of data, for us, is not true. And then the mapping data, this is proprietary. We're pulling snippets of data out of normal production vehicles in order to create this map. That's another thing that we don't see anything else on the market that has any similar properties.
Okay. With that, let's talk about SuperVision. I'm sure on everyone's minds in this room. You know, this is a business that, as we look at it, seems like it's on the precipice of becoming much more diverse in terms of customers and launch activity. And as we look at incremental launches layering in the Q4 into next year, can you just talk about what's layering in in the next 12-18 months, moving beyond Geely to, I guess, your recent awards, especially with Porsche and FAW?
Yeah. Sure. So, you know, our kind of first agreement was with Geely Automotive Group, and it started with one car, the ZEEKR 001. ZEEKR is a start-up EV company that's owned by Geely Automotive, which also owns Volvo and a bunch of other brands. You know, once we got the system, you know, into the cars and things. You know, the hardware was working great, they expanded that to four more vehicles. This was happening in, like, mid-2022. So one of those vehicles launched in March of this year, the ZEEKR 009. We have a car by smart, smart #1, that launched, you know, maybe a month ago. The Polestar 4, which is a global vehicle, is launching in China right now.
It's gonna launch in Europe in Q1, in the U.S. in the summer. So we'll have kind of our first global vehicle with SuperVision. And then, there's a fifth vehicle. It's a Volvo vehicle only for China that's launching in Q1 of next year. So, and then, you know, our kind of most recent, you know, announced agreement was with FAW, a large automaker in China. The first product launch is in Q4 of next year. So yeah, we expect significant growth in 2024 over 2023, mostly just by adding these different models. You know, we're not assuming, you know, significant volume growth of individual models but just kind of layering on more and more.
What about if we start to peak out to 2025, 2026? I guess one surprise for us with the FAW award is that that was a, if I remember, September announcement.
Yeah.
As you mentioned, you're targeting start of production by the end of 2024. I mean, that's really quick based on our historical experience around any sort of auto technology really going onto a platform. I mean, to me, that would suggest that there's a lot of engineering work that's already being done at customers. Is that right? And what does that say in terms of how quickly you can launch business for, you know, future bookings that might be coming?
Yeah. So, really, the, the FAW, you know, there were several Chinese OEMs that, you know, we said at the beginning of the year, we're sort of watching, right? And seeing kind of once this software hit the road with ZEEKR, the expectation was, like, you know, these guys are using 50 TOPS of processing power. You know, XPeng is using 1,000 TOPS of processing power. They don't have any LiDAR. Everybody else has LiDAR. How can they compete? And so I think the Chinese OEMs were kind of waiting to see what the results would be on the road, and the results were really positive. And so we went from, you know, talks and meetings with FAW to an agreement in a, in a matter of a few weeks.
I think the great thing about working with Chinese OEMs is they do move very fast to launch product. So you know, it's really not just FAW. You know, we have kind of discussions with a couple more big Chinese OEMs that, you know, if things come together in the next few months, you know, wouldn't be a late 2024 launch, but maybe an early 2025 launch. So there's still, you know, time to layer in a lot of volume in 2025. You know, we think it's, you know, fairly significant volume with FAW in 2025. And I think that it's a platform, right? So I think if you're willing to do more of a plug-and-play and don't wanna customize too much, it can, it can hit the roads pretty fast.
What about in terms of how you're, you know, how you're launching this business? I think there's some questions out there. You know, you've been acting as a Tier 1 in some cases. Does that impact margins? Or maybe we just talk more generally how that's developing versus the historical Tier 2 model in the partners that you've had.
Sure. So in the core ADAS business, we're a Tier 2, so we sell the chip to an auto supplier, like an Aptiv or a ZF, Magna. There's several in Asia as well, and they're in charge of buying the camera and putting the chip on a circuit board, camera housing, deliver it to the customer. That really helps us to scale, because clearly, you know, there's a lot of work that has to go on, you know, with actually validating the technology before it goes into production with each car calibrating. You know, we don't really wanna do that. With SuperVision, we're agnostic, really. It seems like a lot of OEMs wanna have a more direct relationship with us on these programs. They see them as more strategic.
They, there's many more kinda aspects to these systems that the OEM needs to have a role in. Like, you know, what does the driving experience feel like? Is it aggressive? Is it more cautious? Is it, you know, at what point do I start to brake before a stop sign, so it doesn't feel too harsh? You know, how do I visualize, you know, for the driver, what's happening around the car, the driver monitoring, the parking software? There's a lot of things that the OEM has to be involved with, and you know, and I think it probably speeds time to market as well. So I would say that, you know, all of our kind of announced customers were the Tier 1.
They're, o ut of those 10 OEMs, there's two or three that we would probably be the Tier 2. The difference is essentially, you know, Tier 1, we're not just delivering the two chips, we're delivering the chips on a domain controller, so there's a lot of kind of third-party peripheral components that, you know, we can't charge a big margin on. So, you know, the SuperVision, the target gross margin is 50%, but I would say it's, you know, kinda half the price is, you know, the chips for 75% gross margin, and half the price is, you know, everything else for 25%. If we're Tier 2, then we're just delivering the chips, so the revenue per unit would be lower, but the gross margin would be a lot higher.
We also save money on OpEx because we wouldn't have to have, you know, kind of as big of on-the-ground teams with the OEM.
So let's bring Chauffeur into the discussion as well. So you've announced some Chauffeur awards in the last few months, notable in and of itself, but what was interesting is that with the Polestar 4, you announced SuperVision, and then Chauffeur came in pretty quick succession. We had a joint announcement with FAW. Just how is that dynamic between these two products being related and, you know, essentially modular in nature? How is that evolving? Could we see more joint awards in that group of 10 OEMs going forward?
Yeah. No, I think you could. I think the OEMs, the automakers are looking at this as an evolution, and I think the one consistent thing across the automakers is they all believe that if you can give a consumer a car where, you know, on the highway or maybe off the highway as well, they can do other things while they're on their commute or on a long road trip, that that will have significant value. But obviously, you're gonna have to kinda convince them that they'll be safer if they activate the system rather than not, right? 'Cause there's a trust component here as well. So it's down the road, but this is where the OEMs see the biggest value.
If you could end up with, you know, two offerings, one, which is, you know, eyes off and is more expensive, you know so maybe it's, you know, maybe it's $5,000-$6,000 cost to the automaker, and they need to mark that up to the consumer in some way. And then SuperVision is maybe $2,000 all in to the automaker, and they've got to mark that up, you know. And so you kinda create these offerings. I think that's where we're going. Chauffeur, the difference between SuperVision and Chauffeur is SuperVision is camera only. So there's the only perception system is 11 cameras around the vehicle, giving you a 360-degree view by cameras, and those cameras are capable of kinda supporting driving in all conditions, but the failure rate is probably not better than a human failure rate. So you don't want to allow eyes off.
The only change to Chauffeur is you add a secondary perception system made up of radars and LiDARs that also is capable of driving the car on its own. We have vehicles like this with just radar and LiDAR that are operating in, you know, on public roads in Israel, and other places. If you put those two perception systems on the same car, the chances of both of them failing at the same time, you know, go down exponentially. And so, and, and kind of the benefit to the automaker is by developing SuperVision, you're developing, like, 80% of Chauffeur. So you get to leverage that instead of, like, trying to kinda stuff every sensor you can think of to get to an eyes-on system, and then, you know, once you get there, you're like: "Okay, now, now, what do I do?
I think that this kind of evolutionary path is very positive for us. So yeah, most of the agreements we're working on, you know, include both SuperVision and Chauffeur.
Based on just what you've already booked and the timelines that you've outlined, can we start to pencil in some Chauffeur revenues coming into the P&L late 2025, likely? Just materiality there, I guess if I look back to the beginning of SuperVision in late 2021, even a few units start to move the needle.
Yeah
Meaningfully.
I think that's right. I mean, the target launch date for the Polestar Chauffeur is late 2025. For FAW, it's late 2025. Like, I would give a healthy kind of view that that could slip into 2026. Like, we don't think it will, but, you know, I think that this, it's complex. But yeah, I mean, it moves the needle a lot because you're talking about, you know, maybe $3,000, you know, without thinking about the actual physical sensors. And, you know, we have an imaging radar product that we're very excited about that will probably generate some volume on those types of programs as well, and that could, you know, increase that $3,000.
Yeah. What about the regulatory side for Chauffeur in terms of what conditions you need in your major geographies to actually get a Chauffeur, Chauffeur car out on the road? I know you're actively engaged with regulators and that, RSS, the driving policy, is really structured to be understandable to regulators. Maybe just a quick update on where you stand on that front.
Yeah, so the move from eyes on to eyes off is a big move, right? Because eyes on, it's you know, the driver is still expected to be monitoring the system. You know, you can have sophisticated driver monitoring to make sure that that's the case, and so the liability really never shifts, you know, from the driver to the car.
And we don't see a lot of regulatory kind of interest in this as long as you have that kind of robust driver monitoring system. But eyes off is a completely different story, so the liability would shift from the driver to the car, you know, which creates risk for the automaker, creates risk for us. And then we definitely see, you know, we believe there would be regulatory oversight of this. So we have a plan. The first step of the plan is really to validate kind of the mean time between failure to better than human levels through offline, you know, using data and kinda measuring the accuracy of the sensor systems through simulation, where we can use our REM maps to sort of, they call it hallucinating, kind of a real world.
You know, this is, we're having success in convincing the automakers that we can get to these mean time between failures, but then you actually have to prove it to the regulators, and that will require, you know, having more cars on the road. So production vehicles, you know, if you can have 10,000, 20,000, 30,000 vehicles on the road, maybe you launch them eyes on, and then you're kinda monitoring the system in the background, you can end up with, you know, actually proving a mean time between failure that's better than humans. So, I think that, you know, our, the way we, o ur approach, that's kinda broken down into modules, the RSS that you mentioned, which creates a lot of transparency, the decision-making is governed by specific rules.
We're able to be very transparent with, like, what happened, you know, when there's an incident or ... you know, even if it's in the simulation world, we're able to kind of understand what protocols the car went through to make that decision. It's a lot different than just trying to feed a lot of data into a big neural net, you know, because then you have no clue kinda what's going on within that. So, you know, we do have., we feel like we have a really good approach to satisfy the regulators, but it's definitely something that is important, and it's gonna take work.
About a minute left, I wanna ask about capital allocation. Company has a clean balance sheet. By our estimates, could maybe be close to throwing off as much as $500 million in free cash flow next year. Seems like there could be a, you know, pretty big funnel of options for allocation. Just how do you think about that internally?
We don't see really any holes in our portfolio, so we don't see the likelihood of any type of major, you know, acquisition. You know, could there be tuck-ins here and there? Yes. You know, we also see, like, that this market is still developing, you know, especially the mobility as a service market. Like, how is that gonna play out? Our strategy is to be capital light. You know, that's really our 100% baseline strategy. Could something change in the market that would drive us to maybe own networks of cars? It's possible. We don't see it as likely at all, but it could be a reason to let the cash balance grow a little bit. But at some point, we'll have to start making those decisions about what to do with the cash.
Okay. Well, fortunately, we're out of time. We'll leave it there for now, but Dan will be available for a breakout session in Salon A here in the 12th floor. It starts in five minutes. Please join me in thanking Dan for the presentation.
Thanks. Thanks, Luke. Good job.
Thanks, Dan.