Good morning, everybody. Thank you so much for joining us. Very pleased to kick off day two of the TMT Conference with a fireside chat session with Innoviz to talk about automotive lidar technology, as well as Physical AI. From the company, we're very pleased to have the CFO, Eldar Cegla, with us today.
Thank you.
We'll go through some Q&A. If you do have questions in the room, feel free to raise your hand, and we will certainly get to you. Eldar, great to see you. Thanks so much.
Thank you.
for being back at the conference.
Thank you. Thank you very much for having me.
Great. Maybe just to kick off as a broader question, maybe for those who are also maybe a bit newer, what is the state of the lidar industry today, both in automotive and in non-automotive, and how has it evolved in the last few years?
Yeah. I think you can put it all, maybe jumping directly into the Physical AI theme. In the past, we didn't know that we are doing Physical AI, but it's very clear that you have these kinds of systems that moved from digital systems, that were residing mainly on our computers using data, images, movies, clips, and now turning into systems that need to interact with the world in real- time and need to do it very efficiently, safely. They need to understand the environment. These are basically the Physical AI systems that require very good perception. For very good perception, they require very good sensors. You have already cameras, which are very widely used, but today, you have more and more 3D sensors, which are the lidars.
The benefit of a 3D sensor for an AI system that once it interacts with the world, it can give it a 3D image, pinpointing every object in the environment. We see actually a very strong momentum for lidar, for the need of lidar. It started with the automotive industry. I think the automotive industry went through a few cycles, a hype cycle, and then a slowdown, and I think today we are in a much more mature state, where the industry understands the problem much better. It understands what it needs to do much better. It understands the need of the different sensors in order to reach autonomy, safe autonomous driving. Also you have the ecosystem. You have the different players, like Mobileye, like NVIDIA, that are doing the brains, are using the sensors in order to make smart decisions.
We are seeing this momentum both in Level 4 for commercial application, like robotaxis, shuttles, trucks, and we do see it picking up again in Level 3 application for consumers, for consumer cars, starting with the Europeans, mainly the German OEMs, and we see more and more momentum in the U.S., and I believe eventually it will trickle down also to Japan and Korea.
Outside of the automotive space, again, under this umbrella of Physical AI, we see the need of good sensory, because basically in every application that you think of using a camera, you actually can use a 3D camera, which is the lidar, which again gives you some benefits that a camera doesn't have, or even coupling a camera together with the lidar has a lot of benefits, and we see more and more applications and demand for that from application like construction, smart infrastructure, smart cities, drones, humanoids, and also security and defense, which became a hot topic. Again, automated systems that use different levels of AI, again, Physical AI. Again, they need good sensors that can capture the environment in 3D and pinpoint every object that is around. For lidars, we see a lot of momentum through these kinds of applications.
Terrific. I want to maybe deep dive a little bit here on the non-auto Physical AI.
Sure.
I know it's still emerging. It still seems like it's relatively early. Of the end markets that you described, which are the ones that you're most excited about in the next two or three years? Ultimately, how can investors size these opportunities?
Yeah, sure. Innoviz has announced over a year ago that it started to pivot or to add additional market segments to its marketing efforts and interests. We were in a sort of an exploration mode, understanding what the market needs. Where can lidar do the best work or fit the best? Where is the need? Where are the low-hanging fruits? We focused initially in application like smart cities, ITS, basically smart infrastructure, smart cities. We looked at construction applications and different kinds because we are also constantly approached from different players. Lately, and this is something also we have announced publicly, we see a growing demand from security and defense application, from situational awareness applications, through perimeter security, through automated systems that need to operate in an environment or navigate in an environment.
We see that there is a real sense of urgency in these kinds of applications and an actual missing link or sensory level, which is not accommodated for by cameras and radars because they have certain shortcomings. Not that the world will stop using these kinds of sensors, but they need to add another sensor, and a lidar is a perfect fit, and we see a lot of momentum in security and defense. Our latest partners are from these kinds of sectors. We are not trying to be a Tier 1 as we are in the automotive sector. In the automotive sector, we are competing directly on programs. We know how to run a program, we know how to qualify for it, and so on and so forth.
In the non-automotive space, and specifically in security and defense, there are very strong players, integrators, there we will play as a Tier 2 or Tier 3, providing our products and solution coupled with software, to these kinds of application. Again, we see there the greatest sense of urgency, the biggest market. We try to size, it's very hard to size these kinds of market in terms of TAM, but if we are looking, the market research is mainly on the system as a system or specifically around well-known sensors like cameras and radars. Our understanding of this market is that it's at least $10 billion and beyond TAM for the next few years up to 2030. Very fast-growing, a lot of interest.
Interesting. Within defense security, maybe could we dig a little bit in terms of what the Innoviz lidar is being used for? Is it mostly perimeter? Is a lot of drone detection, which has become obviously a greater focus? Where is that? I think you mentioned software as well. I'm curious your role in software within defense as well.
Yeah. I guess all of the above, basically, and even more. I think, maybe I'll address a few applications, the ones you mentioned. For perimeter security, there are already solutions in place today that are mainly using cameras and radars. The problem is that the reason why they are using cameras and radars, the radar has a wide field of view, and it can scan a wide area, and once it detects something, it can turn the camera, which doesn't have this field of view and needs to be focused around a very specific area. However, a radar unfortunately has a lot of shortcomings, especially when you take a radar that usually looks to the sky, which is very clean and unnoisy environment, and point it to the ground. There you have a lot of noise.
You have trees, you have vegetation, you have metal object, fences, which radars are not very good at. If the vegetation moves or you have wind, then you get a lot of false alarms, and there are other shortcomings. If you replace the radar with the lidar, you solve the problem. The lidar has the resolution, the lidar has the field of view, it has the resiliency, and it can provide for already an existing solution and much better solution, at the same price point. We see a lot of demand in perimeter security, mainly for civilian application like airport defense, utility defense, but it can go also to special government facilities. It can protect borders if needed.
It seems like a very potent application and with the ability to grow very fast because the integrators have already the programs, the project in place. They just need to set the right technology in place. This we see as a very interesting area that we are focusing today. We are working in the U.S. We have a special guy and team that is working in the U.S. for that matter. Usually, you don't provide just the lidar. You need to provide also some kind of software that interfaces with the lidar that gives meaning to what you see. We have partners that we are using in order to provide also the software side. This means, obviously, that the business model can be not only selling lidar, but also selling software, recurring revenues, and so on. This is on the perimeter security.
Situational awareness means many things. One of the hottest topic, I think, in the world today, you see it in Ukraine, you see it in Europe, you see it in the Middle East, is the threat of drones. The reason drones are such a great threat, because nothing can detect it effectively today, because radars, again, they are very good if you fly high, but these kinds of drones practically literally fly under the radar. It's very difficult to identify them. They are small. They fly very close to the ground, so it's a very noisy environment. How do you differentiate between a flying drone and a bird? You don't want to protect yourself against a bird. It's a false alarm. You would like to identify the drone. Again, here, where all other sensing technologies fail, a lidar is a very good solution.
Again, you need a very good lidar with high performance, resolution, enough range, frame rate, ability to work in outdoors rather than indoors. We are a perfect fit for that. This is also a very hot topic, as I said. We see also a lot of applications and a lot of companies are moving towards offering these kinds of solutions with a lidar. Basically, wherever you need sensing, wherever you think of a camera, think of a lidar doing a better job. Might do a better job.
Terrific. Maybe one more, and then we'll go to a audience question, or maybe a two-part question. One, sticking with defense, maybe broader non-auto, maybe talk about the ASP ranges that you can have relative to auto, and of course, the margin profile.
Secondly, what does a sales cycle look like in defense and non-auto? We know auto can take years to go from the first phone call to a production program. Given the urgency, particularly in the defense, what could that sales cycle look like from the moment you're engaging with prospective customers?
First, to answer on the ASPs. In the auto industry, the name of the game on one hand is volumes. You get these very high volumes. It's still the biggest opportunity for Innoviz in terms of volumes and revenues, aggregated revenues. With that comes a very high pressure on pricing. The industry, at the end of the day, or partially, is working with consumer vehicles. You cannot offer a solution that costs hundreds, sorry, thousands of dollars. It doesn't make sense. The industry wants to see that you scale up to high volumes, but then, in return, they want to see very cost-effective solutions. There is a constant price pressure there.
Once you pivot to other applications like security and defense, A, the volumes are different, at least initially, the volumes, if we are talking about in the auto industry, hundreds of thousands per year, and maybe can go up to even to more than that in aggregate. Currently, what we are seeing in security and defense is thousands, tens of thousands, so obviously, the ASP is much higher. If in the auto industry, for high volumes, you sell your lidars at hundreds of dollars, obviously, if you sell somebody a few hundreds or few thousands, you will sell it at much higher ASPs, maybe at a few thousands, which result in a very financially positive contribution to top line and a very positive contribution to the gross margin.
The margin profile is much better, and basically, you can sell lower volumes, have a very meaningful contribution to the top line, and a very meaningful contribution to the gross margin profile. In terms of the cycles, automotive, yes, is a very long cycle. To win an automotive program may take even a few years, and then the development process takes another two to three years. Until you get to SOP, start of production, may take four or five years even from the day you meet the first time the customer. In the non-automotive space, at least from what we see and the sense of urgency that we encounter today, we see much shorter cycles. Even within months, the decisions are being taken from the moment you meet somebody, then you do a pilot, and to deployment agreement. Cycles here can be shorter.
Yes, in the defense industry, sometimes it takes more time, but currently what we are seeing is a great sense of urgency to bring in good solutions with good sensors. You must remember, I'm acting as a Tier 2 or Tier 3, so I'm meeting the program in a much more mature state than when I'm a Tier 1, and I'm starting very early with an RFI and an RFQ, and so on and so forth. There is a good momentum also in terms of timing.
Perfect. Question?
You mentioned that you are serving as a Tier 2, Tier 3 in these non-auto opportunities, and also mentioned that you're providing the hardware and software solutions pertaining to that piece within the tier supply. When we think about the defense and aerial opportunity that exists for you guys from a detection perspective point cloud, is that something that you guys are working on from a classification standpoint, or is that the integrator that would be responsible for taking the point cloud and creating the perception-oriented classification around that?
Basically, Innoviz offers a wide range of software solution. Basically, we can do perception. We have done it for automotive. In these cases, at this point of time, we are not doing it. We have announced a partnership with a company by the name of Cogniteam. They are doing the perception. They are providing the software solution, but we are the Tier 2, they are the Tier 3. They are providing their solution, and we are offering not only the lidar, but the lidar with the software as a solution. This is the role we are trying to play because there are different applications. We try not to be fast and efficient when approaching this market in our go-to-market strategy, so we are trying to focus on what we are good at.
It's lidar and some supporting software, and then bring some partners to do the perception layer for the software.
If you did, you mentioned you have obviously the capabilities to do it. If you did the Tier 3, Tier 2 componentry of it, would there be an ASP inflection relative to the current offering, or would it be just more like an embedded solution downstream?
Yeah, you are correct. It's an add-on, so we are pricing it as an add-on, as an addition. The lidar itself can be a few thousands of dollars, and then you have the software components, and we are selling it together. The software, we provide all the regular software, like services, recurring revenue elements to it, and it's less as a SaaS service, but more the maintenance service and so on. Yeah.
Maybe switching to auto. Wanted to cover your latest observations on the Level 3, Level 4 sourcing pipeline. It seems like we've been waiting for this sourcing inflection, and it's taken a little bit longer, but we have seen some signs from at least a few automakers starting to put L3, including lidar standard on some programs, I think mostly for a few EV makers. What are you seeing? What are your expectation for decisions and wins along these fronts this year?
Okay. I think we need to differentiate between the German OEMs and the American OEMs. Basically, the Germans are usually also higher, and vehicles in the German OEMs, it's already a standard. It's not a question if there will be an autonomous feature in the car. It will be a Level 3. This is true for Mercedes, this is true for Volkswagen, Audi, Porsche cars. This is true for BMW. This is already embedded as part of their offering. The American OEMs, we are seeing now reengaging in these kinds of offering, Level 3 offerings, both Ford, GM, Stellantis, all of them are thinking, and also the newcomers like Rivian and other players are all into it or thinking how, and announcing also, their timelines in order to introduce Level 3 applications. Maybe Stellantis took a step back in that respect.
Definitely Ford and GM have announced publicly, and also Rivian. We see the momentum is back in place. There was a certain delay, a certain setback in the U.S. market. I see it reinstated, and the understanding that this is important for them. This is a differentiator, and if they don't do it, they will lose the competition with other players from the Germans through the Chinese. I think the Japanese and the Koreans will probably adopt it a bit later. We see a lot of momentum in Level 4 applications all over the world. The commercial application makes sense in terms of the business model, it's different. You take a vehicle, you convert it into a Level 4 vehicle that can drive itself. You eliminate the need for a driver, and then you have a platform that you can monetize off 24/7.
It makes a lot of sense to do the investment because you need to load the car with a lot of sensors. You need a stronger compute model inside the car. It makes more sense because the business model makes sense. We see many programs. We had yesterday, a round table, and we saw many players in that, many newcomers, new companies, new OEMs that are playing in this field, like Kodiak, May Mobility, and others. We see a lot of momentum in the Level 4 commercial application, and this is true, by the way, for the ID. Buzz of Volkswagen, which will launch towards the end of this year, beginning of next year, together with Mobileye, and they are putting a lot of strategic value in this platform.
Absolutely. Maybe sticking with Level 3, the pushback sometimes one gets on Level 3 is, well, the automakers will do it, but on very low volume, like the highest trim levels, most expensive vehicles. Again, we've seen at least a couple automakers sort of say, "Hey, you know what? We're going to do this on all of the next generation vehicles," like the R2 from Rivian being one. Are you seeing any of those signs in your pipelines where the programs that are coming may have higher attach rates, or even, maybe standard is going a little bit too far, but even standard on some vehicles, or just still in the next three, four, five years, going to be limited in terms of the penetration of a particular program's volume?
What is different today than two or three years ago? It's the maturity level and the cost. Today, in order to introduce an autonomous feature, there is, A, much more knowledge of what does entail in order to put this in place, what kind of development, what kind of technology you need, A. B, you have the ecosystem, the different parts or some of them already ready, like Mobileye platform, like NVIDIA platform. It's not that it is plug and play, but you have somebody that has solved for you the problem. Chauffeur from the Level 3 solution for Mobileye Chauffeur, you can simply adopt it or integrate it into your platform. You have the lidar as well, much more mature, so the sensor is in place.
With that, with the maturity, also the cost is reduced, which means you can introduce it into more and more model, not only to high-end, maybe start thinking about mid-range. Do I think that Level 3 will be in the next two year very abundant? No. I do think it will continue to trickle from the high end down to the mid-range, and then to be a common standard because volumes will go up, pricing will come down, and this is the regular process. I think we are now starting to see the recovery, but it will take time, as is the case in the automotive industry. There is one exception, which is what happens in China.
China can be an accelerator for all of these processes because if the world will see that the Chinese are starting to push not only EVs, but these EVs are demonstrating Level 3 autonomous features, I think the Western OEMs will have no choice but to accelerate their efforts. That might happen as well.
Great. On this journey, maybe talk about the InnovizThree. I am curious about maybe three parts to that. One, maybe talk about the importance of the behind the windshield. How many automakers are expressing that as a really kind of a must-have in the next several years, two, maybe the cost of InnovizThree versus InnovizTwo, and lastly, any things you can share on recent feedback, conversations as you're kind of developing that next generation?
Yeah. We are saying that behind the windshield is sort of the holy grail of the industry. They don't like to see this bump of the lidar on the rooftop. They don't like to see the lidar in the grill, which is not the best place. You want to have the lidar higher. Behind the windshield is the perfect place. It doesn't change the design of the vehicle. It's behind the windshield, so you have the wipers, you have the embedded cleaning system, and it makes the most sense. All the OEMs would, if they would have had a lidar that can fit in behind the windshield, they would do it. The InnovizThree is exactly that. The form factor is designed in order to fit behind the windshield.
We even offer to integrate the camera inside the lidar, so you save even additional space, and the integration is simpler inside the car. Cost reduced. We reduced the cost by 35% from the InnovizTwo to the InnovizThree. The performance is as good as the InnovizTwo, and we hope that as we continue to develop this platform, maybe we can improve it even more. The form factor, the power consumption, all the resolution, range, all of that, is a very good solution for the OEMs, and they are looking after it. I don't remember the other questions. Sorry.
Just the feedback niche.
The feedbacks are amazing. Every OEM that we show him the InnovizThree or offer him to test it, there is great enthusiasm. Again, this is something that the industry is looking for, and InnovizThree is really a great product with many lesson learns from all our years of experience, and I think it has gotten a great perception so far.
Great. Maybe two rapid fire in the last minute or so. First, maybe a word on the competitive landscape outside of China, in Europe and U.S. and North America. What are you seeing right now in-
In the automotive space, I would say that the competition has diminished greatly. Many companies either disappeared or stopped serving the automotive industry, and I think there are basically two or three companies that are serving, one of them being Innoviz. We always predicted that the industry will consolidate, and it happened exactly as we predicted, and we hope that Innoviz can keep its seat around the table and be one of the two or three companies that will serve the industry. Outside the automotive space, there are many players which are offering different solution. I think once you bring an automotive-grade lidar with the performance of automotive-grade, range, resolution, power consumption, form factor, it's a very potent competitor in the market. I think specifically for security and defense, there is almost no other lidar that can bring the same performance as we do.
Great. Maybe at a high level, and without providing guidance, how would you frame the 2030 opportunity, whether it's units, penetration, if our investors to think about what the next few years might look like?
At least our projection is hundreds of thousands of units a year. Maybe north of 500,000 , but we still need to win these programs and aggregate the volumes, but I do believe that we are on this kind of trajectory of hundreds of thousands of units towards 2030.
Terrific. Great conversation. I think we have to leave it there, but thank you so much.
Thank you.
for being back in the conference, and thank you all for your participation.
Thank you very much.
Pleasure to have you here. Thank you.