Innoviz Technologies Ltd.

Thank you for standing by. Our conference will begin shortly.

Good morning. This is Rob Moffitt, Vice President of Corporate Development and Investor Relations at Innoviz, and I want to welcome you to our earnings conference call. Joining us today are Omer Kalev, Chief Executive Officer, and Elder Segla, Chief Financial Officer. Following their opening remarks, we will open the call to your questions. I would like to remind everyone that this call is being recorded and will be available on the Investor Relations section of our website at ir.innoviz.tech. Before we begin, I would like to remind you that our discussion today will include forward-looking statements that are subject to risks and uncertainties relating to future events and the future financial performance of InnoVis. Actual results could differ materially from those anticipated in the forward-looking statements. Forward-looking statements made today speak only to our expectations as of today, and we undertake no obligation to publicly update or revise them. For discussion of some important risk factors that could cause actual results to differ materially from any forward-looking statements, Please see the risk factor section of our form 20F filed with the SEC on March 9th, 2023. I will now turn the call over to Omer. Please go ahead.

Thank you, Rob, and good morning, everyone, and thank you for joining us. 2Q was a really, really big quarter for InnoVis. While the quarter itself was only 90 days long, you could say it was actually five years in the making. This quarter we delivered on what was easily the biggest milestone in our company's history. we shipped our first production units ever. They are on their way to our customer and close partner, BMW, and will be installed in the grill of the beautiful 7 Series, which will include our InnoVis-1 LiDAR sensor and our AI-enabled perception software package. I know many of you saw this already, but BMW Blog did a wonderful article and video summarizing BMW Group's plans for the Level 3 7 Series. If you haven't watched it yet, I encourage you to go to the bmwblog.com or search for it on YouTube. In the video, you can see the 7 Series driving seamlessly through a variety of scenarios and hear more about BMW's plans and timeline for deploying the technology. We've been saying for a while now that the platform was expected to SOP in the back half of 2023. So to get the first production units out in the earlier part of the third quarter feels like a nice win. While there's always going to be some level of uncertainty in the library industry, we are building a track record of credibility that includes delivering on our goals. With the 7 Series SOP now underway, we are turning our attention to additional models and markets. While the 7 Series was planned to be our initial flagship launch for the Innobis One, we've said in the past that our technology was certified on several models and variants within the BMW group. We are pursuing integration with additional models. This includes supporting the recently announced localized development efforts of Level 3 autonomous driving, including within China. While all of the news about the first generation Innobis platform, Innobis One platform is obviously exciting, the bigger news today is around this morning announcement that we are entering an all-new development phase of a second generation ladder platform for the BMW Group that is built around the Innobis 2 sensor and our AI-enabled perception software stack as an opportunity for a second platform with BMW. Our first generation program included selling Innobis One components and perception software to BMW. The lighter components were shipped to Magna, the tier one on that program, and then manufactured into a finished lighter and shipped to BMW. But the second generation platform that we're developing is remarkably different. First of all, we are quoting the program now as a tier one. This means that instead of selling LIDAR components and only getting part of the economics, we could have the opportunity to sell the entire package, including the LIDAR and the AI compute model directly to the customer. And since this new platform is being developed around the much more advanced InnoVis2 LIDAR sensor, and it's planned to run on our newest custom ASIC, which I'll talk more about shortly, It is expected to be able to offer a much more robust perception software package, including several new AI-enabled features. Without going into too much detail, this stronger version of our perception software is designed to do much more than earlier versions. One of the best example of that is the new Minimum Risk Maneuver, or MRM, software that we plan on developing with BMW. And we plan for this more robust software to sit on a dedicated AI compute model offering more compute power than we've ever had before. I will go into this in more details in the coming slides. The BMW Group has been an amazing partner for us. They operate with some of the highest engineering and design standards in the industry. We have set a high bar together with the 7 Series launch, and I believe we are both excited to continue that work on developing a next-generation platform. As part of this process, BMW is making a meaningful investment in terms of time and resources and is embedding a dedicated team of engineers with Innobis so that we can move quickly on this next phase of development. is planned to target a much broader range of models than the first generation program. And with a higher install rates that should result in higher and larger volumes. We believe that ladder technology and level three driving is going to become much more common over the next decade as it trickles down from cutting edge EVs and luxury, ultra luxury models and more into mainstream vehicles. This opportunity for Innobis doesn't exist just at BMW or Volkswagen, but across the entire industry. We can see the level of opportunity that is out there in our RFI and RFQ pipeline. Level three vehicles are increasingly going to become a reality. Next, I want to introduce you to the InnoVis Core, our new AI compute model. This is the dedicated compute model that I mentioned earlier as part of the solution that is being developed for BMW. I want you to think about the Core as a new base for us to grow our suite of AI machine learning powered software. It offers decentralized compute power dedicated entirely to the LiDAR and the LiDAR-based software solutions. By leveraging the core, we can reduce the total compute load of the central ECU. This can reduce the cost of the central ECU to the OEM and free up compute resources for other applications, all while creating a dedicated ECU for the LiDAR and other LiDAR-based solutions. This is a critical component of the BMW solution that we are developing due to the more robust perception software and the entirely new application built off it, such as the new MRM software. Having a second compute resource should allow for more total processing power, which can unlock analysis at higher resolution and increase frame rates. And with that, reduce the overall bandwidth and simplify the system architecture. It also separates the hardware from the central ECU, adding a layer of redundancy that is critical for a level 3 or higher system. the biggest operating system. This can enable a number of functions, including over-the-air updates, data sharing, and high-definition 3D mapping. It also enables connectivity through dedicated ports to other sensors, such as the camera or radar. While we are not integrating camera or radar data into our software offering today, with the InnoVis Core as a platform, it should be possible to begin working on solutions that can integrate this functionality in the future. This gives us a path for ongoing future software product development and extensions. I want to revisit our MRM solution. Last quarter, we introduced the concept of the MRM to you. In this quarter, we announced that we have begun development of the solution with BMW as part of our offering. As a quick reminder, the MLM system is software that is designed to sit on top of the InnoVis Core AI compute model within a vehicle and operate as a backup system. In the event of a complication with the primary system, the MLM could take over control of the vehicle, offering a transition period for the driver to retake control of the vehicle and offer the ability to safely pull the vehicle to the side of the road if the driver does not retake control within a specific timeframe. MRM systems are not new. They have been around for several years, but historically, they have been camera-based. We believe that operating a ladder-based system should offer key structural advantages over camera systems, including a true 3D image, along with reduced risk in low light and extreme sun situation, as well as an environmental consideration like rain or snow. Customers have told us that in order to be able to drive at increasingly faster highway speeds, the MRM system likely needs to be based on a higher resolution 3D sensing modality, replacing the 2D solution being used today. The InnoVis MRM solution is developed to fuse AI, deep learning, and classic computer processing for object detection and classification, landmark and lane marking detection, ego motion estimation, path planning, and maneuvering. The dedicated parallel processing pipeline is expected to ensure fail-safe execution of any needed risk maneuver. Successfully building out this product category would be a natural extension of the success we have already demonstrated in perception software and would help us to move further up the stack, potentially offering additional incremental opportunities down the road. The benefits of having a larger software offering are clear. First, they can build upon and further expand the value that our ladder sensor technology brings to the table. And second, the gross margin profile on the software is much higher than hardware. In the automotive end market, where you have more than 19 million units of volume per year, you can generate meaningful leverage and strong returns on investment capital from a growing software suite. As we're talking about the intersection between hardware and software, I want to talk a little bit about our second generation custom ASIC, which was taped out a few months ago and is now in our labs. An ASIC is an application-specific integrated circuit. At a very high level, it's a custom-built chip designed specifically for one use. And for us, it's a 16-nanometer chip mixed-signal automotive process design that has nearly 50 processors with multiple dedicated proprietary DSPs to process dozens of pixels simultaneously at the speed of light. Our first-generation ASIC was taped out in 2019. That chip has clearly served its purpose, having won multiple series production awards since then and building a substantial RFI and RFQ pipeline for us. This is the chip that is going into SOP because this year, but we believe this second-generation chip is going to take our performance to the next level, leveraging several additional years of learning and R&D. There are two key benefits to this second generation ASIC. The first, that it unlocks much more range, taking our maximum range detection range to over 450 meters, up from 300 meters before. And the second is to have better resolution. The more powerful chip enables us to produce millions of more points. In fact, we were able to nearly double the total number of points that we can process per second. This can power the higher density point cloud with even better resolution than before. Together, this dramatically higher range and resolution translates into more points of data. And that data is what fuels our AI tools, including the neural networks that are a critical part of our software development. Faster neural network training translates into better perception software. Better perception software can unlock new features like enabling vehicles to travel at higher speed or operate in more complex environments. Moving faster than competitors and unlocking features like this can drive higher service production wins rates for our sensors and higher dollar value for our software packages, all while deepening our embedded relationships with customers. While we are on the topic of resolution, I want to take you through a quick trip through our company's history. On the right hand side of the slide, you can see the progression of our point cloud over time. As I mentioned earlier, we didn't have the first generation custom ASIC until 2019. So the point clouds before then were much more limited. And as you can see, in 2020, with the addition of our first custom ASIC, coupled with our improvements in the optics, hardware, and software, we took a big step forward. In 2022, we had another big leap forward with the development of the InnoVis 2. We didn't stop there. The video you see on the screen is one of the latest point clouds from the InnoVis 2 with the first-generation ASIC. You can see it has continued to improve since last year, and it is at a point where we are approaching high-definition camera-like resolution with the added benefits of a true 3D point cloud. I'm confident that once we get this second generation ASIC fully integrated, we'll take another major step forward and broaden the gap between us and the competitors. In terms of our timeline, we expect the first samples with this new chip to ship this fall to our customers, including Volkswagen and BMW, including our recently announced new light commercial vehicle program partner. We are beginning automotive qualification of the part, which should be ready for mass production by mid-late 24. And while we are on the topic, I just want to make one more point. And that's that there are a number of other players out there in the LiDAR industry that are still using FPGAs instead of custom ACID. While it's true that you can build a LiDAR with FPGAs, at Innoviz we strongly believe that you won't be able to scale a LiDAR company that way. If we try to produce the Innoviz 2 with FPGAs, it would take somewhere between four to six very expensive FPGAs, adding hundreds of dollars of cost to the finished product. It would be heavier, take more space, produce more heat and consume much more power. And while not really matching our current performance, since only a custom ASIC allows for the power processing of dozens of channels of data, which is the salt of the higher pixel rate and improved resolution. I believe that our first generation custom ASIC has been a major source of the competitive advantage that we have demonstrated thus far. And this next chip is going to take our lead to the next level. One of the other benefits of our custom ASIC that I wanted to touch on is the ability to support multiple LIDAR configuration with just one architecture. We designed our chip to have the flexibility to be able to offer a wide area of configurations across all different types of ranges, resolutions, field of views, and frame rates. We can tweak the product with only minor changes that can mostly be done through the software. As a result, we can create end-to-end solution for a variety of use cases from long haul commercial track that need ultra long range for highway speeds all the way down to wide field of view, static smart city applications. With this benefit, we can go to the market with one architecture built around our already cost-advantaged 905-nanometer-based single laser, single detector solution and fulfill almost any use case across automotive, trucking, robot taxis, buses, shuttles, smart cities, and more. All with the same optics, the same hardware, and the same ASIC. This should enable us to run much higher volumes through the same components, rapidly lowering our cost per unit. driving attractive unit cost economics and delivering potentially a meaningful structural cost advantage. Next, I want to give a quick update on the new light commercial vehicle program that we announced last quarter. As you recall, we said that this program where we are moving on a faster than normal timeline due to the fact that we're displacing a development stage competitor mid-project. I'm happy to share that we've already begun shipping sample units to the customer with the first units being sent towards the end of the second quarter with additional shipments expected to run through the back half of the year. With those units now in the customer hands, We expect that you will be able to see the new test vehicles on the roads in the coming months with our ladder, offering additional tangible evidence of Innove's ongoing commercial momentum. Moving on to our pipeline, even with the two of the platform moving from our pipeline to our awards and programs category, year to date, we're still solidly in the middle of the 10 to 15 range of programs in an official RFQ process. and we still have roughly half in the more advanced RFQ stage. And while the first half of the year has been defined by success, primarily with existing customers, we believe that the back half of the year and the fourth quarter in particular could be defined by new customers. One interesting insight that I have to offer from some of our RFQ work this quarter is that one customer has indicated they are only considering 905 nanometer solutions and another has decided to shortlist his 905 solutions to only 905 solutions. For both, this follows their previous experience with other solutions and the challenges they face with power consumption and resolution. As we disclosed in the past, we are working on more than five active RFQs in parallel, which is the highest level of activity in the company's history. There appears to be a very robust level of activity, all targeting the 25 to 2028 OEMs want to compete on the technology. And we believe the next big differentiation factor is going to be level three driving, exactly like what BMW is bringing to the roads now. And between the programs we've already announced, to 15 in the rfi and rfq pipeline we've either already won business or actively quoting new awards with eight out of the top 10 global automakers we ultimately believe that this is likely going to be a winner takes most kind of market the technology is safety critical there are very high levels of tech differentiations and the player that wins the most business is ultimately going to have a scale and cost leadership advantage that is likely going to be difficult to match Given the fact that most of these programs will be on the road for 8 to 10 years, we believe that a major portion of the industry market share is going to be determined in the next 12 to 18 months. And with that, I will turn the call over to Elda.