Innoviz Technologies Ltd. (INVZ) Earnings Call Transcript & Summary

We'll get started here for day 3 with our hardware companies. I'm Samik Chatterjee, I cover tech hardware at JPMorgan. I have the pleasure of hosting Innoviz. And from Innoviz, we have Oren Buskila, who is the Chief R&D Officer at Innoviz. I also have Rob Moffatt here from the Investor Relations team. Thank you both for coming to the conference.

Let's start with a longer-term question. Obviously, you're a LiDAR company and there's been a lot of discussion around LiDAR and its position on a vehicle. But we've obviously seen for a longer duration companies that have camera-based ADAS solution sort of have a position on the vehicle as well. What are you seeing from OEM customers today? How much of willingness are you seeing to have camera and LiDAR side by side on the vehicle? And today, if an OEM is saying, I choose to remain camera only, what is that argument that they're making to remain camera only rather than adopt the LiDAR side by side with the camera?

So obviously, cameras -- automotive cameras are very mature technology, and they have been very useful in bringing ADAS to the level that it is today. But going forward to reach level higher of automation, Level 2+ and Level 3, which is actual self driving of the car, there's a practical consensus that is just not practical to do with a camera, mainly because you just -- you can't get to the level of accuracy in identifying obstacles and reaching the safety that is needed for Level 3 applications. So basically, we're seeing practically all OEMs now working to integrate LiDAR in the future ADAS Level 2 -- Level 2+ systems with the one exception that everyone is familiar with.

I mean just to clarify then, is your understanding today that if somebody -- an OEM is staying with Level 2 with camera, they're just targeting more of a Level 2 solution in that case?

Precisely. I don't think anyone now is targeting Level 3 with the camera-only based system, you need the redundancy. You need a camera and a LiDAR and a radar to reach that level of safety.

Okay. Okay. Maybe I can follow up on that. Just another point that comes to mind is what is that level of sort of performance difference or safety difference you've been able to prove out when you put LiDAR side by side with the camera? What is that magnitude of difference on the safety aspect that it makes?

So the capabilities of ADAS systems have been improving over the years. But really, the limit is being able to transfer the control of the car to the vehicle itself, to allow the driver to not take part in the driving at all under certain conditions. And this is exactly the line between Level 2 and Level 3. Level 3 is real self-driving, real autonomous driving, where the car assumes the responsibility and does not require the driver to pay attention to the road. And this is only enabled by LiDAR, which is -- which delivers a 3D high-resolution imaging of the scene unlike a camera or a radar.

Okay. I'll come back to that topic in a bit once we go through a few more questions. But let's move to some of the recent wins that you've talked about based -- I think, on the last earnings call, Omer mentioned this is based on some of the conversations you have with customers, you're finding further reinforcement of the advantages of 905 technology over 1550. Can you outline what you're hearing from customers on that front exactly, how much of that is a pricing decision, which we obviously understand 1550 is more expensive to manufacture versus a performance decision?

Sure. So OEMs don't really care about the wavelength off of the LiDAR. They don't care about whether it's 905 or 1550 or 1330. It doesn't really -- it doesn't matter to them. What matters is performance, it's cost, it's power consumption, it's size and reliability. And up until now, 905-nanometer LiDARs had most of the advantages, particularly cost and size and power consumption. And in the past, there was a certain gap in performance for 1550-nanometer LiDARs did reach higher range, which is important, but that is no longer the case. So with the InnovizTwo LiDAR from Innoviz, we've actually solved the only challenge that stood between 905 and reaching the top level of performance. So the InnovizTwo now has the performance of any 1550-nanometer LiDAR, but with all of the other advantages. So this is why we're seeing now OEMs, they don't need to choose between performance and cost.

That's interesting to mention that because I think for a long time the argument has been that 905 doesn't really have the same range as 1550 because you can't put the same amount of power in it because it's closer to the visible sort of wavelength, right? So how do you get around then overcoming that hurdle that you -- that 905 has relative to 1550?

The issue was indeed eye safety, that the human eye is more sensitive to 905-nanometers, which is true. Innoviz has a patented technology that allow us to transmit basically almost as much power as we -- laser power as we want or need. So we've solved this. The exact way in which we do it can be read in the patent basically. But we will ultimately now not limit it in terms of the transmission power and we can reach the same level, same performance and even higher than 1550-nanometer systems.

Okay. Just to then go further down 1 level on that same sort of question about price versus performance. At the same time, there have been arguments about price for 1550 not being able to scale down as quickly over the sort of next few generations as 905. Where do you think you are today in terms of what the price of the sort of the bill of materials looks like, how quickly can it scale down relative to 1550, that still seems to be a bit more challenged on that front?

So the -- there is simply no way of building a 1550-nanometer LiDAR today at the same cost of a 905, with a central reason being the cost of the laser itself, the 905-nanometer laser is simply a diode. It's a single component, a semiconductor component, costs in the range of a few dollars, sub-$10. And for comparison, a 1550-nanometer laser is a system by itself. It's -- there's a spool of fiber, and there are several charge pump lasers. So alone, it costs above $500 at scale. So there's just no way to reach the same cost. And this is how we're able to offer InnovizTwo for a price lower than $1,000 per unit and even as low as $500 depending on volume.

Samik, let me just jump in too because I mean I talked to Omer about this the other day. And the conversations that we had, going back to the program where we displaced the competitor, performance is binary. Either you have the performance or you don't. And price is a secondary factor after that, right, because price gets down to product-market fit. And from what we heard from the OEM where we displaced this development stage competitor, everything that they talked about was on the performance side, not on the price side. And this is a Level 4 program where you're putting 3 to 4 LiDARs on the vehicle, right? So if anything, price would be magnified in a situation like this, and the conversation wasn't about price, it was about performance.

Let me add something on top of what Rob said. 1550 is known for its range, right, because you can transmit high laser power, but not everyone is familiar with the more specific requirements from OEMs, there's much more than just range. There is resolution and field of view and frame rates. And if you go and compare those, the whole set of KPIs between our 1550-nanometer competitors and us, you'll find that there's really no -- there's no competition, in fact, InnovizTwo is much, much more powerful.

Okay. Let's talk about R&D priorities for the company today. The reason to ask is, obviously, we are in this sort of tough macro backdrop, and you have to maintain a balance between spending too much relative to also trying to maintain the leadership that you have on the technology side. So how do you sort of think about the priorities in that context of what you're spending most time and resources on?

So our top priority is now the launch -- the upcoming launch of the InnovizOne with the BMW Series 7. This is happening this year. So there's a lot of effort going towards this. We're really at the final stages. Secondly, we are working on the industrialization of the InnovizTwo. So we have already -- we're manufacturing samples at pretty high volume actually today of the InnovizTwo delivering to customers. And we're working on basically designing or improving, tweaking the design for mass production, both the LiDAR itself and also the -- all of the supporting machinery for assembling and calibrating and testing the LiDAR. This is what we are now mostly engaged with. And the third area is the software development where we're working on our perception software based on the InnovizTwo. The software package that we're delivering with InnovizTwo is much more capable than what the InnovizOne had, it was much more complex. And on top of that, since we are providing the InnovizTwo as a Tier 1 company, we're also responsible for all of the, I'd say, automotive standard software known as AUTOSAR. So this is a big, big increase in our -- what we need to do. And I think not many people know, but the Innoviz R&D is split almost half and half between hardware and software. So there's a big, big software effort towards that.

And the recent MRM software that you sort of talked about, is that -- now when you sort of think about it, is that part of the perception software you classified? Or is it more of an automotive standard sort of AUTOSAR software? Where would you really classify it? What led to that sort of building that product to cater to that -- those OEM customers?

So MRM stands for minimum risk maneuver. So basically, it's a feature where the car needs to perform some maneuver to bring it to a safe stop or a safe state. And we are now getting -- we have gotten recently requests from certain OEMs to develop MRM based solely on the LiDAR. This is as a backup to the primary system. And we actually need to develop both sides of the software, both the software, the perception-related features required for the MRM, and also the automotive standard software, which is -- it's the first time for us needs to actually actuate the car controls, the steering wheel, the brakes, the gas. This is the first time we're going to do that, and it's a big challenge, but we're happy to do that.

And I'll just add to that. It was part of an RFQ, right? So earlier in the year, we pointed to more than 1 RFQ. We just delivered a program on our 1Q earnings. This was part of one of the other programs that we thought was very close. They added it to the RFQ process. This is why it's taking a little bit longer than we thought. But it's very encouraging to have your customer ask you to develop something as a part of the RFQ process, right? It's a signal that the RFQ is going well. It's a signal that they believe very highly in the technology, especially if this is the first instance of LiDAR doing an MRM. And as a result, we're now introducing it to other automakers as well that are in the RFI and RFQ process.

Interesting. And sorry for digressing here a bit, but when you're thinking about, firstly, you said this is -- the MRM is based on the LiDAR system independently as the backup option clearly implies that the OEM is looking to sort of put LiDAR side by side with another sensor system and use LiDAR as the backup. So that's sort of how I'm reading it. Now secondly, so when you think about a safe stop, eventually, you will also have to sort of integrate with some level of mapping data to find where the safe spot is, if I'm thinking about it correctly. Maybe sort of what are your plans around that? Are you continuing to go to partner with different sort of mapping partners on that? Or would you go the route that some of your competitors have where they're developing their own mapping solutions?

So specifically for MRM, mapping is not really required. This is a maneuver that needs to be done in a matter of seconds. So the only thing -- think as a driver, right, when you -- when the car suddenly stops in front of you, you need to react very quickly. You don't really care about whether you're 5 miles or 6 miles from the destination, right? You just need to stop safely. So this is what the MRM needs to know about, right? What's ahead of you, what's next to you? And this is -- so mapping is not really part of it. We do need to develop also special hardware for that because the hardware needs to be -- besides the LiDAR itself, there's another hardware component or box where the MRM is going to run on. It's an ECU. So this is another part of the project. As for mapping in general, we are working with customers who are developing mapping applications based on our LiDAR. And for now at least, we're not planning to take part in this development ourselves.

Moving to the next topic. The win that you have with Volkswagen on the CARIAD platform. How are you thinking about the opportunity in relation to expanding across different brands, different price deals and also different parts of the technology stack now that you have like additional software that you can offer?

Yes. So the win -- design win with CARIAD and VW is obviously a great milestone for Innoviz. And VW, Volkswagen is one of the top 2 automotive players in the market. And CARIAD is the company inside the VW corporate that is responsible for software development for all of the brands, many brands inside of VW such as Audi, Porsche and many more. And being chosen by CARIAD means that the LiDAR is now going to be the standard for any model in VW that's based on that platform. So that nomination was actually for several brands and for several models of each brand. And this is still just the first platform. We now have the opportunity of expanding to further platforms by VW. So that shuttle program, for example, that we announced is another such platform and there are many more. And we see that those brands benefit a lot from the fact that there's a single sensor that's now common between them, so they don't need to duplicate the efforts, development effort, and they can reuse whatever CARIAD has been using. But despite that, we actually work also directly with some of the brands because each one might have some specific requirements. Think about how Porsche looks differently than how an Audi looks like. So they have special requirements as for where to install the LiDAR, what it should look like. So we're actually working with several of them to make them the best fit.

And maybe more on the price tiers. When you think about getting those wins across and sort of having the LiDAR across vehicles at different price tiers, do you see more of a difference in terms of how many LiDARs per vehicle the content opportunities? Or do you see more of like the OEMs are deciding this is going to be either a standard on this vehicle or option based on the price tier, not really changing the LiDAR opportunity on a per vehicle basis?