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

To introduce our next session. We're very happy to have Innoviz back with us at the Citi Tech Conference for a fireside chat. From the company, to my left, I'm very pleased to have Eldar Cegla, the company's Chief Financial Officer; on his left, Rob Moffatt, the VP of Investor Relations and Corporate Development. Lots to talk about in the world of ADAS autonomous and, of course, LiDAR. We'll keep this to a fireside chat session. If you do have questions, feel free to raise your hand. We'll get you a mic and we'll get your questions answered. So with that, Eldar, Rob, great to see you both. Thank you so much for being here.

Thank you.

Let's get to the -- maybe start off with the first question a lot of companies in the LiDAR space get, which is the latest read both on demand broadly as well as the competitive dynamics, what you're seeing and how you're seeing the competitive dynamics involved.

So first of all, regarding the market, we see maturity, the processes of the industry, the OEMs who are sourcing technologies that will support autonomous driving is maturing. We see RFIs, the process of going from RFI, RFQ, and then nomination and start of production, we see the industry is going towards the RFQs. The RFQs process, which is usually a finite process, takes 6 to 9 months, are now being in the midst -- we are in the midst of the process. We see it progressing very nicely. A lot of due diligence. And we believe that a lot of decision will be made in the next months in the industry. So we are seeing the volumes. We are seeing the clients converging. And we believe there will be decisions. In terms of competition, I think, Innoviz is, generally speaking, in a very good shape. I think on the technology side, on the cost side, Innoviz has a lot to offer. I think, I will say modestly, we have maybe the best-in-class LiDAR in terms of performance, from resolution, range, size, power consumption and so on, I think we offer the best in class. I think also the cost, which the industry is very sensitive to cost, we are at the right place where -- for what the industry is looking for today, Level 2+, Level 3, one LiDAR per vehicle, some programs that are more like shuttle program like commercial vehicles, also we provide solutions there. They use more than 1 LiDAR, less sensitive to cost. So we are in a good position. I think the competition -- where we see more competition, is our position as a Tier 1. This, I think, is a more challenging aspect that we took upon ourselves. We became a Tier 1 more than a year ago. Volkswagen nominated us as a Tier 1. Since then, we are competing on programs as a direct supplier. This comes with a few challenges. The OEMs expect us to be not only the best-in-class technology, best quality production and so on, they need to see a company that is robust, a company that can support their program for many years, from the development stages to, later on, the SOP, the start of production and supporting the vehicles once they run around. So -- and in that respect, we see our main competition rather from other LiDAR companies or pure-play LiDAR companies to Tier 1, the main one, we usually mention it, is Valero. We see Valero in many of the RFIs, RFQs that we are competing on. And we are hoping that Innoviz is able to solve both on the technology side but also be -- show that it can be a very strong and reliable Tier 1 for the industry. And it connects maybe to the -- one of your next questions on the follow-on.

I'll double down on a little bit on the maturation piece because it's something that I think a lot of investors are missing, right? Because for -- [Itay], we've been talking about LiDAR for how long, right? 10 years ago, and it was Velodyne and they were spinners, and they cost $75,000. And there's a lot of investors, when they think LiDAR, they're still thinking about that. Then we kind of had the first generation awards for us, it was BMW and others, it was Volvo and GM, et cetera. And those programs came in and a lot of them came in a little bit slower than expected and the volumes weren't as high as expected. But what we're seeing right now is the pivot from RFI to RFQ, that's one aspect of it, right? Once a program is in an RFQ, it's going to be -- it's a vehicle that's going to be on the road, right? An RFI, you're not sure whether it's going to end up, the time line is a little bit more nebulous. Once you get into the RFQ process, there's a decision being made actively by an OEM on a LiDAR. The other thing that we're seeing is that the RFQs that are out now are incorporating more vehicles, right? They're going to deploy LIDAR more broadly and at higher volumes. That's another important step in that maturation curve that I think investors might not appreciate because they're not seeing the RFQs themselves directly.

Terrific. Maybe while on the topic of the RFQ, we can dig a little bit more into that, maybe continuing on the competition side for a moment. How many other players typically are involved in the RFQ process? Is it 2 or 3? You mentioned 1 competitor, I don't know if you want to mention more. But I mean, what is it now in terms of the players getting to that final stage?

So usually, when you start the process with the RFI, the request for information, usually, you see, I would say, 6, maybe 8 LiDAR companies, the OEMs try to find out what is available, what kind of information they can get from the industry in order to understand where they need to target the RFQ. Once you get to the RFQ, usually, you get shortlisted. It means they take you -- your solution seriously. Usually, you see 2, max 3, competitors in that -- in the RFQ stages. And by then -- the RFQ is a very detailed process where you undergo many, many audits on different aspects from the technical and you need to answer from the technical, to production, quality, and you need to meet those -- meet or not meet those requirements. We feel they are very comfortable. We answer -- we have in our pipeline many RFQs. We are used to -- we have a lot of experience answering these RFQs. At least on our part, we know that in some cases, we overspec -- our LiDAR overspecs the requirements of the industry. So we are comfortable there. And as I said, the main challenge is being a Tier 1, convincing the OEMs that we are there for them, we can support their programs. And I think we've done it better and better over time. And I think now that these processes are maturing, we are able to at least have a good chance to winning many of these programs.

Got it. So you mentioned a couple of times, so the challenge competing in Tier 1, convincing the OEM that you're there for them. Was that part of the rationale for the recent capital raise that you undertook? Is it trying to sort of bolster the position as part of these audits? Or is that separate from that?

No, it's exactly for that. I think the follow-on was very important for our customers. So when we are doing these -- or participating in these RFQs, I'm being audited, the financial side of the business is being audited very carefully. And it's always an issue. It was an issue with BMW, it's an issue with Volkswagen. It's an issue that is debated with the RFQs that we are now competing on, some very serious OEMs. And the fact that we were able to secure additional funding, strengthening our balance sheet, signal to the market to our customers two things. A, obviously, we have a stronger balance sheet. But B, we have the access to capital markets, the story is heard well and is being well accepted by the market. And we can always, if needed, tap into the capital markets and get additional funding for the company. So we got very positive feedbacks on this follow-on. It relaxes many of the processes that we are in right now. And I think it was well balanced between the size that we chose to target. Actually, we were able to raise more than we targeted, and the needs that -- or the signaling that we did to our clients. So it was well balanced, well received, and I think it's...

I'll layer onto that a little bit, right? And I think it's important to think about the cadence of the year, right? Because coming in the first quarter, one of the most important messages on our earnings call, again, was the step-up from RFIs to RFQs. The time line that Eldar laid out, 6 to 9 months, if something goes a little bit slower, maybe it's 12. If you think about that time line to where we are today, it's fair to assume that we're moving into the later stages of some of these conversations going into the fall and toward the end of the year. right? When you get into the later stages, you have a financial due diligence aspect to that conversation that you're having with the OEM customer. We wanted to go into that as strong as we could. We showed 2 things. One was the quantum. We topped off the balance sheet. The more important aspect was showing that we have access to capital and our investors support us. And that was an important message for the OEMs themselves to hear.

Terrific. So it sounds like this is sort of a late-stage audits and maybe we're months away perhaps from these decisions.

We are. We've been communicating. We're very close. If you look at our guidance this year, we're targeting 2 new brands to bring in. That's part of our targets. And if you think about it, let's even go back to one of the earlier mentions, right? There is a competitor that we find ourselves going up against more often than not. That's the only competitor that we lose any sleep over at night, and it's not on the technology side. We're not concerned with their tech. The thing that competitor brings to the table is the fact that they've been around 100 years, and they have a big balance sheet and access to capital. Part of our thought process was if we could remove any concern on the balance sheet and access to capital, we could shift these conversations entirely back to the technology side, where we have a high degree of confidence that we can convert these OEMs into customers.

Got it. What can you tell us about these programs to the best of your ability in terms of regions? Maybe it's too early to ask, but whether LiDAR comes in as optional, and maybe the SOP or targeted SOP of some of these RFQs.

So also on this aspect, we see the industry maturing. Back when we won the BMW program, which is launching this year, this is an optional upgrade that you can have with the BMW cars that you can purchase towards the end of this year. Going to the next program we won, the big program we won, Volkswagen, it's part of it, part of the brands and part of the car models will be as a standard part probably as an upgrade. So this is starting -- the technology starts to mature. The OEMs getting more and more confident that they need it. This is an important differentiator. Looking at our current pipeline, so you got, I would say, all the names that matter in terms of OEMs. We said it in the past, a major part of the 10 leading OEMs are now in pursuit of autonomous driving, of sourcing LiDAR sensors, the whole platform. So it's very no-name, important name, and the geographies, basically everybody, additional European OEM, American OEMs, Japanese, Korean, China. So we are very confident that, again, as I said at the beginning, the market is there. The market has matured. It has taken a strategic decision going towards autonomous. And we are hopeful that we can bring in these programs as much as possible and be a clear leader there.

And a question I think that often comes up in the RFQ discussion is that there are -- a few other companies would talk about, we have x amount of automakers in our pipeline. Are you aligned with all the sort of compute providers out there? Is there 1 or 2 that we should think about Innoviz sort of wins with these? Or is it kind of spread out pretty evenly?

You mean the Qualcomms and the NVIDIA. So there are 3 main platforms that are leading the industry. You have Mobileye, you have Qualcomm, you have NVIDIA. Each of them, well, has programs with different OEMs. Some of the OEMs use more than 1 platform. So we have already announced that we are well integrated or have a program with Qualcomm. This is when we won the Volkswagen deal. Volkswagen works with more than 1 platform. We have very close ties with both of -- the 2 others which we are working on. We didn't say which of them we have a program that we have won and work -- or started the integration. So with Qualcomm, we have the integration, with another of the 2, we have integration, and we have very close ties. We announced that NVIDIA we started working more closely in order to win additional programs. So I think this is very important for the OEMs because the OEMs care about cost, obviously. They care about the time lines. So how long does it take to integrate a new platform into the system. And they care about risk. Risk reduction is a big factor in everything they do. And if you're already working with one of these platforms, this means a tremendous risk reduction. Obviously, if you already are integrated to one, the next decision is much more easier. And this is another benefit that we bring to the table.

I think one of the key messages to it, [Itay] we have series production wins with 2 out of 3, right? There's a lot of other LiDAR companies out there and they're saying, we're working with or we have a project or we have a program. What you need to clarify is if it's a series production win, right? That means it's a vehicle that's going on the road, that's going to be produced in very large volumes and sold to customers. It's a very different hurdle. It's a very different signal than saying, we're working with so and so. And that's something I think a lot of investors sometimes don't fully appreciate.

Yes, absolutely. And one more question in terms of thinking about the next few months in the RFQs and I want to get into software. How important does your current ramp of the InnovizOne play into decisions on these RFQs for the InnovizTwo in terms of just smoothness and just making sure like, hey, they're doing with -- well with the first program, so that the proof point or is it just such a different manufacturing process and product that is almost independent?

They are not completely independent in the technology respect. A lot of carryovers from InnovizOne to InnovizTwo. So a customer that is working with InnovizOne knows that for the next-generation program, it doesn't not need to start from scratch. There are a lot of commonalities between the 2 technologies. And InnovizOne we have 2 programs, the BMW, which is launching and the shuttle program, which is planned to launch this year as well. As of the Volkswagen deal, we are competing, we are offering the InnovizTwo basically to all our current clients or potential clients. The reason for that is it's a cheaper, more cost-effective product and high performing. So it makes a lot of sense for the industry to take InnovizTwo rather to stay with the InnovizOne. With the InnovizOne, we played as a Tier 2. So Magna was the Tier 1. Magna is doing the manufacturing, we are selling the components, the ASIC, the MEMS, the detector. Magna is manufacturing LiDAR and selling it on to BMW. Becoming a Tier 1 means that we are providing the whole LiDAR. And we can discuss about a little bit more the offering that we bring to the table, the offering per car is increased. We are offering not only the LiDAR, more software, more hardware.

It's just a dramatically better product, right? We talked in the past about the 70% cost reduction from InnovizOne to InnovizTwo, right? A big part of that was moving from 4 lasers, 4 detectors, down to 1, and some other engineering changes that we've made. We've also talked about a 30x performance in terms of range, resolution, et cetera. And that was before our second-generation custom ASIC, which we just announced last quarter. We finished shaping it out, that's going to extend our range on the InnovizTwo from 300 meters to over 450 meters. And in terms of resolution, we don't know the exact number yet. We have to get it into the LiDAR and test it, but we think it can bring us basically to the equivalent of a high-def camera in terms of resolution, with the range and the benefit of a true 3D map. So we're looking forward to sharing more on that very soon.

Yes, definitely look forward to that. Maybe a good segue into software. In your last earnings call, you talked about MRM. Maybe just review that for people and the significance of that. And then maybe -- and then a question at the end of that is, is that also coming into your conversations on the RFQs at this stage? Or is it kind of too early to go that far?

So the MRM system is a system that we're developing for the BMW -- the new BMW project. It's a minimum risk maneuver system, which means that the car that drives autonomously, if the main ECU, the main computer fails for some reason, the car cannot be out of control. Something needs to bring the car into safety, either stop or make aware the driver that he needs to take control. So there is a separate ECU unit which runs this software. MRM system usually ran or used the camera as the main sensor in order to bring the car to safety in case of an emergency. BMW decided that a preferable setup would be using a LiDAR which is a 3D sensor. It brings some benefits that maybe the camera cannot resolve. And this is part of our offering. So we have the LiDAR and we have a compute box that we are providing as well. On this compute box, we are running our software -- our perception software, and the MRM software. So the offering per unit is much larger. Regarding talks with other OEMs, this is starting to be promoted just now. Actually, the demand or the ask came from BMW. I think that this is very -- a process that you can expect. Once you are Tier 1 and you are a technology player, you can provide more technologies the OEM -- BMW approached us, knowing us very well, knowing that we can provide -- develop for them this piece of technology. And hopefully, we can offer it to other OEMs as well.

And so now you -- you've always done a perception stack, computer vision. Now you've got, I guess, initial slice of the driving policy stack. How far does that go in the future for Innoviz?

So currently, we are trying to be very much focused around the LiDAR. If we can offer more and more services that are benefiting from us being the LiDAR producer, from us being very intimate with the technology, with the output from the technology, we are trying to focus there. We are trying to broaden our offering. Obviously, we can think of going up the stack, providing more and more into the driving decision. I think at this point of time, it's too early. We are trying to focus on the things we do best, the LiDAR, the 3D, the decisions or the understanding of the themes based on the LiDAR. And if there are more opportunities that are related, we probably will do it or think of doing it. So it makes sense. By the way, to this MRM system, one of the other sensor is actually connected. So the radar is connected to the MRM system as well. So to some extent, we are starting to get inputs also from radars.

It gives us a platform, a potential beachhead to expand into other sensors down the road or into other forms of software. Software is a critical part of our longer-term strategy, right? And we know in automotive volumes go up tremendously over time, tech is typically deflationary. But we can grow our software content per vehicle. This hopefully is just the first step in terms of what we're doing. That embeds you even further with the customer. It makes you stickier. It leads to longer-term relationships. And it's a major differentiator because it's a critical input for the safety of the product.

Interesting. So radar could be part of the longer-term road map here.

Has the potential.

Has the potential. Currently, we are just funneling the radar through the system, but there is a good reason why they want to have the radar connected to the MRM system.

Interesting. And maybe one more on software. Talk about data that you're getting from your existing customer right now, both prelaunch? And are you getting data post the vehicle launch that you can kind of use from real world kind of ground truth examples that kind of improve your software offering?

Yes. So the software is a deep learning algorithm, the perception software, that is getting better over time as data pours in. The main thing you need to solve for is edge cases. And edge cases basically are endless. You always have another edge case. You never -- you can never cover everything. You can cover a good portion of it, but never cover everything. So once -- at the beginning of the program or until the SOP, we get data from the OEM, they drive the cars millions of kilometers, gather data. Some of it is offloaded to us. We use it in order to train our system. Once the car is on the road, it caches in some of the information, not all of the information because they don't have enough storage in each car, but it does remember the last occurrences that the car did. So if there is some events, some edge case, an accident, we are getting the data -- the exact data in order to assess if, A, the LiDAR needs to have something to do with it, or the LiDAR gets -- the inform -- or we get the information just to train our systems in order to improve it such that we can identify better this edge case. So the fact that we have programs and that they are launching, it means basically that we have crowdsourcing of edge cases that we will get the tickets from and we can improve our system. So this is something very important also for the future and the upgrades of the systems.

We have our cars on the road, the OEMs have their cars on the road, and now we're going to have customer cars on the road. It's critical because 99.9% of the driving that happens with our test vehicles and OEM test vehicles, it's almost not usable, right? You're just driving, nothing really important happens. It's the 0.01% where the software gets confused or there's almost an accident, those are the things that we plug into our neural networks, which feeds our software. It's an AI tool that we use in conjunction with our human engineers. And it's going to speed up the pace of development. And we think our lead versus our competitors on the software front.

Perfect. Maybe to put it all together, we talked about the sensor itself, InnovizTwo, the MRM compute module, incremental software opportunities. How should investors think about the journey of kind of the ASP, or content per vehicle for the company both in the next kind of 2 or 3 years as you roll out the InnovizTwo and then longer term for modeling purposes?

So definitely, the -- and obviously, the content per vehicle increases as we offer more and more to the customers, another compute box, another software. And if we -- we discussed it in the past, the ASP of the InnovizTwo or anything around $850 to $500, depending on the year and the volume. So basically the ASP is able to go significantly down over time. Now, so let's say, it's $1,000 minus, this is the offering per vehicle. Now that we have additional software, we are getting per vehicle additional dollars. So it's $1,000 plus. And meaning we have -- we are getting more and more per vehicle. We're getting a bigger piece of the pie. And I think we will be able to offer over time more services, more pieces of hardware. And it means that per vehicle, we will be able to generate more revenues.

And driving down the price of the LiDAR itself, the sensor is critical to driving volumes higher. It's going to drive adoption, which drives volumes, which is how you make money in automotive. It also increases the opportunity for us on the software side, right? That's where the content growth can come from. And then if you go beyond that, it also opens the door potentially for a recurring revenue model, right? We -- on the software side, we want to evolve into a recurring revenue model. For that to happen, the OEMs need their own recurring revenue. And we think getting into true Level 3, Level 4 autonomy opens the door for the OEMs. And OEMs have been trying to develop recurring revenue models for years, and some of them have been good, Telematics, et cetera. But if you look at autonomy, the potential to sell a car that's potentially Level 2+, but with a subscription service can be a true Level 3 or Level 4, that opens the door for their recurring revenue, which opens the door for us on the software side longer term.

Absolutely. So it sounds like kind of all-in packet, I know it's early days, over $1,000 is sort of how to think about the potential.

Yes.

Perfect. Wanted to maybe swing kind of into an industry LiDAR adoption question. We often -- we've done this conversation, tend to associate LiDAR with L2+, L3, L4, but it's obviously just a broader safety element just to make vehicles safer, that the regulations historically do tend to follow. I think we've seen some recent development in terms of nighttime detection. What do you see on the regulatory side that maybe gives you more or less sort of optimism longer term about LiDAR being just there as a safety piece of content required maybe one day for many if not all cars?

So first of all, I think the regulation is getting there, not necessarily specifies LiDAR. It specifies, A, you need redundancy, you need -- in order to get from Level 2 to Level 3, you need 3 sensors; in order to have these sensors, these sensors have to be self-aware. What does it mean, self-aware? It means they need to know that they are degradated. The camera can know that it is degradated, LiDAR can know it's degradated, meaning it has some -- let's say it has some mud on the window, it knows, it can't see well, it notifies the system something is wrong, we need to do something about this. Radar is maybe less self-aware. So the regulation gives -- I think, is giving a broad requirements of how do you -- what do you need in order to get to a certain level of safety, a certain level of autonomy. I think it supports the LiDAR story, which is a sensor, different modality -- physical modality than a camera, different than a radar. So I think for us, it's good news. I think also in terms of safety, we heard that there are some regulations that concern stopping the car in the case of emergency. So this also supports installing LiDARs even if it's not Level 3. So all in all, I think with safety in our minds, and the regulator mind, it supports more and more sensors be installed in a car.

If you do have regulations that require lower level ADAS like systems like automated emergency braking, that adds cost to a vehicle, which is obvious, but it also lowers the gap between that kind of an offering and a true LiDAR-based Level 3 or Level 2+ system, right? So if you need to deploy some ADAS technologies and it's $200 to $300 cost per vehicle, it lowers the delta between that and a LiDAR-based system, and it could be a critical catalyst for driving volumes and adoption much faster than what people are expecting today, if it happens.

Right. Absolutely. And then wanted to spend a minute or 2 just on the nonauto side, on the Innoviz360 product. I think it was a couple of -- yes, maybe a year or so ago, you introduced that product. How is that going? Kind of what's the focus on non-auto as well as like how you're seeing the competitive environment develop there?

So first of all, LiDARs are not exclusive to automotive. The only -- we are focused on the automotive because this is what we believe the biggest opportunity out there. There is no other sector that can provide in one contract hundreds of thousands of LiDAR. So we are focused there, with the belief that once you solve the automotive LiDAR problem, call it, or be able to provide to the automotive space, will be able to provide to any other industry that make good LiDARs. And good LiDARs are being sourced or are required in other industries, in many aspects: smart cities, securities, security applications, and so on. Part of it is the strategy that we have the 360. 360 is basically InnovizTwo with a different scanning mechanism. So it sits on the same technology base. And we understand that there is no reason not to leverage from the fact that we have good LiDARs and they are already available to the market not to monetize out of it. So in the automotive space, the cycle time is much longer. And in the nonautomotive space, you can tap into additional revenues in relatively short time span. This is part of what we do. We -- if somebody wants to buy for a non-automotive application, we are happy to sell.

We made a bet, right? I mean I think you saw with early LiDAR companies they had a choice. Do you focus on industrial and go after earlier revenues, higher gross margins now? Or do you go after automotive? And companies went both ways, right? We made a bet that by focusing on automotive, we can secure the volumes, right? We're not talking about thousands of units per year. In automotive, you're talking about the potential for millions of units per year. So there's 2 parts to the bet. There's one, secure the volumes, which is going to lower your cost and allow us to come into the nonautomotive space later with a massive structural cost advantage. The other aspect is it's much easier to downregulate the technology, right? When you're working with BMW and you're working with Volkswagen and you're trying to solve for a car going down the highway at 85 miles an hour, it's a very complex use case with a lot of edge cases. If you can solve for that 2, 3, 4, 5 years from now, we can downregulate that technology and put it on a forklift. It's not going to be that challenging. If you're solving for the forklift today and saying that you're going to solve for an 85-mile an hour vehicle down the road, the chances of that working out, I think, are much slimmer. So that's the bet that we made.

Perfect. And then maybe in the last few minutes, a couple of financial questions. We talked about auto and content per vehicle, we talked about non-auto. Putting it all together, how should investors think about the gross margin profile of the company as you ramp production to InnovizTwo and get some of these other products out there as well.

So there are a few stages in how we, A, finance the company; B, get to profitability. So in the earlier years, this year, next year, the following years, part of our strategy becoming a Tier 1 means that we are tapping into NREs these services that we provide, the automakers, during the development of the platform. And I mean tens of millions of dollars per program. And this helps us to balance off our expenses. In addition to that, you sell samples, and high pricing, thousands of dollars, and again, serves the company to generate more and more revenues. Obviously, our aim and our goal is to get to mass production, to get to mass volumes in order to get the margins. So -- and this -- the strategy here is to aggregate volumes. So we, today, are competing almost exclusively, not almost, exclusively with InnovizTwo, meaning each program that we win means another client to the InnovizTwo, which means the aggregation of volumes. Once you get to these hundreds of thousands of units per year and even more, you can get to very favorable gross margins and get to this tipping point of being a profitable company, breakeven, cash flow positive.

Got it. In terms of other funding options, how are you thinking there, as to just to the capital raise? But most that you're kind of [indiscernible] NREs in the near term?

Yes. So first of all, we have a few options of financing the company on the -- on ongoing basis. Obviously, NREs this is in the short term. Once we ramp up the volumes, we get the revenues from the programs. We are launching 2 programs this year, which are providing additional revenues. So all of this aggregates to what we expect increased revenues for the company. Obviously, once we will see this hockey stick materialize, we can finance the company in a non-dilutive way, working capital and so on. So this is how we see financing for the company. We believe we can -- we have a strong balance sheet. There are the NREs. And we have enough runway to get to the inflection point.

Perfect. And I'll ask one last one, and we're out of time, but I'll make it very quick. If all goes well and you win a lot of awards with InnovizTwo, should investors think about maybe some R&D creep next year or significant amount? Or is it mostly H1 products so you can kind of kind of leverage those?

So Innoviz is quite mature company. It has what it needs. It's already running as a Tier 1 for over a year. We are pitching InnovizTwo, as I said before. So it's the same product for everybody with minimum adjustment per client. So basically, our expectation is to stay stable on the expense side. Yes, if we win additional program, we might need to cater this, add additional headcount, but it's not dramatic. So if you look at our current headcount, we have in Germany something like 11 or 12 people supporting 2 very big clients. So let's say we get another German OEM, I don't believe there will be a very big increase in headcount. So...

The big step-up in our cost happened when we transitioned to a Tier 1. That was for Volkswagen, that was more than a year ago at this point. So from here, it's much smaller. No major step-ups.

Very clear, very helpful. Eldar, Rob, thank you so much. Great conversation. We appreciate you participating in the conference. Thank you very much for joining us well. We'll end it here. Thanks again.

Thank you. Thank you very much.