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

My name is Winnie Dong, and I work in the U.S. auto and technology [indiscernible]. Innoviz is a manufacturer of high-performance solid-state LiDAR sensor and perception. The company is looking to enable the mass production of autonomy -- autonomous vehicles and has won large production contracts with companies like BMW, Volkswagen, among others. I'm very pleased to be joined by Omer Keilaf, CEO, for a presentation and Q&A. But before we dive into the Q&A, I think, Omer, you have a presentation that you'd like to share with the audience.

Yes. Okay. So hi, everyone. I'm Omer. Happy to be here. So I'll start by saying Innoviz is acting as a Tier 1 direct supplier to the automotive space. I'll touch some of the benefits that comes from there. But that's kind of our, I would say, a unique proposition and advantage in this market. I'm jumping into the product. I think when talking about the product without showing it, it's kind of meaningless. LiDARs have gone through a lot of progress in the last few years. This is our second generation that we are currently offering to the market. You can see, I would say the very high details that you can get from our sensor. This is coming from a single sensor. We are using a 905-nanometer. You can see here the recent version running at 20 frames per second, very high details, 0.05 degree solution, 250 meters. This is coming from a single 905-nanometer laser. I mean, 5 years ago, people would say 905 can never get to this performance. And the benefit of using 905 is because of the low cost. The cost of 905 versus the solutions based on 1550 is significantly lower. Having the ability to unlock the performance of 905 is obviously kind of the holy grail. This is allowing us to meet all of the requirements we get in the market at a price point that no one can match. Here is another video for sensor running at 20 frames per second. You can see the details. There's no other LiDAR that meets this performance. Now when we, Innoviz, is offering our solution, we are not only offering the LiDAR, we also offer the perception software. That's the software that analyzes the point cloud and translate it into object detection and classification. The Level 3 system that eventually is fed from the different sensors, the LiDAR, the camera and the RADAR, eventually is doing the sensor fusion and driving the vehicle. And if this system fails, then you need to have redundancy also for the driving platform. So in a program that we're currently closing for, we are also offering an MRM system. MRM system is used as a secondary compute platform that will drive the vehicle but only for safety. So you don't want to duplicate the entire platform to keep the Level 3 active. But you do need to have redundancy in the case this platform fails. You want to have a backup that drives -- that still drives the vehicle. In Level 3, the driver is not looking at the road, is not holding the wheel. You need something to control the vehicle, and this is the MRM system. It is simplified because it's only using the LiDAR data. And in this program, we are offering the LiDAR. We're also offering the compute platform, our perception software and the MRM software. Those -- that's an additional, that's an incremental offering that we are now doing. In general, we have, so far, 5 production wins. I'll touch the details of them, where 3 of them are for Level 3, 2 of them are for Level 4. In 4 of these cases, we're acting as a direct supplier of Tier 1, meaning that we are selling the LiDARs ourselves to the customer. This is different than -- our first win in the market was to BMW as a customer, but we were a Tier 2. We were working with a Tier 1, Magna. Magna is producing today LiDARs for BMW using our components and our production tools, but we're acting as a Tier 2. And from that point on, all the programs that we engage and won was as a direct supplier, and I'll touch some of the benefits. And of course, from 5 of those -- those 5 programs that we already awarded, there is an extensive order book, $6.9 billion. And obviously, with a car company like Volkswagen, you can imagine it's a huge win for us. In our pipeline, other than those 5 programs, we have 10 to 15 additional platforms that we're competing on. Eight, I would say that Innoviz either already awarded or in the process of an RFI or an RFQ with 8 of the top 10 OEMs in the market, which is, I think, a huge, I would say, validation that the market is consolidating, is in a process to choose a LiDAR. And we could not have been in a better position with the fact that our LiDAR that we are offering now to any customer with our second generation with a validation coming from several customers that already chose this product, and that's really a big benefit they get in terms of lowering their risk. Another point that we started communicating to the market is related to the NREs. So people would ask, okay, how are you going to fund the company? There are several years until the, I would say, the ramp-up of the different customers. As a direct supplier of Tier 1, we are entitled to get NREs for our programs. This can vary between $20 million and even $60 million. There is a program right now that we are waiting for a decision for, which is actually at the top of that range. And that's obviously a very meaningful contribution to our ability to fund our activities. This is only for one program. We are competing on almost 15 programs. Each one will provide us NREs. And we have a cash equivalent of almost $160 million. That's roughly 15 months of work. We are expecting to collect this year $30 million mostly from NREs. That will give us another quarter. This $60 million gives us another runway. You understand that every win we bring extends our runway and, of course, support our work. In terms of the different programs, so we have the BMW, the one I mentioned at the beginning. This is with InnovizOne, our first generation working with Magna. We're selling the LiDAR and the perception software. BMW is going to launch this year with their Level 3 based on our LiDAR and our perception software. So you can imagine not only that our LiDAR needs to be validated and tested for start-up production, also our software. So we are driving, and BMW is driving and Magna is driving hundreds of thousands of kilometers and testing the software. And one of our -- really, one of the many advantages that we have is that our software is already validated for this automotive grade, and we are offering it now also to other customers. Our second program is the shuttle, which is also launching this year using multiple of our LiDAR. This is InnovizOne. Actually, we are in discussion with them to go to the InnovizTwo at a later time. Volkswagen, our biggest award so far, we announced it last year. Volkswagen is the second largest OEM in the market. It was the biggest RFQ we've seen for a very long time. Now we see more, I would say, RFQs that are that size. But that was a very unique, I would say, opportunity for us. And generally, Volkswagen is a, I would say, technology leader. Many carmakers see the validation we got, becoming a direct supplier of Tier 1. So a company like Volkswagen is a big, I would say, big validation opportunity for us that opened many opportunities in the automotive market. Just 3 months later, we got another award from an Asian OEM that saw the decision made by Volkswagen and generally understood that, okay, this is a great opportunity for them to piggyback on the decision that Volkswagen made knowing that they will work with us closely to bring the product to the market. And now just recently, we announced another program, which is a Level 4 platform using multiples of LiDARs. This is coming from an existing customer. It's, I would say, a validation of the spaces that once you're working with a customer, it opens new opportunities. And again, there are 10 to 15 programs that we are working on. I would say some of them are relatively going to be decided soon. There are many programs that were in RFI stage. It converted to an RFQ, and maybe I'll spend a few minutes on -- a second maybe on that. Generally, on the automotive process, it's done by RFIs and RFQs. RFI is a process where the customer is asking a lot of questions, it could be 10 suppliers to understand what they can offer. But eventually, that process can take a long time. Because unless there is a platform that they need to make a decision for, they don't need to move to the RFQ stage. An RFQ stage is where they limit the number of suppliers they eventually negotiate with. But at the end of the process, it could be 6 months or 9 months. Eventually, there is a decision for a specific time when they launch with specific compute platform and integration and window curve and cleaning system and everything you can imagine for a product that eventually need to be mass produced. And many RFIs conducted RFQ in the last quarter. We've never had so many RFQs dealing in parallel. We have 10 to 15 programs. It's actually the far end of that. And we said that more than half of that is in an RFQ stage. That's a -- it's a really -- first, we had so many RFQs to deal with in parallel. Yes. So I think that we talked about the different opportunities, the 10 to 15 programs, the 8 to -- or the top 10 OEMs, I think we can skip on that. This slide talks about the momentum. So we started in 2016. We're not the oldest LiDAR company in the market, but I think we're moving the fastest. We see -- 2 years after we just started, we got our BMW award. And then you see that it took a lot of time for the market to converge. But now you can see in the last year, we got 3 programs. And we're still seeing many opportunities for us to grow this year with additional decisions that are going to be made. In terms of our technology pipeline, so we had InnovizOne, our first generation. This is SOP. This year, we've been doubling the shuttle program. InnovizTwo, I showed you the product. It really is amazing. It's like 30x improvement compared to the first product and 70% cost reduced. Amazing step. Innoviz360, it's our, I would say, our development for the future when it comes to Level 4. When thinking about where the market is going, we knew Level 3 is going to be the fastest growth. And we really need to bring the best LiDAR for that kind of market. But in the future, we do expect to see Level 4. And from that perspective, we think that a wider field of view is needed. Innoviz360, it's a very low cost with very high performance that there's no 360 sensor that is capable of, and we see that as a good opportunity for us going forward. There is also an InnovizFour, you can say, that is in development for a later time. So we talked about the cost reduction and performance improvements, really a huge jump between the different products. You can see here kind of a summary of the last 7 years. You see the 5 steps of the front load in LiDAR. There's also the 360 somewhere below this. You can see myself 7 years ago with the LiDAR. It's a table top setup, which we kind of like wanted to test the technology. You see one year later, we packed everything so we can carry it on a car, and you see the performance has improved. And then we actually industrialized the product so we can start selling. We called it InnovizPro. We called it InnovizPro, it was Innoviz Prototype, and then we shortened it to InnovizPro. And you see the performance has improved. Then InnovizOne, again, users improvement in performance and cost reduction. And then InnovizTwo, which I think is an amazing step. In a way, I really think this is a good fit for Level 3, very high speed and unlocking not just the premium vehicles. Talking about the motivation of becoming a Tier 1. So we started as a Tier 2. We were a very young company. BMW would never consider us as a direct supplier. Thinking about the Tier 1s in this space, Bosch, Continental, Magna and Aptiv, like we were a very young company. We could not approach BMW as a Tier 1. So you can see the difficulty when working as a Tier 2. Like you have the Tier 1. And Tier 2, we need to split the profit. There is the program cost of its size. Then moving to the InnovizTwo, and we saw this as an opportunity to reset the situation, let's see if we can become a Tier 1 ourself. We started that dialogue with Audi. They saw the motivation, right? I mean you -- once you do the vertical integration, you can cut a lot of cost from your offering. Of course, with the bill of material cuts of the 70% adds to a very attractive price point while keeping a very healthy gross margin for the company. So of course, Audi, that was doing the sourcing decision for Volkswagen, they had to feel comfortable with us becoming a Tier 1. It was not a simple task. I mean, eventually, when making a decision of such a unique component with a single source for a very high volume where the industrialization and quality are big topics for these customers, the production, the shortage of components, et cetera, they need to be very certain that we can act as a Tier 1. But we were able to convince them that we can. Because even in the BMW program, we acted as a Tier 2, but we took many of the roles of the Tier 1. We were responsible for the production and development or the production tools and the software and the supplier management. And in a way, by showing them everything that we've done also for InnovizOne, we were able to convince them that they feel comfortable with us acting as a Tier 1. And you can see obviously the benefit on that. We are bidding now today to any customer as a Tier 1 without the dependency on a Tier 2 -- a Tier 1 that might want to act on the program and might actually want to benefit from it significantly higher than us. So that's maybe the biggest step we've done in the last year. Our software is related to the computer vision. So you're seeing here InnovizOne, the product that is launching this year with the perception software, object detection and classification. We're collecting a lot of data and of course, training our algorithms. And those trainings are going to be carried -- are carried over for InnovizTwo. Our program with Volkswagen is with InnovizTwo. Our software -- our perception software is embedded on their Qualcomm platform for their Level 3. And of course, we benefit from having a lot of validation done by BMW and Magna. The MRM, which I talked about earlier. I think it's a very interesting situation where this specific customer decided that in their next platform, the MRM will be based on the LiDAR and not the camera. It used to be the camera. The MRM system will be based on camera. This is something that mobilize offering. They decided they want the MRM to be based on the LiDAR because they see the LiDAR as a primary sensor, much more resilient, seems much better, and that's respectable for weather conditions and ambient light, et cetera. Also because of their desire to have very high speed of Level 3, relying on the camera as an MRM system at that level reflects too much risk. And therefore, they asked us to be part of the development of the MRM system based on the LiDAR. Okay. So we still have time?

You have.

Okay. I think this is important because usually, people are asking us about the different customers, how are you doing with this customer or that customer. And of course, eventually, the OEM, who is the one who makes the source of decision, okay? No one else. They need to see the light of their platform, their pricing, the quality, et cetera. But eventually, each one of those customers is using 1 of 3, these are the major players, of the compute platform. And we need to integrate with them. Now of course, once you are already working with one of those platforms, and in this case, we already have 2. We have design wins with 2 of these 3 platforms. It's a huge benefit because now, when we're competing on a program with the customer, we always try to understand which platform are you using. And we allow -- we show them that there -- the risk reduction they get from working with us because we were already integrated, there is obviously the cost involved in doing additional and not unnecessary work is obviously in favor. So working with these platforms is super important and gives us advantage. We are working on multiple -- I mean in our 10 to 15 programs that we're competing on, obviously, these are the players that are relevant for us. And the fact that we have design wins with 2 of those 3 are important, and we expect to have a design win with the third as well. Now we talk about automotive, and this is really our focus because that's where the growth would come. There is no other markets beyond automotive right now that I can work on a deal and get a contract for millions of LiDAR. There is nothing like that. Of course, there are many applications. But because -- and that's why we focus on it. We invested a lot of effort and time in order to bring the best-in-class [ on-looking ] LiDAR for Level 3 knowing that that's going to be the growth. But leveraging on the existing production line and capacity, of course, we opened the opportunity to sell to other applications. We just yesterday announced on another contract with an Israeli company that is deploying our LiDARs on highways to avoid collisions into bridges. It's that -- it's really the tip of the fork in terms of like the different applications that the 3D sensor can be utilized. So that exposes us to additional revenues that I think would be interesting in the coming years. So in summary, I think that the market is really consolidating. There's many decisions that we are expecting to see in the next year. And I think we've never been in the best setup in order to continue our momentum. We just had several wins. Any customer who needs to -- who wants to choose their solution for the coming years would not want to take a risk and take a product that no one else took. It's a huge risk. And if you look at kind of like which products meets their performance requirements in terms of resolution, range, field of view, the pricing of the product, the perception software that we can offer, the automotive requirements that we can already show our experience and the Tier 1 proposition, they will only source it with Tier 1. They will not risk a very meaningful component through a technology company. And I think from that perspective, we've never -- I mean, we are really in the best setup, and we expect to continue our momentum. I'll pause here for questions.

All right. Thank you so much for all of that.

I was stopping.

No worries. I guess -- so thanks to all of that information. But taking a step back and just looking at the industry, when you see auto ADAS and autonomous driving market, what do you see that inflecting in terms of LiDAR penetration? Do you think it will only be used in Level 3 and above applications? And then which end market you actually see the technology penetrating perhaps faster than the other one?

Of course, I mean, I can refer to those programs that were either won or in a process to win, right? You see we have right now 3 programs for Level 3 and 2 for Level 4. I would say the mass majority of those programs that we're competing are for Level 3. Some of them want to launch as a Level 2+. They want to benefit from the fact that you have a better sensor in order to improve the safety of Level 2 and gain confidence until they switch the software and get to Level 3. I think that every price point unlocks a certain penetration rate, okay? I think with the price point that we are at, which is about $500, around that, we're pricing at $850 to $500, depends on volume and time. And that unlocks a quite meaningful, I would say, market. I can tell you that I had a discussion with one of the largest OEMs in the world telling me that the price point for the [ LiDAR for it, ] I would say, a standard for all vehicles is actually not that far from that. So I think that when you think about the motivation of carmakers, why do they want to add a LiDAR. Of course, it's safety. But it also unlocks their opportunity to monetize. Once they have that Level 3 capabilities, they can start charging for the services and they can benefit from it. And of course, adding a LiDAR at $500 is meaningless in that term. When you think about FSD, it cost $15,000, then, of course, adding LiDAR at $500 is definitely not the friction.

So with so many competing LiDAR makers in the market now, how do you see the competitive landscape evolving and materializing over the course of the decade? Do you think you're consolidating to a few players across end markets or have numerous players that have perhaps more specialized approach with more -- based on more optimal use cases?

I think it's already consolidated. I think the chances that an OEM now will choose a LiDAR that hasn't been picked before is 0, okay? And I think that when we're looking at programs that we're competing on, where customers, our customers, eventually, they see the automotive requirements, not only from the LiDAR kind of performance and price. Eventually, they are very aware on the hurdles of industrialization when producing millions of LiDARs. And they want to pick a partner they can rely on. I mean, they are making a bet, in a way. Like they want to launch in 3 years, they want to know that you're capable of launching in scale, in the quality they need for Level 3 as a passenger vehicle. It's 15 years, right, of quality that you need to maintain. So they choose their partners not just by the, I would say, the technology. I think we are currently ahead of the curve in terms of the performance, and we'll continue to push that. But the fact that we were only selected by Volkswagen and BMW and others is the best that will flywheel effect that we want to have. So I think right now, when we are competing on a -- in a program that we're competing right now with the MRM and everything, you can imagine that taking the role, providing the LiDAR, providing the compute platform, the MRM, they cannot rely on a company that they don't see that can continue to grow and succeed. They want to be -- and basically, we're competing versus other Tier 1s. So the competition, when you think about the number of Tier 1s today that offer a LiDAR at the right performance and price point, there aren't many.

So with the start of production for BMW and the Shuttle slated for this year, which you sort of talked about in your slides, what remains to be done ahead of the production timeline? And then what is the expected ASP and volume perhaps for these programs? Is there any opportunity to take the brand across more programs with these existing customers?

Yes. I'll start with the second, definitely. I already said, I mean, we're in the initial program, where all the discussion about moving to the InnovizTwo, which is great. Generally, we are just doing a lot of testing. I mean, not us. I mean the customer is doing a lot of testing, mostly on the software side. I mean they are driving many vehicles around the world and testing the perception software. And if there is an issue, they open a ticket. We fight with them a little bit. But generally, that's kind of like the effort that is left to do.

Okay. You alluded to the difference between InnovizTwo versus One and in the future products as well. Can you perhaps go into a bit more technical details about sort of like the exact improvements that we're seeing? What's driving those...

It’s my favorite question.

The times of improvement, how does this compare to other sensors currently on the market?

Yes. I think a good way to kind of explain. How do I play with this? Sorry. One second. Okay. We talked about -- it's not playing. Okay. Never mind. It does. Okay. So InnovizOne, this is what we're launching this year. One of the things you can actually see, there are 4 lasers here. You can actually see the stitches. Like you have 1, 2, 3, 4 lasers. We use 4 lasers because our scanner capture 30 degrees by 30 degrees. And we had to multiply the field of view. So we use 4 lasers pointing in different directions. InnovizTwo, I think you can see here. Okay. You can see here. There is no stitches, okay? Generally, we reduced the number of components by 75% just by reducing 3 lasers and 3 detectors and their corresponding optics aligned to it. And obviously, the assembly is much simpler. You can see we've done so because we were able to capture the entire scene with a new scanner. So we designed a new scanner that captures the entire scene. And that's allowing us to use a single laser and a single detector, 905-nanometer, single laser, single detector. Reminding you that LiDARs were using tens of lasers and hundreds of lasers using a single one. And you can see the performance, right? Now of course, we were able to improve the performance by more than 30x while reducing the number of components. That takes a huge step in the technology. So eventually, the LiDAR performance is determined with the signal-to-noise ratio that the system has. The LiDAR is like a communication system. You transmit and you receive, right? And your performance is determined by how much light you're emitting and how much light you're receiving, and how much of the light that you're receiving you were able to actually identify, okay? So in order to improve the performance of the sensor, we improved several factors of it. Of course, you have also the noise figure of the system. So we improved the laser efficiency. We improved the aperture of the system that eventually determines how much light we're collecting. We improved the detector efficiency by several times that improved our ability to convert every photon to an electron. And on the noise level, we improved the system resiliency to noise. So those knobs that we're playing with and we are developing these different components in our LiDAR, that's what allowed us to move. And you can see here how many orders of magnitude that we were able to improve year-over-year. I mean you cannot improve 40x or 20x a year by just clicking on a certain technology multiple times. You cannot, right? You do that by having multiple contributors that you are able to improve twice, 3x, 5x every year. And our team is very multidisciplinary and is doing that by improving different elements in the LiDAR.

Awesome. So as you scale, do you think you'll continue to lean on contract manufacturing partners? Or will you start to build out some [indiscernible] capabilities in-house? And then on that note, can you also remind us what your current capacity is for manufacturing?

Sure. So I think the benefit of working with a contract manufacturer is that you're not tied to a certain geography. And there are customers in different areas. And your ability to scale your manufacturing is faster. Also, eventually, the production line, we do design. So we design the production lines. It's eventually a set of machines that produce the LiDAR. There is no one that I need to train in order to produce my LiDAR. And once my line is active, I can duplicate it and place it in different locations. It takes a month to train the operators on how to operate the LiDAR. Even in InnovizOne, we design the production. All of the automation of the production was done by us. We copied that production that we have in Germany that we designed, we copied it to Michigan Holly Magna facility. And in a month, we show them how to operate. And now they're producing LiDARs with our technology. I think the investment of building a factory in a single geography is wrong. Its huge investment, doesn't make sense, especially when the prices are still allowing us. I mean, the bid of material of the LiDAR is very low because of the [indiscernible]. Our ability to add the contract manufacturer part here is very small. So our ability to leverage on that without having the investment is really great long term. And like we did the vertical integration to become a Tier 1. Whether we do a vertical integration to do the assembly, possibly. I think it's not the right time. I mean, obviously, we will have also additional LiDARs and additional variants and different customers, and then it makes sense.

You touched around [ OEM ]. So I guess -- and also the $500 average like pricing. What is your ultimate pricing target per unit? And how much do you need to flex on [ OEM ] to get this to happen? How do you plan to get there?

I mean we are there. I mean we are using a very simple design. We're using a single laser and a single detector silicon. There is no rocket science behind that. I mean the rocket science is in design of those chips, but the processes in production are simple. That's the -- I would say that's the benefit when you use standard processes. So generally, we are targeting $850 to $500 over volume and time to maintain our 30% gross margin. And we are taking into account warranty and liability and the contract manufacturer.

Okay. And how should we think about, I guess, your cash burn in 2Q?

So actually, I touched it earlier, right? We have about $160 million beginning of the last quarter. We're expecting to collect $30 million this year from different NREs that we're collecting from customers. There is one program that I would say we are in -- I would say, we're close to a decision. And we're -- I believe that we have in a good position there that will tap additional -- between $50 million and $60 million of NRE. And that obviously extends our runway. And there are other programs that we're competing on. And we have sales. I mean every customer we win also generate sales to us, right? I mean we sell samples that are priced at high thousands of and several thousands of dollars and with more for piece price. So generally, we're in a good position.

Awesome. One of your largest wins have been the $4 billion LiDAR business win for, obviously, Volkswagen CARIAD. How is your development work there tracking? And has there been any notable shifts given what has happened?

Yes. Actually, I think that we try to address those kind of questions. We had a press release with CARIAD saying things are going well and we are on track. Other than that, we said that we are talking with Volkswagen on additional platforms that will be using the LiDAR. Volkswagen is a big company. They have different brands, different platforms. They have commercial vehicles. They have different, I would say, opportunities for us. And whatever platform who use Level 3 -- sorry, LiDAR for Level 3 or Level 4 will use our LiDAR. And from that perspective, I would say that the potential is even bigger than the figure you mentioned. I mean think about it. They are selling 10 million of cars a year, okay? That's a huge number. And the adoption rate of LiDARs would grow even at kind of like the 4 billion figure expected from 10 years of development of, I would say, deployment, I would say it could become also a very conservative number because talking about 10 years of a company that produces 10 million cars, I would say there is much more potential there.

Before I move on to potentially my last question, any questions from the audience? So you're now in discussion with NVIDIA about integration onto their Hyperion platform. Can you tell us about how that differs from Luminar's agreement with the Hyperion Drive platform? And what competitive advantage is that -- will be able to supply versus that?

Yes. Without referring to Luminar, I would say, eventually, carmakers that are using Hyperion or planning to use Hyperion eventually have the ability to choose which LiDAR they want to use. And in those RFQs that we're dealing with where NVIDIA is in the platform, working with the customer and with NVIDIA in order to fulfill their requirements, we have the, I would say, NVIDIA are supporting the process. That's an opportunity for us.

Okay.

And I would say that, obviously, the performance that we're -- I mean with the resolution and the range, the power consumption, the size of the product, I think that we have a very good team.

So with that, I think we're up on time. Thank you so much, Omer, for joining us.

Thank you.

Thank you, audience.