Hey, everyone. I'm George Gianarikas, one of Canaccord Genuity's sustainability analysts, and thank you so much for joining our first sustainability summit, and we're very excited, enthusiastic about having Dan Galves here from Mobileye, Chief Communications Officer. Dan, thanks so much for joining us.
Thanks, George. Appreciate you having me.
So since we have just 20 minutes, maybe to focus first on the momentum that you could be seeing with SuperVision and Chauffeur wins in addition to Surround ADAS. Where are you in that process? I know we talked about it a lot on your earnings call and at your Analyst Day, but any updates there would be appreciated.
Yeah, maybe starting with Surround ADAS, it's kind of more of a kind of typical cadence to RFQ process. There's competitors in there, so we kind of understand the process. It's been more predictable. It's been more predictable in terms of the timeline, and we feel really good about this segment, which we kind of see as the next generation of ADAS. And I think competitive aspects, like from the Chinese and kind of looking to kind of standardize this type of feature set, is, I think, really important and makes it definitely feel like the early days of ADAS, where you started to get companies making it standard instead of optional, and that really drove adoption to higher levels pretty quickly. So yeah, no, we feel good about kind of where we are on the Surround ADAS timelines. And like I said, they're more predictable.
SuperVision and Chauffeur, I think it's the same. Nothing's changed from kind of what we talked about at the Capital Markets Day. Again, the timing is a little bit harder to predict because these are not really normal RFQ processes. There's typically not another competitor. It's more you're competing with internal software development. You're competing, in some cases, with, do we want to do this or not? What's the timeline? What types of vehicles? What price point? How are we going to market it? So I think it's a strategic decision by the OEM, so it makes some sense that they would take a while, but we're very eager, obviously, to get these over the line.
Have product portfolios from the OEM stabilized enough now? I mean, they wanted to go fully EV by 2030, like two years ago. That's obviously changed. I would guess that that change in product portfolio caused also some delays in choosing who their autonomous provider would be. Do you see those OEM product portfolio decisions as having stabilized now?
I think it has stabilized. That period where product portfolios were in flux, like what powertrain mix do we want? What platform should come when? Should it be a combination of powertrains? I think really put the progress at a standstill for a number of months. This was late 2023, quite a bit into 2024. I think the other impact that even though now they're stable, a lot of these kind of strategic platforms, I think, have been pushed out a bit. So kind of the sense of urgency to develop a next-gen ADAS or AV or whatever you want to call it for a 2026 platform that was strategic for the company, maybe that platform's in 2028 now. So it took away a little bit of that sense of urgency to kind of get the decision made.
But I think we are stable, so we're not seeing that as a headwind anymore.
In terms of these decisions from the OEMs, I mean, there's so much change in technology now, just specifically within autonomy, right? There's all this debate around end-to-end neural networks versus a hybrid approach. Which sensors do you want to use? Waymo's had some incremental success from a robotaxi perspective. Are you seeing maybe some of these insurgent competitors sort of in some of the bake-offs or some of the competitive dynamics that you're seeing within the OEMs? And importantly, to the extent, let's just say this decision gets pushed from a 2028 product to a 2029 product, or maybe there's someone else that wins a deal with OEM X, Y, or Z. Where does that put you from a competitive perspective? Do you worry that if you don't win this set of five deals, whatever it is, that somehow you're on the back foot for the next 10 years?
Yeah. So I think let's put Surround ADAS aside because, like I said, the competitive landscape is pretty similar to ADAS in a lot of ways. And I think it's very critical for us to win these deals because I think it's right in our sweet spot. You're talking about high volume. You're talking about ability to execute, kind of probability. You're talking about needs to be very cost-optimized. So I would say if we don't win Surround ADAS deals, to me, it would actually be more concerning than SuperVision, Chauffeur, what I would call delays. I don't see the OEMs that we're working with. I don't see them choosing a different direction. I think it's possible, to be totally frank, that some would maybe push the decision or say, "Not sure if we need this right now. Maybe we should give our internal team more time.
Let's kind of see what happens." To be very clear, that's not the messaging that we're getting from the OEMs. So I think if companies decided to delay even further, then I don't see it as a big negative for us because I think we have the Porsche Audi launches coming. That's going to be a huge proof point for the industry in 2026. So you would see kind of the results of the EyeQ6-based platform and all the new software engines that we have and kind of the execution collaboration with these very kind of high-level teams at Porsche and Audi. So I don't see it as something really negative. We're not seeing kind of, I mean, we see startups as maybe inserting a bit of noise into the process, especially at the high level of companies.
Like, "Hey, should we look at this new technology or these guys more kind of on the front foot?" What we hear from kind of the product teams within these companies is, "We don't see the ability to execute here," and it's like a little bit of a chicken and an egg situation where those companies, in order to prove execution, you need a customer, and that's really tough to get, we can tell you, right? So yeah, I think that we're very confident that we'll get kind of announcements and design wins on SuperVision this year. We're extremely encouraged by the progress we're making with both the Porsche development team on SuperVision and the Audi development team on Chauffeur. We're seeing very kind of positive indications in the test environment with the new software for EyeQ6 High, so we feel really good.
Yeah, I don't think that there's anything that's a real winner-take-all aspect of this.
I'd like to discuss your sensor approach because you've had some changes, I think, over the last several months. You decided to, I think, more or less abandon FMCW LiDAR for now, and you're very excited about your Imaging Radar solution. So how do you look at the sensor set progressing? Forget the next few years, into the late 2020s and 2030 and beyond. Do you need LiDAR at all? And will you be strictly relying on the Imaging Radar to do a lot of that work?
Yeah. So I think I would say, first of all, we're camera primary, camera-centric, and we're going to remain that way. And I think you can do, and kind of new AI has really led to the ability to do more with cameras than maybe you have in the past. And I'm talking more about creating a variety of software engines and software approaches within the computer vision system to set it up so that you have to have kind of two failures within the same task in order for even the camera system to fail. So we talk about redundancy, and most people think of it as camera plus radar, but there's also a lot of redundancies within the camera system. And you can essentially take buckets of edge cases, like the unusual-looking vehicle that's not really in your training set.
If you have an appearance-based software engine and kind of the end-to-end AI you're using doesn't recognize this vehicle as an object to be avoided, the appearance-based system will, right? So I think that camera is primary for us. It's the lowest-cost sensor. It's the highest resolution. And we have kind of figured out ways to do more with camera than we probably envisioned two or three years ago. But we feel very committed that you also need a second perception system that has different failure modes than the camera system. And this is kind of primarily why radar is more important to us than LiDAR, because radar really has kind of very complementary or independent failure modes. Weather is the easiest one. Again, weather is like an edge case on its own, right? And so if you're having an issue, camera is affected by weather.
LiDAR can be affected by weather. Radar is not, right? So having that kind of high-resolution radar through the imaging radar, the ability to create a kind of environmental model, a sensing state just from a radar technology means when you do encounter bad weather, you've got this really robust backup system so it doesn't impact the system overall. We also think LiDAR is important, but kind of the success and kind of what we're seeing in kind of the development of the imaging radar, the fact that it's meeting the specifications that we were targeting to begin with in more of a production environment gives us more confidence that we aren't going to need that much from LiDAR, which gave us the confidence that we wouldn't need a next-generation LiDAR technology and could rely on off-the-shelf time of flight. Not really off-the-shelf.
I mean, there's definitely LiDAR companies that are better than others and we feel like there's two or three non-Chinese ones that we can use and kind of get the information we need to make the system work. Yeah, I think the imaging radar plus camera is kind of the mainstream direction. LiDAR provides that extra distance, another high-resolution sensor. We're certainly going to take advantage of it, but it's probably not as important as the other two.
On your recent earnings call, you talked about sort of a stabilization, excuse me, in business trends. Obviously, 2024 was a little bit tumultuous in terms of just how you started the year and then in June. How comfortable do you feel now that at least some of the order rates and the guidance and the business can sort of inflect higher from these levels?
Yeah. I think that there's maybe three dynamics to talk about. One is SuperVision. We were not expecting our kind of main customer, Zeekr, to replace us during 2024. They did. That created a lot of noise. That created downside in kind of our expectations. We've kind of removed that risk from the table this year. Domestic China OEMs, we do feel like we have more of a committed kind of volume number for this year based on some work we did with the customers last year to sort of combine support for the export business with commitment to a certain level of volume in the domestic market, and so again, we feel like that's more stable, less kind of apt for surprise.
And then I think the thing that we needed to kind of come to terms with is we need to incorporate that our core customers could continue to underperform mostly in China. And so kind of the magnitude of headwind that we encountered last year is not baked into third-party forecasts for this year. So we baked it into our own guidance. And so I think so far we've seen good stability in the market in terms of production levels. The order flows have been stable. And so kind of we're feeling good about where we are in that we could do better than expected if things stay this way. But if there's deterioration kind of in conditions, we can also withstand a certain level of that as well.
I'd like to ask about the recent announcement from BYD and their God's Eye product. What do you think that means for the market? What system are they using? Any additional details and maybe business implications would be helpful.
Yeah. So our view, and I think we have pretty good information, is this is almost 100% kind of an in-house developed system. I think that they may have gotten some help from kind of a software provider, but not much. It's the functionality of kind of the base system, which we think is going to be where 80%-90% of the volume comes, is very similar to a Surround ADAS feature set, right? Kind of on-highway, control steering, control speed, automatic lane changes, but probably not very scalable to off-highway, not scalable to eyes-off. So it's really kind of what we've envisioned for this Surround ADAS product and cost-optimized, right? Kind of quite vertically integrated. They've got good scale, so they get good component costs.
And I think for our customers, in order to compete, you need to take the same philosophy in terms of taking advantage of scale. And we have the scale, right? Because I think that we have the connections with the camera suppliers. And if this Surround ADAS product becomes millions and millions of vehicles a year, then that leads to better pricing for EyeQ6 High and for all the components within the system. And so in comparison to current Level 2+ highway systems, which are four or five different ECUs, multiple suppliers, every system works a little bit differently, you don't get the benefits of your industry scale there. So I think we're actually quite encouraged. I think it highlights how, in some ways, it highlights the speed of the Chinese and kind of the aggressiveness of the Chinese.
It may kind of impact negatively kind of the perception of our customers. But I think in reality, what this means is it's kind of a shot across the bow that this type of functionality needs to be standard across many different vehicles, not just sold as an option or targeted here or there. So it's too early to see any real impact. But I think our view is that this could lead to kind of increased standardization of this type of product set. And we are very ready to kind of serve the market with a very kind of cost-optimized, efficient, high-quality, high-performance system.
Can you remind us of where you expect your Surround ADAS ASPs to land?
In the $150-$200 per unit for an EyeQ6 High. Our intent is to go to market as a tier two. So there would be incremental content provided by the tier one, like in terms of the ECU, which we'd be very kind of involved with the design, the cameras, which we are the ones who kind of choose. So a lot of it is driven by our reference design. But overall, we think that this can support somewhere in kind of a $600-$700 overall system cost, which I think is very optimized to compete very well with the Chinese.
Maybe last question, the Tesla FSD version 13.2 question. When your SuperVision product hits the road, I mean, I know that that's in the future, and for Tesla, that's now. But how competitive do you expect that product to be relative to Tesla's, at least current version of FSD?
I mean, I think a pretty common kind of comment we got at the Analyst Day in December in Munich was like, "Wow, we kind of expected this to act like one of the kind of OEM-directed systems, but it's just as good as Tesla FSD, just as capable, just as kind of broad in terms of the environment it can operate in, the human-like character of it, the assertiveness." And this was like an EyeQ5-based product, which was developed like 2019, 2020 in terms of the core technology. So I think all of the new software engines, the EyeQ6 High processing power, the transformer-based architectures are going to lead to a leap in performance with the EyeQ6-based system for SuperVision and Chauffeur. Obviously, Tesla will get better as well over the next year, year and a half.
But we feel very capable of competing and having just as good functionality. And this is kind of one of the key questions that our OEM partners are asking or evaluating is how will it kind of compare to Tesla FSD? So Tesla FSD is a great system. We have tons of respect for what they've accomplished, and it will improve. But we feel like ours will be just as good, if not better, when it launches with Porsche and Audi.
That's a great place to stop. Thank you so much for joining us, Dan. We appreciate it.
Thanks, George. You too.