Luminar Technologies, Inc. (LAZRQ) Earnings Call Transcript & Summary

Before we begin, please be reminded that during the presentation, we may refer to GAAP and non-GAAP measures. Today's discussion also contains forward-looking statements based on the environment as we currently see it, and as such, does include risks and uncertainties. Please refer to our press release and business update presentation for more information on the specific risk factors that could cause actual results to differ materially as well as a reconciliation of GAAP to non-GAAP measures. [Presentation]

Please welcome Chief Legal Officer, Al Prescott.

Welcome. Welcome, everybody. Good afternoon. I'm Al Prescott. And it's my pleasure to introduce you to Luminar's first-ever Luminar Day. I think it's going to be a great opportunity today for you to hear about our business, our great current and upcoming near-term products and our deep, deep leadership bench. We've got a really terrific speaker lineup that's going to be coming up from both inside the company and some of the powerhouses of auto tech. We have an agenda so long, they all don't fit there. There'll be additional content on manufacturing and industrialization. And after we get through this, after the close of market, we'll be releasing our earnings results. And then Tom Fennimore is going to take the stage here and he's going to go over the financials. For those of you who are still here after about 5:00, we'll be taking live demonstrations at our long-range testing facility. When I say demonstrations, you don't have to be in front of the car, you can be in the car. I think you'll have a lot of fun with it and really see some of the real dramatic improvement we can make. For those who aren't here in person, online, we'll be posting the webcast and the presentation after the close of the event. I hope -- after today, I hope you'll be as excited about Luminar as I am. I've been with Luminar about 2 years. I was formerly the General Counsel at Tesla. And I came here because I see in Tesla a lot -- I see in Luminar, a lot of what I saw in Tesla, which is a young company that hadn't made it to scale with some great product that's going to do great things and make a lot of money along the way. And here at Luminar, I hope you're going to find what I do, which is we have a tremendous opportunity to make really, really dramatic improvements in safety for people today of real cars that people are driving in today and not robotaxis in the future. And along the way, we can be wildly financial successful. So with that, I would now like to introduce you to Dieter Zetsche, the Chairman of Mercedes-Benz and -- Former Chairman of Mercedes-Benz and a Luminar Advisory Board member.

My name is Dieter Zetsche. I'm the former Chairman of the Management Board of Mercedes-Benz. Looking at the development of the automotive industry in the last 5 to 10 years, probably has seen the most change ever. This is driven by 3 major trends: one is electrification; the second one is autonomous driving; and the third one is digitalization. Luminar definitely is in a very special spot in this regard. Many suppliers thought that, talking about autonomous driving, it's all about robotaxis. But in reality, for the industry is much more important to go for other systems, that means assistance systems in the different levels from 2 to 5. And Luminar has an absolutely leading position. That starts with the hardware, where the lidar of Luminar is absolutely second to none. It goes on with the software where Luminar has understood to embed the lidar into an overall software system, which basically can provide for a full autonomous driving system or an assistance system towards autonomous driving by itself. Beyond that, Luminar understood early on that it makes a lot of sense to go for the development together with OEMs. And Luminar is almost the only supplier who does exactly that. Taking all that together, I think Luminar has a fantastic opportunity to support the automotive industry in following this transformation successfully and being a very important partner to the major is. After more than 4 decades in the automotive industry, we, especially at Mercedes, have always dreamed of a world where there are no accidents anymore or at least no harm to the passengers. This was a dream. Now it's about come reality based on this fantastic technology which is now becoming available, not the least, based on what Luminar has developed, this very powerful lidar and all the algorithms, which allow us to really detect the danger before it occurs and ultimately, save the life of the passenger. This is, I think, the most important development beyond electrification of the automotive industry because it really makes the car absolutely positive thing all around.

Please welcome Founder and CEO of Luminar, Austin Russell.

All right. Well, thank you, Dieter. What a legend. That's awesome. Well, I appreciate everyone coming out here, and this is so great to see, a full crowd all around. So wow, great to have you guys here. Thanks again for making the trip. And happy to be hosting this at our new grand opening for headquarter building in Orlando, Florida. So here we are. Like we said, we've got a packed agenda today, but start to be able to go through this and want to be able to give a little bit of an overview in setting the stage as we dive a little bit deeper into today in the presentation. So kicking it off, I want to be able to just take a step back and just as a reminder of the holistic problem that we're solving here and what we're going for. And I think it's really just shocking to see, to be able to take in how big of a holistic problem the -- vehicle accidents are, generally causing over 1.3 million fatalities annually. And if you actually add that up statistically, even with all of us here, it's about, on average, 1 in 100 of us will tragically lose our life in an accident. So statistically, at least 1 person in this room will die as a result of a vehicle fatality. Very real problem. When it comes to accidents, fortunately, most of them are not fatal. But it doesn't change that it can be incredibly negatively impactful, of course, with the average person even in the U.S., for example, experiencing 4 collisions as a driver over the course of their life. So that causes ultimately over 50 million injuries globally from vehicle accidents. So that probably adds up to explain why $1 trillion will be spent per year to auto insurance companies by 2027. It's sort of mind-blowing to think about it, I think it's actually on the order of around $750 billion today. And we're going to talk a little bit more about the implications of insurance and what Luminar's plans are holistically with some news from an insurance standpoint. But I think this is 1 aspect of the holistic journey that's very underappreciated and just how impactful this is. I mean, it's literally like what, over 1% of like global GDP is spent because our cars aren't able to prevent us from getting into accidents. So that's the -- these are the second order opportunities and effects of what can happen when you substantially change vehicle safety. So you would think that despite all of these advancements in assisted driving systems and new technologies on vehicles that we're seeing over the course of the past decade, that there would be a dramatic reduction in the vehicle fatalities holistically as well as per capita. But the challenge is, is that what we've seen in terms of the real-world data is that, that actually hasn't happened at all. In fact, not only has it not gone down, it's not even stayed flat. It's actually gone up, of all things, which is extremely counterintuitive and almost shocking to be able to say the least. And I think when it comes down to it, what's clear is that there are 2 things that are happening. One is assisted driving systems that are out there today are able to have an improved impact for -- first, from a safety perspective. But it's not a panacea that solves vehicle accidents altogether. It only solves a small fraction and a portion of holistic vehicle accidents. The second part of it is that, I think, as time has progressed, people are continuing to become more and more distracted. Maybe there's an argument people are just driving worse and worse. Not totally clear but I think the data speaks for itself, that this is not a problem that's going to be solved on its own, and it's headed in the wrong direction. And by the way, even in the past couple of years, same thing. No precipitous drop and continued this trend. So vehicles equipped with these current camera, radar-based assisted driving systems, if you take like the specific, for example, pedestrian scenario, they don't end up working the vast majority of times in the scenarios where you end up needing them most. And I think this is important to know because it's also not just about the people that are in the vehicles that are affected by these things, but also -- as it's called, the industry vulnerable road users or pedestrians and bicyclists and other folks that can tragically lose their lives from vehicle accidents as well. The other hard part is that where the biggest challenges is generally for higher-speed scenarios. But the higher-speed scenarios are usually the ones that end up causing the fatalities in the first place. So this is a problem that needs to be solved, and we need a step change in technology to make this happen. So if you take a look at it, what I'm saying here is nothing fundamentally new around autonomous vehicles holistically and the efforts for the industry and what we want -- what everybody wanted to make happen from the beginning. Maybe a little bit more color on the significance and severity in the trends over this, over the course of the past decade that I think is surprising. But at the most fundamental level, you take a look at where the autonomous vehicle industry was, and the question is sort of, what happened? There was all this grand vision, if you take a look at the original presentations from companies that we're trying to be able to create driverless vehicles, what, 5, 10 years ago, that the whole point was to be able to save lives. It all comes back to saving lives. But the challenge is, is that you replacing drivers is very, very hard. As Al and Dieter and other folks have mentioned there, our whole vision from the beginning was always about enhancing the driver, not replacing the driver. And that was the big bet, because replacing a driver is very, very hard. You have to cover exponentially orders of magnitude more edge cases that can be extremely challenging even with the best lidar and even with Luminar lidar in the vehicle. That does not make it a solved problem. It's still very, very hard even with those capabilities that make it theoretically possible. The thing is, is that our whole goal is realizing this and taking it out of R&D and bringing it into production vehicles. And that's where you see the transition from cars that look like that, to cars that look like this, into consumers' hands. Again, we're all about enhancing the driver, not replacing the driver. That's the most important part. And with tens of millions of consumer vehicles shipped annually, representing trillions of dollars per year in sales, this is an existing, established market that isn't going anywhere for that matter. This is something that is a massive opportunity for us to be able to directly address, of course, in addition to the trucking market. So what does that lead up to, is our holistic vision. And I've said this before when we first launched this vision at -- not this past the CES, but the prior CES, laying out what we're going to do and everything of what this company leads up to over the next 100 years. And that's the opportunity to save as many as 100 million lives and 100 trillion hours out on the road over the next 100 years. And I think this is something that everyone in Luminar is just incredibly passionate about. And everyone from -- all the way from our customers, of course, to our suppliers, and importantly, the investor audience that we have here as well, like I said, it's just incredible having you guys along for this journey and we're all a part of this to be able to make this vision happen. I think this is, of course, in addition to, as Al mentioned, making a ton of money. We're probably one of the most philanthropic things that we could try to do for humanity and society generally, just in terms of the implications of what this can mean. And that mission is what drives everyone at Luminar to make this happen. But that all sounds grandiose, and everything. Okay, like what's the tangible reality? There is a master plan to achieve this in 4 easy steps. Just kidding, they're not easy. But the step -- first step for us, part of this, number one, is creating the world's best lidar for production cars and trucks. That's what we spent the better part of the past decade doing, and that's exactly what we've done. So we actually -- that's a check. Great to be able to have that. We're now, of course, with Iris, we developed these 4 series production consumer vehicles and we hit -- started production in the fourth quarter of last year. So step two, launch and enable the safest and most advanced cars and trucks on the road starting with high-end models. It's basically like flagship vehicles for these automakers. And that's generally what we've been on. That's what people have planned us into initially and originally. And I think that we have generally been incredibly successful at positioning Luminar as a desirable future and capability for consumers generally when working with premium and luxury automakers. But of course, it's not -- you're never going to get to where you need to go to save many lives or build a $1 trillion business at the end of the day by just sticking to flagship vehicle models. The important part here is that you can start to move towards mainstream vehicle models by launching and enabling proactive safety and highway autonomy systems on those vehicles. And that is the transformation that we're starting to go through now. Our recent announcement last week with Mercedes-Benz, I think, is perfectly exemplified of that, of something where we can move from a single vehicle model to now, in their words, increasing that by more than an order of magnitude to be able to enable this vast opportunity ahead and moving towards mainstream vehicles. But of course, it's not just them, Mercedes, you could argue that any Mercedes vehicle is a great vehicle. But we start -- when we started to have partners, everyone from Nissan to SAIC in China, those start to see -- you start to see that vision for how that opportunity is ahead when people are saying they want to be able to standardize this across their lineup by the end of the decade. So that's a critical part for the next step. And that's what we're going through in the throes of and in the transition of right now. And for step two, for that matter, we will be, I think, checking that box over the course of the next year as we begin launching on these flagship vehicle models, starting with folks like Volvo and Polestar and then ultimately, to the other automakers that we've talked about historically, and Mercedes, of course, too. The last step is to democratize advanced safety for everyone. And I think that ultimately, we see Luminar as something that should be a fundamental right, not just an option to all drivers out on the road, given how vast of a problem this is and how important this is. That said, I think that there continues to be additional significant opportunity for monetizing the upside for comfort features, in addition to that, when it comes to autonomous capabilities, software services, and we'll talk a little bit more today on the insurance front. So there you go, 4 steps. So one significant bit of news here. If when you add everything up, is that Luminar is now planned into over 20 production vehicle models and variants across the globe from major automakers. This goes to show how we -- we're making that leap from that second step from what we're talking about into that third step and starting to go much more mainstream. This is no longer something like I said, that's just meant as a novelty for a flagship vehicle, this is something that is exponentially accelerating and becoming more and more widespread. So it's funny, even when we added all of the different models and everything that we've had, even I was surprised by the end of the day of just taking a step back of what we've been able to achieve. So again, a huge congrats to the team across the board in helping enable this, to make this happen. Of course, we have to be able to continue and execute, and that's what a lot of the speakers that we have here today as part of our team and our leadership will be speaking to and how we're making this happen and how we're delivering on this. But the important part is the OEMs are there, the business is there, the contracts are there, people are there, the execution is there. We have everything set up for success. Now of course, it's clear how this transitions into a multibillion dollar order book for us as a company. But it's not just about future promise. I think this, all of this, as these models get phased in at different years throughout the next handful of years and throughout this decade, Luminar is very uniquely positioned to be able to deliver real revenue, scalable profits and exponential growth from our multibillion-dollar order book over both the near term and the long term, and, I think, frankly, have the biggest opportunity to be the first profitable company holistically in this space and industry. Or maybe even just autonomous vehicles, generally for that matter, by delivering real product into real cars. We transitioned from just thinking that, before, when we were going public, a single vehicle model out to these 20 different vehicle models. And that's driving now what I think is really kicking into high gear this year as an inflection point. So that order book starts translating into revenue. And that transition is enabling us to expect triple-digit revenue growth every year for the next handful of years and continuing to exponentially accelerate. Tom will be speaking more to specific details on financial guidance and what we're looking ahead for. But when you guys know the laws of exponentials there too, you start getting to some pretty massive numbers when you're at these breakneck growth rates holistically as this translates across the board from, like I said, order book into revenue. So transitioning a little bit to industrialization and execution at a global scale. I want to be able to talk a little bit more about how we're making that happen and also how we're supporting this to be able to execute on these programs, scale accordingly and give a little bit more insight and a couple of updates on those fronts. So just taking a step back, over the past handful of years, Luminar has developed a truly global footprint. And this is a very important part in establishing ourselves among the automotive ecosystem and significant part, because automakers are very geographically diversified. It's very, very industry-specific in this world of automakers, but we are able to successfully execute and support customers from across the globe. And I think never more exemplified by us constantly -- and the leadership that we have constantly traveling around the world to make this successful. So when it comes down to it, I want to be able to speak a little bit more to Mexico and specifically, this new high-volume automated factory that we have that will ultimately be coming online in the not-too-distant future for -- in Monterrey, Mexico. But also talking a little bit about the team composition, I think one thing that's important to note that's just very unique to Luminar is we have this combination of both deep technology experience and at the same time, deep automotive experience, and sort of architected the company from the beginning to be able to be at a merger of both of these worlds. And I think this is sort of exemplified in a very unique way, probably more than anything, that we've sort of defined ourselves as a sort of new class of automotive technology company that has the opportunity to work and partner directly with automakers to see through this technology into the real world, into consumers' hands and build a huge scalable business out of it. Normally, technology companies have to work very closely with or license a product to a Tier 1 or a legacy automotive Tier 1 supplier to be able to have any shot at being able to theoretically get put into a car. But the problem is that you don't end up controlling your own destiny, you don't end up controlling your own product, and it's not clear that there's any viable path or solution with that approach to be able to make a great experience or a great product generally for the end user. And so we're leveraging our huge array of -- I think we have like a millennia of -- many years' worth of talent that we have, even in the lidar systems domain and optical systems domain as well and technology generally as well as software with automotive and that talent base. So moving to Monterrey, Mexico. We're happy to be able to say that we now have the majority of our new production lines installed as part of this capacity plan and as part of this new factory build-out that we've been doing to be able to scale up from our more manual lower-volume facility in Mexico to this dedicated high-volume facility that we're doing to build out of. And this is where we get to start seeing all these huge industrialization efforts pay off from both a product standpoint as well as an automation standpoint in this new factory. So we'll have Debbie Poppas and Jeff Jaisle will be able to speak more detail to this later on. But the important part is that this factory is now expected to be able to come online ahead of our prior guidance for the second half of the year that we previously guided to. Now, we expect it to come online in Q2. And throughout the course of the rest of the year, we'll be focused on the validation and qualification of the facility to be able to prepare for global series production vehicle launches, including Volvo for their charter production. So taking a step back, I think it's important to be able to take a look at the end-to-end automotive vertical and ecosystem here. And the important distinction with Luminar and everything that we have today is that we're operating literally all the way from the semiconductor level, the chip level up through the stack, through this trend to the optical transceiver, building out the lidar system, being able to produce this in series production and working closely with our contract manufacturing partners to make that successful and happen. The software stack, of course, for what it takes from a base software standpoint as well as the advanced software layers that we're developing on top of this. OEM partners, again, working directly with them to make this successful. Consumers, being able to get consumer recognition of Luminar and the benefits for that, that's obviously what we've been making a big push for and already is starting to gain huge traction and will only be accelerating as OEMs start launching vehicles with Luminar on a global scale over the course of the coming year. And at the same time, insurance. That's something that we want to talk a little bit more about. And at the same time, we have some news on the semiconductor front with the creation of a Luminar semiconductor entity with leveraging all Luminar's expertise in subsidiaries in the semiconductor world more generally, which I think is maybe an underappreciated aspect of Luminar and something that we have continued huge growth opportunities in more generally as part of a core engine of what powers us. But literally, end to end, that's our opportunity. That's our TAM, that's our ecosystem. People may see us as a lidar company, which is true in some ways as part of the foundations for what we've had. But there's a lot to this story holistically when it comes to a transformation as a solutions company. And that's exactly what we've done. We're going to speak a little bit more in monetizing the ecosystem. But one of the things that I think is going to be critical to that. And the thing that we want to talk about today is insurance. We've established a new exclusive partnership with Swiss Re, leading provider of global reinsurance and I think actually the second largest in the world, to be able to better quantify the benefits of Luminar's lidar and the benefits of proactive safety and highway assistant autonomy capabilities for vehicles, to be able to translate those safety savings into real numbers that they can provide to insurance companies globally. At the same time, we have an opportunity and actually will be launching the Luminar insurance product ourselves to be able to capitalize on that opportunity by already, off the bat, underwriting some additional cost reduction for consumers, understanding that even a tiny difference in safety makes a massive difference in economics. And we have an opportunity to capture upside in that $1 trillion market. It's not something that you'd think about or expect, but the economics are actually shocking when it comes down to this and the second order effects of what Luminar can enable. So what's on the table is we have -- the average cost of auto insurance in the U.S. is about $1,750 per year. If you add that up over the course of a 12-year vehicle life span with the average life span, that ends up being over $20,000. I mean just think about it, that's like a material portion of the overall cost of the car. And all of that is just going towards -- or at least the vast majority is going towards the overall notion that vehicles will not actually prevent accidents in the first place, it's because vehicle accidents happen, it's because collisions happen, that's the whole point of insurance. So if you imagine a scenario where Luminar comes in, we're disrupting the industry even just take a base case of what, for example, Volvo and their safety engineers, which I think are generally conservative, said, were proud of the opportunity to improve safety and collisions right off the bat at 20%. We think we can ultimately, the theoretical limit is up to 7x improvement in safety. But even at 20% of these numbers is massive. And what ends up happening is that the technology even despite the notion that we're being adopted for the benefits and the features and the savings capability -- life-saving capability today, the technology will ultimately end up paying for itself and potentially even multiple times over along the way. So we look forward to being able to see that realized in the industry as well as having the opportunity to capture some of that upside ourselves. So enough about me talking, we have an incredible lineup of guest speakers here today from across the board in the industry. And some of the best figures of industry have fortunately, been able to take a moment and take some time to be able to actually speak on behalf of the vision for what they have and how Luminar is transforming that holistically for their company and beyond. So fantastic lineup. We're going to go through a little bit more detail. Not all of these will be here with me, but this is across the board for the company. So first up, we have Volvo. Jim Rowan, the CEO of Volvo, will be able to come on and say a word.

Around 5 years ago, we launched the Volvo Cars Technology Fund. It's mission was simple: invest in early high-potential technology start-ups with the hope that those firms would help us accelerate our own technology development. Luminar was one of the first companies that we invested in, in June 2018. It was a company with promising technology for the future, led by people that we believed in. And now 5 years later, our partnership with Luminar is stronger than ever. Lidar will come as standard on the new Volvo EX90, our new all-electric SUV. That means for the first time, a global production car is equipped with high-performance lidar and the related software. Together, they enable a new generation of smart and safe Volvo cars that sets a new bar for our industry. Lidar's ability to detect objects up to 250 meters away in the glare of bright sunlight as well as in complete darkness makes a big difference as we look to take safety to the next level. We see lidar as an icon of 21st century safety automotive, just like the 3-point seat belt was in the last century, it can be groundbreaking. With its centimeter level accuracy and high resolution, it can estimate shape and size of objects, map the geometry of the road ahead and even estimate the size of speed bumps and potholes ahead. And all of this detail is instantly processed by Volvo's new core compute technology, allowing the car to act accordingly. In other words, an important step forward in our journey towards even safer cars and the safe introduction of autonomous drive technology. Our partnership with Luminar is a great example of our approach to developing new technologies. Luminar embedded technology is one more reason to look forward to our next generation of Volvo cars, so stay tuned.

Awesome. Thank you, Jim. That Was awesome. And what a fantastic partnership, more generally and awesome be able to see what a difference it will make for you and your customers. All right. So next up, we have Daimler Truck, the world's largest producer of commercial vehicles. And the CEO of Daimler Truck is here to be able to say a word about the state of the industry in Luminar.

My name is Martin Daum, and I am the CEO of Daimler Truck. Daimler Truck is the world's leading truck manufacturer in 2022. We've sold 520,000 vehicles, and we have a strong presence in all regions. In North America, for example, we are the clear market leader with our Freightliner and Western Star brands. But we do not want to lead our industry only today. We also want to lead it tomorrow. That is to say, we aim to lead the technology transformation towards zero emission and automation. We are fully committed to autonomous trucking because we see huge potential in this technology. There's more safety, more efficiency and a very attractive business case. We aim to launch autonomous trucks by the end of this decade. To achieve this, the right technology, of course, is crucial. An autonomous truck must be able to recognize its environment in a reliable way. And this cannot be done just by cameras. We therefore focus on a combination of lidar, radar and camera technology. Lidar's capability of long distance perception is key to make autonomous trucking work. And there is no doubt that Luminar is one of the key players in the lidar industry. We therefore decided to partner with Luminar and we even invested in the company. We did so together with our partner subsidiary, Torc Robotics. Torc's autonomous fleet is equipped with Luminar technology. This technology is part of our sensor suite, and it is providing us with a reliable basis for the development of the autonomous driver. Let me conclude by saying how glad I am to see so much interest in autonomous trucking. I'm sure it will be the next big thing, and at Daimler Truck, we are all in to make it a reality. With that, I wish you a very successful Luminar Day. Thank you so much for your attention.

Awesome. Thank you, Martin, too. That's fantastic. And yes, I think the whole commercial trucking space is -- there's huge opportunities ahead. And I think also one underappreciated aspect of what's possible. I mean it's just as important to talk about collisions with consumer cars, collisions with big rigs are that much more impactful when it comes down to it. Okay. Next up, we have Polestar and Thomas, their CEO.

I'm Thomas Ingenlath, CEO of Polestar, a global pure-play EV brand with a clear mission to improve society by accelerating the shift towards sustainable mobility. Our core pillars are design, technology and sustainability. And our ambition is to create electric performance cars that push the boundaries of the automotive industry at large. Since day 1, we embraced collaborating with industry leaders who are experts in their field. That includes Luminar. Today, Luminar supplies the lidar for Polestar 3, the SUV for the electric age that we launched last October and have now opened for orders with lidar. This significantly enhances the car's advanced driver assistance systems to create a superb assisted driving experience, but it also supports the future of autonomous driving as that technology continues to evolve. In January, we announced an expansion of our partnership, taking our relationship to the next level. The partnership now includes scope to work on integration of lidar in our future cars, starting with Polestar 5, which we plan to launch in 2024. We are pioneers and so is Luminar. Luminar and Polestar have developed a close relationship, and the dialogue about future integration of lidar as a design-driven and beautiful piece of innovative tech. As different as our companies may be, we feel strongly aligned in spirit and ambition as young and upcoming leaders in the electric automotive future. We look forward to combining our R&D and product design expertise with Luminar's spearheading innovation. We believe that Luminar is at the forefront of long-range lidar, and our closer collaboration will allow for greater innovation in our cars to come.

Fantastic. Awesome. Thank you, Thomas there, and with a great partnership. All right. Yes. And Polestar, for those that don't know, is a leader in the new EV revolution there, and is already producing what, tens of thousands and, what, soon to be hundreds of thousands of electric vehicles out there. And excited to be implemented across vehicles in their next-generation lineup. So next up, we have Mercedes and hot off the heels of the recent substantially expanded deal announcement that we have with Mercedes, I'd like to be able to welcome Markus, the CTO and on the Board of Management of Mercedes-Benz.

Hello, ladies and gentlemen. My name is Markus Schafer, and I'm Chief Technology Officer responsible for development and procurement at Mercedes-Benz. I'm sure most of you have already heard by now, but I'm happy to repeat the good news. Mercedes-Benz has found a valued partner in Luminar. Luminar slider sensors are helping us to enable Level 3 conditionally automated driving at our next-generation Mercedes-Benz models. Lidar allows us to build an additional redundancy in sensing modalities, braking, steering and the power supply. And by using the best-in-class lidar technology from Luminar, we can achieve high range even for the smallest items with low reflectivity in the infrared spectrum. This will help ensure both high Mercedes and safety standards and customer satisfaction. Customers will benefit from Level 3 automated driving at higher speeds. We aim up to 130 kilometers per hour or around 80 miles per hour at our ultimate stage. Functionality is likely to include automatic lane change and highway-to-highway transfer, and we intend to roll out these features worldwide. We are glad to have found a great partner in Luminar and look forward to the amazing achievements we can accomplish together.

Fantastic. Thank you, Markus. And I think all of these just go to show, again, the exemplified transformation between Luminar as a company holistically from what we were as -- from a consideration of a commodity supplier even a supplier generally, if you notice when these companies are talking when the -- not only is it at the highest level and at the strategic level for the overall business for these automakers but also speaking about us as a partner. And that's the most important aspect for how we can successfully collaborate, deploy these systems, scale and work with that whole truly vertical ecosystem that we talked about earlier. So a word on the product that we have for Mercedes, which is going to be a next generation of Iris. But the quick history, and we'll have others talk it through a little more. Back in, what, 2016, 2017, we originally launched out of stealth mode with this lidar center called Model G. Took us a lot of letters to get to G in the first place there in the previous 5 years, as you could probably figure alphabetically, but ultimately transitioned to Hydra. Hydra is what has sort of established us in the industry and supplying it to autonomous test and development vehicles and initially automakers to be able to have the first look at what Luminar's technology is, the first scalable semiconductors and chips in the product and showing breakthrough performance in something that can be manufacturable. Iris, of course, is our first series production product. That's what we developed over the course of the years for automakers to be able to enable it into series production consumer vehicles. Now what's next is what we are now launching, which is Iris+. Iris+ is the next generation of product in this Iris-lidar family that significantly builds upon what we did initially with Iris and furthering the holistic capabilities of the product. So what you're seeing in terms of benefits from that, right off the bat, you're going to have as much as 3x the performance of what Iris had. So that's a combination of both range, resolution for the product. And the key is being able to see resolution at range for even those hard-to-see use cases like the things that Markus, the Mercedes CTO was talking about, that uniquely and only Luminar can fulfill. That, of course, translates to further improved safety in terms of covering even more types of scenarios for the vehicle. It has a 20% slimmer profile as well to be able to further provide seamless integration with automakers from an implementation and aesthetic standpoint and an opportunity to be able to improve upon that by likely as much as another 20% in addition to that product from better integration modalities and capabilities that sort of integrated in partnership with these different automakers on the different vehicle models. Lastly, this translates into a highly manufacturable product. We're expecting Iris+ to be as much as twice the efficiency of Iris from a manufacturing capacity and as well as capital efficiency standpoint for launching the product. And that translates through design for manufacturability and better scalability into millions of Iris+ product that will be deployed across consumer vehicles globally in the second half of the decade or by 2027. It's noted that Iris+ will first become available for consumer vehicle production models as well as truck production models for start of production beginning in 2025. So if you actually zoom out and take a look, it's -- holistically, we've been able to leverage all the benefits of what the Iris architecture has, but upgrading each of the individual components, both at the semiconductor level as well as at the subsystem level to be able to further enhance the performance and to be able to improve the overall efficiency and capabilities of this for maximum safety and scalability of the product across vehicle lines. But we're not even stopping at Iris+. And given this is sort of the first opportunity that we've had where we've even talked about product road map capabilities and I think the first opportunity, well, I think as a public company that Tom will talk about long-term models and capabilities as well, I want to be able to give a little bit more insight in terms of what's beyond. So in January, we acquired the lidar division of Seagate Technology. Seagate is the largest producer of hard drives and data storage solutions globally, which, what, they produce like tens of millions of these hard drive systems on an annualized basis. And there's actually an interesting and surprising amount of similarity from a component and supply chain perspective for what Seagate's done from a complex optical and optoelectronics systems perspective as well as what's possible on a lidar. And so this deal included -- the assets of this, including IP and specific IP that was developed over a period of years, and exclusively licensing other IP that Seagate has developed over decades of time. This will be leveraged to substantially enhance our development effort for the next generation of lidar even beyond Iris+ that will enable an additional step function in terms of capabilities and cost downs that will allow this to go from, call it, millions of vehicles that Iris+ is on to tens of millions of vehicles. So with that, like I said, a huge thank you for all being here today. There's a lot more throughout what we have in terms of the presentations that various members of leadership will be giving. And excited to be able to hand this off to Aaron Jefferson, our Vice President of Product Management. Thank you, everyone.

Hi, I'm Aaron Jefferson. Happy to be here today. Thank you for that warm welcome. I'm here to talk to you today about our product strategy and our product road map. And Austin mentioned a lot about safety. Safety is where I've spent most of my career, and it's really the foundation of why we're here today. If we take a look at our markets, our goal is really to deliver 1 singular product, 1 singular SKU into all of our different markets, consumer vehicle being the key market that gives us the volume, gives us the scale, gives us the learning to be able to provide reliable, robust solutions into that market and then scale across into trucking, into robotaxi, into the adjacent market. And so we utilize these really to grow ourselves and understand exactly what technology is needed. There's varying requirements across these 4 different markets. But ironically, our solution is very robust, whether it's vibration, whether it's EMC, magnetic interference, our solution is scalable across all of these, and we're able to really focus on that solution for the market and then deliver into all of our consumer bases. Now we've talked about Iris as a foundation. It's the reason we're here, it's the wonderful work of the likes of Jason -- Dr. Eichenholz and Dr. Weed and Austin, of course. But Iris is that foundation. And as an organization, Luminar quickly recognized that you just can't deliver that Iris lidar. You have to understand what that data is, what it provides and how to understand that data and deliver the understanding to the vehicle so that you can control the vehicle in the right way. So perception is essentially taking all that point cloud data, that sensing data and making sense of the way you and I do when we drive. Is this car going to pull over? Is it slowing down? Is it speeding up? So it does this without all of the drowsiness and people trying to read and do all the other crazy things people do when they drive. And so we utilize artificial intelligence and machine learning, which Dr. Weed, again, and Dr. Annie Guan will talk about as well. So basically, have a seen understanding and try to give that information to the vehicle so that it knows exactly what to do. From there, we purchased a company called Civil Maps last year. And mapping comes into play, not only for autonomy in terms of localization, understanding where you are, where everything is around you such that you can better control the vehicle, but it also allows you to update maps to understand when an intersection has changed from a 4-stop to a roundabout. So you can understand that detection, you can understand that change, you can detect it. Basically collect that change and then deliver that change to mapping companies, to our partners as well as our customers and our OEM providers. So again, that is a new area of space. If you look at our product software offering, it is just a very complementary piece that helps with our perception software, but also helps shore up our product offering as we grow into the stack. And then Sentinel is essentially taking all that information and deciding when to brake the vehicle, when to steer the vehicle and being able to understand and basically deliver that control. And so we want the capability to be able to deliver this full stack solution to our customers. But then also for me as a product person, I want the ability of our team to be able to understand what our lidar can deliver, how good our perception is such that we're not dependent on any of our customers to relinquish this into the market before we can say, hey, we can achieve this. So what we've shown at CES is proactive safety. We've shown highway autonomy. We can show the capability of our lidar without waiting on something to enter into production. And that gives us a very powerful message in terms of our capability and again, to be able to demonstrate, not on a PowerPoint, but what you'll see later today at the test facility, what really can be achieved with vehicles. So we achieved our Sentinel beta solution, which was improved proactive safety. When we talk about proactive safety, we're talking about -- when you look at Euro NCAP today and crossing scenarios and cycles and pedestrians, there's a lot that comes with that. But that's still limited, that's still in the daytime, that's still under a certain miles per hour. Our goal is really to deliver on the increased speed, being able to detect objects, small objects, and to be able to control the vehicle in a very safe way and deliver the performance necessary to increase the safety, but deliver those numbers that I also mentioned earlier in terms of the number of deaths and the number of accidents. Here's what we demonstrated at CES and what you'll see a bit later. This is a 2-year-old on a body car. Again, so we have the ability to detect this and basically stop the vehicle. The last scene that you saw where you come behind, you come from a curve and that small child is coming from behind a vehicle, that's a real scenario that no car on the market today will stop for, okay? And the nice thing about this is it doesn't matter if this -- ADAS technology, this technology is really the first technology, it has to be future-proof. When I first started working on this -- I won't speak to my age, but when I first started working on this, there were no electric scooters on roads where people were driving by. So it doesn't matter if you're on a scooter or if you're on a Jetson hovercraft, it doesn't matter. Our sensor is going to detect that and it's going to stop the vehicle, no matter the scenario, which is really nice. Next, you go from a simple safety scenario such as this to driving on highways. We're going to show as a lower-speed, 40-mile per hour small object detection. But imagine you're utilizing systems today. You're taking all this input, you're understanding the Super Cruise systems that are out there, autopilot systems that are out there, and you're driving 80 miles on a highway. And now your hands free and there's a tire or a rear wheel in the center of the road, and you might damage your car, you might cause a huge pile up, you could do all kinds of things. So it's important to be able to detect those small objects. And that's the direction that the industry is going. And that's how we build our road map to, again, increase that performance. Here, you're going to see stopping for a tire, broad daylight, and then you're going to see it at night. So again, if you're driving on the highway at high speeds, let's just say you have a broken head light, you still want your car to be as safe as possible. And the fact that lidar can detect these things and perform in advance of causing an issue is very key for us. Again, we talked about perception. I won't go deep into it. But one of the nice things about lidar is you get the best scene context that's available in 3D. And so if you think about camera, camera is really good at understanding the scene, while lidar is good at understanding the scene, but also understanding where everything is distance-wise and position and how that changes over time. And so where systems fail today is understanding that and they don't have to cross check with other sensors. Within this singular component, we have a complete understanding of the scene, we understand road edge, lanes, free space, objects moving, whether it's pedestrian, cyclist or vehicles. And we know where we are, and then we're able to basically control the vehicle. And again, this perception software is able to think like you and I even better and make good decisions for the vehicle. And like I said, Dr. Annie Guan will talk about how we do that and why that's important. Next is HD maps. We mentioned that we purchased a company and so what we did at CES was we were driving around in Vegas, and we were able to essentially see our vehicle drive around, collect data and build maps real-time. And as I mentioned before, the goal of this is really to deliver the functionality such that as you drive around, you're collecting data, you're taking all the necessary data to be able to deliver functionality to the vehicle, to better position yourselves in the lane as well as provide updates. And so now, imagine we have 1 million of these vehicles on the road between North America and Europe and globally. Essentially, we can map the world with the precise data that you need to be able to provide those updates and again, making driving safer. And the only sensing that can really give you the granularity and the precision data that you need is lidar. So here, you think about what systems are today, hands off, you want to be able to be comfortable. You want to take that safety guard down, put your hands down on your lap, relax and trust that your vehicle is going to be able to perform correctly. Our job is to make sure that we're doing safe implementation of our lidar and our software systems so that you have that comfort level and you feel safe and that the consumer market feels safe to deliver -- to purchase and to drive those vehicles. Next, I'll talk about our lidar road map. Again, we talk a lot about execution today, making sure that we're successful in how we deliver this technology into SAIC already, in Volvo and Mercedes in the future, but just as important as our road map. We have to maintain our competitive advantage from a performance and technical standpoint, but we also have to make sure that we address size, cost, power and we get this thing to the point where it can go in any segment vehicle at any price point, again, to deliver the safety that Austin talked about. So SOP in 2022 last year was [ SIC ] for Iris. Later this year, we expect to go into production with Volvo. We talked about Iris+ for Mercedes which will be in the next-generation platforms by mid-decade. And then our next-generation development, which is key, and we'll hear more about that from our advanced team from Taner and the others, is taking our key technologies and really driving the market. Another advantage of our organization is the component group, you'll hear from Mike and others later, where we have these key components that really give us the differentiation for our product. So with those together, we're not waiting on the supply base to come up with new technology, we're driving that new technology. We are basically setting our own destiny and delivering the performance necessary to keep our advantage to get size down, cost down, maintain our performance or improve upon our performance. And so that's really exciting in terms of being at an organization that can really drive and deliver that. It's a huge differentiator that Austin mentioned, gets overlooked, but we're going to emphasize that more and more as we go forward. Iris, again, the first 3D automotive-grade lidar to enable highway-speed autonomy. You heard Markus Schafer mention 130 kph , 80 miles per hour, to be able to do that, there's no system on the road that can do that today. And one of the bigger topics was really around getting the lidar accepted into the roof line. Designers of vehicles are some of the most difficult individuals to convince that technology needs to be in a particular place. And so that design is their baby. They're very eccentric. But what we've been able to do is allow the data to speak for why you want that vantage point. You want to be able to detect lanes, objects, free space. You want to see as much of the road as possible. The same way we are in position in the vehicle, you want that lidar there as well. So that is a huge paradigm shift, if you will. We saw some of this in China, but in terms of Volvo really being the one to accept it. And then all of our customers now, we have fewer and fewer discussions around where to put the lidar. And now it's about how do we integrate and how do we maximize performance, and that's the direction we needed to head in. And then you have Iris+, Austin mentioned 20% smaller and slimmer. You'd be amazed at how much a few millimeters makes a huge difference in terms of fuel economy and importance at an OEM, but we're delivering that today. And we're able to see further small objects. So again, increasing our performance for higher capability, higher speed capability and safer implementation and then also designing such that we have higher levels of production capacity. Again, as we go from generation to generation, we are going to -- we expect to go from hundreds of thousands of these sensors to millions of these sensors. And so the manufacturability aspect of that is extremely important. So having the capability we have in-house, working with our design teams, our teams sitting closely together and working together is key for us to deliver that. And then lastly is our next-generation lidar. One of my jobs here is to take all these different inputs. What does Mercedes like, what does GM like, whatever customer, the consumer market and bring all those requirements in and say, okay, what's going to be the best product for the market? How do we work with our R&D teams, get all of these inputs, and we have to make something that is scalable, that meets the, say, mass market customers as well as the premium customers, delivers the functionality. We will not sacrifice our performance and dumb down our product just to be another me-too. We really -- it really is important that we have that technical advantage, but also address the size, cost power such that this is easily adoptable into all of our customers. And that's really what we focus on for our next-generation product. So next up will be Dr. Weed to come and talk about -- dive a little bit deeper into the technology, why Luminar, our competitive advantage and a little bit about our software. Again, safety is not a privilege, it is supposed to be standard, and we're working toward that. Thank you for your time.

All right. So again, thanks, everybody, for the attention today. I have the opportunity to start the pivot into peeling the onion back. So we've talked a lot about the what that Luminar is doing, and we'll start talking a little bit more about the how. So I'm going to set a little bit of context particularly focused on differentiation. Obviously, as an investor, this is rather important. There's a lot of folks out there selling LiDAR. When it comes to differentiation, there's 2 things that may be not obvious. And this is what Luminar focuses on. So we're not going to get buried in the details of specific sensor specifications but we'll show some proof points of where this all leads up, but there's 2 areas in which we differentiate and it's important and it's very deliberate. We set out to future-proof our customers' vehicles. This has already been mentioned, and I'll show some detail into what this really means, how we future-proof their vehicles. We're also on the mission to get this tech into every car on the road. This means we need to cut down and flatten all the barriers of adoption, not just high-end luxury cars, but also vehicles that anybody can afford out on the road. So what do we mean by future proofing. From a sensing perspective, this means, we have to be able to do more than what the vehicle needs to do on the day it drives off the lot. This has been mentioned and it's going to be a continuingly important message in the automotive space into the future. You've heard some of the leaders of the biggest automakers in the planet start to talk about this. But how Luminar does this is by delivering sensing technology that can detect, track objects at significantly longer ranges, enabling much higher speeds of operation. We can do this while simultaneously understanding the threat assessment level of these objects. This is this idea of a contextual detection. Where is the thing I'm detecting? Is it actually a threat? How is it moving? So all of this information is really important to be able to know, is there a thing? Do I care as a car? And then how can I plan around it? And then thirdly, we need to be able to make this technology, this data stream be available in all conditions. Sunlight, setting sun, nighttime, shadows as well as in inclement weather. Of course, any sensor, including the human eye is going to degrade as weather gets worse and worse and worse, but we need to allow that sensor data to be still available and give as much capability to the vehicle as possible in all these conditions because it's when you need it the most. And so how a lot of this comes together? We show an example. It was mentioned this tire in the road. You'll probably see this increasingly present in the broader landscape because at the end of the day, as we move to autonomy, everything that could cause harm to the vehicle or disrupt the drive becomes collision relevant, even a tire on the road, a brick. And so we need to be able to think about these things. And it's a good case of how a lot of this comes together. So what we show here is a data example. Our Luminar car is highlighted in green. It's going to be driving to the left, and we show an aerial top-down view of data. Highlighted in red here is the tire on the road that we are driving up on and doing a stopping scenario for. And so this is a very small object right? In this data stream, it's called an occupancy grid. We're basically looking and mapping the data in 3D and looking down on it from above and we can understand where the road surface is, kind of subtract that out and highlight anything that is collision relevant in front of the vehicle. That's why it's black in front of the vehicle and kind of there's objects and barriers off to the -- up and below to the left and right of the vehicle. But we can kind of see how the environment is in front of us. So we can make sure that regardless of who's in control of the vehicle, whether it be a person or an automated system, we know not to hit things. And that's really, really important as subtle as it seems. And we need to be able to do this at significantly longer and longer ranges the faster the vehicle wants to go. So what you see here, a little bit of nerdiness we can dive into the white curve that sweeps through here and makes a function of detection range versus vehicle speed, specifically safe vehicle speed is from a third-party estimation on this from UNESCE. Now there's some references you can look up later. It's actually a really interesting report that assesses how far away you need to know where something is to be able to safely handle it. So we can map against this how far away we can see things and then figure out how fast we can safely drive. So if the goal is 130 KPH, again, 80, 85 miles an hour, we need to be up above 150, 160 meters. It's quite far, particularly when you start caring about things like a tire. Cars are easy. Pedestrians are actually really pretty easy too for LiDAR, but small still collision other things are what's difficult. And that's where our differentiation comes. It's unlikely that even the best of our customers are going to hit the road on day 1, unlocking 130 KPH, they're going to build to it. But we need to allow them that opportunity to build to it. The sensor can't be the linchpin. It can't be the -- I should say it can't be the limiting factor. The software needs to be proven, validated and developed. And we need to give them that road map. And so while the rest of the LiDAR landscape is really kind of hitting up against this kind of wall that exists in capability of legacy technologies, we give this huge roadmap, and that's really a big point of differentiation in our customers' eyes. And it gives them the ability to provide their consumers a more long-term experience where the vehicle is better every day it is out on the road as opposed to today, where the best day is when it leaves the lot. So you don't have to take my word for it, take our word for it. We have pretty fantastic partners and to show some remarks here from a customer or a partner that couldn't be more aligned to our mission, Nissan. And they've done extensive work validating and benchmarking what's necessary to achieve the same things we're talking about as proactive safety within their platform. [Presentation]

Yes. So as you can see, braking is really just scratching the surface, right? There are so many strange scenarios that lead to these millions and tens of millions of collisions every year on the road. And with partners like Nissan, we're going to address these. And in a way that can be deployed across fleets of cars that have been sold years ago, and that's a really exciting opportunity. So the next step of how we differentiate and how we're going to achieve all these things we talk about is through cutting down the barriers to adoption. You'll hear a ton of information in coming talks about what we're doing at the hardware component level, driving the cost and -- the cost down and the performance up of the actual piece of hardware that goes in the car. Critically important, something we've been focused on from day 1. We are a company that is anchored in hardware technology. The other half of hardware technology though was integrating it, right? We already heard some comments on this, but it gives some context to why we've spent so much time investing in partnerships in the ecosystem. We've partnered with a lot of leading roof suppliers like Webasto and Inalfa. We're starting to work with others in the ecosystem to provide a more holistic hardware solution so it's easier to adopt this technology, right? We want it to perform as much as possible and set it up for success, not just hide it away somewhere in the vehicle and say, done. But the last area here is critically important in an increasingly important way, we deliver software, not necessarily so that we can go sell software, although that is the nice upside, but that we can accelerate the adoption of LiDAR-based processing in vehicles. This is new, LiDAR isn't a ubiquitous technology out in this field. And most of our customers haven't already spent billions of dollars developing it. While some have, not all have. And at the end of the day, our biggest opportunities in the marketplace have a little to do with our LiDAR competitors and more to do with how do we accelerate their adoption of the technology itself. And so this means delivering software across a huge gamut of types all the way down to how we control and make the most out of the data at the LiDAR itself, through things like perception where we're extracting understanding from the 3D point cloud all the way through to controlling the vehicle. And we provide all of these pieces so that we can partner on platforms like those shown here. Partner with customer automaker systems and allow them to solve their problem with all these pieces and really accelerate the adoption. And so to dive a little bit deeper into the real heart of a lot of the perception activities is one of our colleagues to talk a little bit about Luminar's AI engine, Annie Guan. So welcome her to the stage. And thank you very much.

Hi. My name is Annie Guan. I'm leading Machine Learning and AI team at Luminar. Today, I'm going to share with you how Luminar is using our AI engine to deliver the next-generation product safety and autonomy. So why do we need to use AI and machine learning, with only the raw point-cloud from the sensor you only see the point in space and reluctant -- reflectance, but you don't know where the points are and how they behave in the future. In order to develop a product safety in highway autonomy features, it is essential to identify what those points are and how they will behave over time. In order to classify those points into specific objects such as people, cars and cyclists and predict the possible future movement based on the classification, we need to leverage the power of our AI engine. In this image, you can see that now we have applied Luminar AI engine, we have gone from just pure raw point-cloud into a rich environment model, showing all different types of objects in space with their exact location, height, width and heading. By leveraging the inherent information contained with the 3D POINT cloud data, we achieved a much higher level of accuracy than if we were to infer this information from only the camera images. The rich environment model that is produced using the machine learning techniques allow us to understand the surroundings and also the potential hazards on the road. With this in mind, let's take a closer look at the system and technology that power the transformation from the raw point-cloud to rich environment model. Here, we can see the different elements of Luminar AI engine pipeline that is capable of consuming a raw point-cloud and generate 3D bounding boxes for objects. 3D lane points, barrier points, road points and also the drivable space. At the core of the Luminar AI engine, we have 2 neural networks. One is the 3D object detection for dynamic objects. And second is semantic segmentation for road, lane and barrier classification. To train the neural networks, we rely on a large amount of label training data. We're also building a training infrastructure, which utilizes a data engine and distributed training where data in the model are paralyzed across different GPUs. To deploy the neural networks on the vehicle edge device, we must optimize network for fast, efficient influencing using AI acceleration techniques. Now let's dive into the details of the neural networks. Let's first take a look at the 3D object detection neural network. The model consumes multiple frames of raw point-cloud and outputs a 3D bounding box with object classification. First, we performed a Dynamic 3D voxelization which divides the point-cloud into a grid of voxels and map points. The 3D voxelization step preserves all the raw points. There is no information loss in this process. Features are then generated from each point which are fed into the network backbone where we do sparse convolution. The sparse convolution step only processes non-empty voxels rather than processing all voxels in the grid, which reduces inference time. Different objects have different sizes in the point-cloud. In order to address the challenges, of varying sizes, we apply a multi-scale feature aggregation where we create feature maps of different scales. For large objects such as vehicles, feature maps with a coarser resolution are used. While for small objects such as pedestrian, feature maps with a finer resolution are used to capture the more detailed information. Finally, we do a regression and classification step to determine the 3D bounding box's dimension, orientation, object classification and confidence score. With this network, we can detect a vehicle at 200-meter with centimeter -- 200 meters with centimeter level of accuracy. The second network we have is semantic segmentation. The model takes the raw point cloud and outputs classification for the 3D points for road, lane and barriers. We first extract point-wise features, which are then projected into a bird's eye view. The feature map from birds eye view then goes through an encoder-decoded network backbone to get point-wise confidence score. Finally, we reproject this into the 3D space to determine the 3D point-wise classification. Now life-saving proactive CP function requires both high confidence detection and extremely fast inference time. In order to achieve this rapid response, the neural network needs to run on edge device in the vehicle fast and efficiently. This is where AI acceleration and inference comes into play. The goal of AI acceleration is to reduce the inference latency while minimizing any potential impact on the model accuracy. At Luminar, we have been working on various techniques to achieve this goal. Some of the approaches we have been working on, including quantization, pruning, mixed precision training and knowledge distillation. By using those techniques, the model can be run much faster for [ SEP ] critical applications while still maintaining high accuracy. With these techniques, we have reduced the run time by a factor of 2x to 4x and can run on automotive-grade SoC in under 20 millisecond. Machine learning and AI-based models require large amounts of high-quality label data in order to train the models. In our case, we have partnered with Scale AI to provide ground truth labeling across different ontologies covering both dynamic and static objects. With that being said, I will now turn back over to Austin for a conversation with Alexandr Wang, CEO and Founder of Scale AI on AI and machine learning. Thank you.

All right. Well, thank you, Aaron, Matt and Annie. That was awesome. And obviously great to be able to see the latest as it relates to our AI, and we've been sort of developing behind the scenes, but I think great to be able to put it into the spotlight at a pivotal time for the industry. So we have Alex here, the Founder and CEO of Scale AI, which is our key and exclusive partner to be able to help enable this whole ecosystem around all things, AI. But maybe for those who aren't familiar, Alex, do you want to say a word on your company?

Yes. So I founded this company is Scale AI. We're a deep partner of Luminar so really excited to be here and pretty amazing to see all the stuff that you and your team members have been working on. We're an AI company based in San Francisco, California. Our core product is all around enabling data sets for AI and machine learning across a wide variety of applications. So as an example, we helped build ChatGPT with Open AI. We've worked with them since 2019. So I like to say it's been a 4-year overnight success on that one. We've worked across the entire autonomous vehicle and automotive spectrum, working with folks from General Motors to Toyota, to many of the tech players and have an exclusive partnership on the LiDAR side with Luminar. I will -- I'm sure we'll talk about this in a bit, but one of the reasons for that we see is just the, a, the commercial traction that the Luminar folks have is just incredible across all of the auto OEMs. We do business with them as well, and we know it's not easy to do business with them. But -- and then the data, the quality of the data at a core fundamental level is just -- is dramatically better than the competition. And so we just see this incredible opportunity to do so much more with the data, which we're going to talk about. We also work -- we see -- we work across a wide variety of industries as well across large e-commerce players, like Instacart and Grab to large -- to the government, working with the U.S. Army and the U.S. Air Force. We're based in San Francisco, and it's a really exciting partnership that we've built here.

Yes. Awesome. And that's to say the least. I think we've been -- it's been a few years now there too since originally starting to collaborate on all of these things and getting it together. So -- maybe just a question on the data side. Like how do you see the importance of this data, the fidelity of the data and volume of data as it relates to Luminar and maybe even more generally, why Luminar?

Yes. So I think if you take a big step back on the sort of arc of artificial intelligence, and Annie certainly got to this in her recent presentation, if you take a big step back the 2 major drivers of the entire revolution of artificial intelligence have been massive increases in data and massive increases in computational capability. Computational capability has been powered by Moore's Law. And then data and data availability has been the sort of like secret driver behind a lot of this improvement. So if you look at ChatGPT, for example, or the sort of language models that are -- that everyone's kids here are using to cheat on their homework, that -- those models are trained on literally 1 trillion tokens of data from the Internet as well as whatever data that they can get a hold of. There's stories of these AI companies, they go hire people to go like scan old books and like and literally scan these old books for data to be able to train their algorithms on. So data in many ways, is the lifeblood and the sort of data hunger of modern AI algorithms is really -- is very difficult to appreciate, but it is -- they are massively, massively hungry. And you don't get diminishing marginal returns. Like what you see with ChatGPT, for example, or the progression on the language side, which is sort of all the hotness is that -- as you keep getting -- putting more and more data, the models just become smarter in very unexpected and quite profound ways. And so when you look at how that translates into the autonomous vehicle and automotive ecosystem, the same thing holds. If you look at across much of the autonomous vehicle development today, the data volumes, despite -- it is a lot of data that's being produced, but that pales in comparison to what -- the amount of data that will be produced from production rollouts of the technology and production fleets. When many, many consumers are buying vehicles with high-quality commercial LiDARs, the amount of data that will be produced from those sensors is just -- is astronomical. And I think we're going to see, in many ways, the same trend that we saw on the language and -- on the language side of AI in the sort of autonomous vehicle side of AI, which is once you're able to amass these huge amounts of data, you get just dramatically new and different performance from the algorithms, which is -- so at a core level, that's exciting. And then I think, okay, so what, if you have much stronger AI, what do you do? And anything you build -- obviously, there's autonomy, which is one Holy Grail. I think you have a lot of interesting and exciting opportunities in other parts of your business as well. So the insurance business. Insurance is obviously a place where if you have very fine-grained data and very fine grained algorithms, you could underwrite significantly better, you could drive a lot -- at this point, I'm pitching your business plan, but -- but I think across like every part of the automotive stack, huge amounts of data become incredibly, incredibly valuable. And this is one of the reasons why this partnership is exciting for us because there's just not that many companies that are going to have this much data. It's going to be rare to have this much data, and we're excited to see what incredible things we can build together.

Yes. I know it's awesome, and that's for sure. And I think this is really what you're seeing today with the Luminar AI engine being relevant for detection, recognition, being able to adapt collision avoidance or assisted driving capabilities. As you said, that's just the start. And actually, yes, it's funny that you alluded to that even like on the insurance side and other things that as you -- as we go through that whole vertical ecosystem there too, there's relevancy literally all the way from the semiconductor level all the way up through the stack to insurance and stuff. But that's great. And well, I mean, you guys are -- what you've got to be the leader holistically in this space now, too, from all the wins that you've established and volume that you're processing there too. So that's a huge -- and I'll say congrats to yourself, too. So what do you think about from the importance of the fidelity of data as well because that's an important one of we have this huge volume that I mean, for everyone's reference here, historically, autonomous test vehicles and development vehicles, I mean, you're only talking like fleets of hundreds of cars that have LiDAR on them that are collecting data and leveraging it. The whole point is now we're able to get out there with millions of vehicles. And you don't even have to wait till then. It's like tens hundreds of thousands of vehicles in the more immediate term that we're scaling through. But there's been -- obviously, there's vehicles out there today that have other kind of more basic sensors. But what do you think about the importance of the 3D data and the 3D aspect of it and the fidelity of that data in creating a holistic system and solution?

Totally. Well, I think it's -- I think, a, it's critical to safety. If you don't know how far something is away from you, you're -- you might run into it, which would be really bad. I think that's relatively intuitive. But -- I mean, I think a lot of it -- you're right, there are some cars with some basic LiDARs that have like a few lines. And that doesn't really give you a whole lot to do any sort of meaningful machine learning or deep learning on top of. But if you have on some level, if -- let's say, if my eyes were LiDARs and I saw this room, and I had like a perfect 3D scan of everyone in this room, I would be able to discern almost strictly more than if I were just -- if I had like -- my optical sensors, which are my eyes. So the sort of the full amount of information that you can get out of the data, which is what a lot of these AI algorithms are trying to like get ours like how do you get the full amount of information that exists implicitly within all the data is just far, far more through, I think, the incredible sensors that you all have built. And there's -- again, there's kind of this like a clear theme in AI, which is once people build the mechanisms by which you can collect massive amounts of data, the Internet being like one of these collection mechanisms for getting a huge amount of data from humans. But once you build these mechanisms, sensors being a core component, then you get -- you have the ability then to glean so much more out of the AI algorithms. And I think that if you go use case by use case, you could start with just autonomy functionality. Autonomy, contingent upon autonomous systems are going to be safe autonomous systems. And you're just -- it's very, very, very challenging to get requisite levels of safety without very high fidelity 3D sensing. I think the other piece that is exciting is I think there's components that you might -- that I think we'll discover over time. There's additional upside, which is that we're -- Annie kind of alluded to this. But like we're -- as a machine learning AI community, we're very early on in the sort of journey of understanding and processing and building AI systems on top of LiDAR data and 3D data. Even if you take an example, ChatGPT that's -- this is like decades into using machine learning AI on natural language processing and all of a sudden, we have these huge breakthroughs. And so I think there's a lot of upside as the machine learning and AI community continue to improve upon the algorithm techniques to be able to process all this data, which is, I think, very exciting. I mean I think that in many ways, the sort of -- one of the exciting things about our partnership is will we have a ChatGPT moment as it pertains to automotive data.

Yes, absolutely. And I think we've had pretty much exactly that moment when it comes to the LiDAR itself, and now it's already starting when it comes to the software. And that's critical, working collaboratively, of course, with our partners on these things holistically as well as not just the automakers, but the platform providers, too, that we'll be talking a little bit more and Taner will be talking about in his section as part of this holistic ecosystem for what we have. So that's relevant. And I think the interesting part, there's also seemingly a lot of synergies in terms of overlap of opportunities and customers and everything there that I think should hopefully accelerate that as you guys are providing sort of the back end behind what we're doing for the Luminar AI engine, but also working with some of those folks yourself. So maybe I'd say like what do you see in terms of additional opportunities in advancing the ball forward for the industry, like, for example, anything from cross collaborations, mapping, other kinds of functionality. You mentioned insurance.

Totally. Well, I think the exciting thing is that, in many ways, one of the ways that I view our partnership is that Luminar has built this incredible platform upon which there will be one of the greatest data sets for autonomy and driving data produced. And then it's contingent on both of us, how can we -- how -- what are the exciting things we can build out of this data set as well as how do we work together? I think we're both very much so in the automotive business to get to this outcome of much safer vehicles, right? I think that like at the core, that's that certainly what drives us and myself and our teams. That's what drives you and your teams. And I think we've already seen that our teams are going to be able to -- have been able to and will continue to be able to work together quite effectively in accelerating that mission, right? How do we get like every automaker in a position where they're shipping safer vehicles. And sometimes they require some muscle from our sales team, sometimes require some muscle from your sales team. But at the end of the day, I think we're making great progress along that journey. So -- and then yes, to your point, I mean, there's so much to do with the data, insurance, there's mapping, building a live 3D map of the entire world that's constantly up to date. This is not something that exists today. I think we all -- like Google Maps gets really out of date quite quickly, actually, not the fidelity that you need for autonomous driving. So there's just so much to do with the data that I think I think we're going to discover new opportunities in future years as well.

Yes. That's awesome, man. Well, thanks for taking the time here too. And feel free to speak to Alexandr in the audience during the day, great to be able to have you in person and awesome to see you guys leading this whole next AI revolution and providing all the back-end infrastructure to make the Luminar AI engine successful.

Yes. Excited to -- thanks for having me.

All right. Shoots man..

Time for a 10-minute break, please. 10 minutes, we will start again right away after 10 minutes. Thank you. [Break]

Luminar Day will meeting will resume in 5 minutes. Please take your seats and remember to silence all mobile devices. We're going to start again in 5 minutes. Thank you. Luminar Day will resume the meeting in 2 minutes, please, 2 minutes. Please find your way back to your seats. Thank you. We'll limit our day. Please take your seats. We're going to begin in under a minute. Please take your seats. Thank you. Please welcome Co-Founder, Luminar, Dr. Jason Eichenholz.

All right. Good afternoon. Thank you, everyone. Hopefully, everyone is getting their sugar high from those brownies going. So as Aaron and Dr. Weed talked about, there's really 4 components to the LiDAR system. Now we knew from very early days, back when we were developing the Model G that we needed to move into the eye-safe region or as you hear about 1550 nm. We knew that because we could use 10x the laser power, 17x the photon budget or 1 million times the energy and still be eye-safe. But people told us it couldn't be done, but we were able to make it happen because we developed what we call our chip-level up strategy. Now many people focus on the economics to make these 4 components of the LiDAR system work together. But the reality is you need technology innovations that allow you to unlock the autonomy, unlock the performance. In order to do that, you can't do that with commercially available off-the-shelf components. And as a small startup, we weren't able to get the attention of the industry. Now we are but we also have vertically integrated and developed our chip-level up strategy. When we got the technology to work in the performance, the economics came along for the ride for free. So if you take a look at our technology, and here is an example of the technology of our Black Forest Engineering ASICs and our OptoGration APDs and I'm going to hold them up here for -- to give you all a sense of scale of what you're looking at. You're looking at this stuff on a microscope slide. And it's on a microscope slide because you need a microscope to see it. So we've got our subsidiary, Black Forest Engineering that produces the mixed-signal ASIC, OptoGration that makes the high sensitivity in a high dynamic range receivers, and we broke away from the limitations of silicon. We're able to have these hybrid systems that enable the eye-safe 1550 window, and they became accessible to all. Now as these devices got smaller and smaller as you see them and they're underneath the microscope, we're also able to do some pretty cool things physics-wise to unlock the performance. And if you take a look at U1 and you see the Black Forest Engineering chip and there's a very small square where you see the APD from OptoGration, you'll notice that there's no wire bond. So this gold wire bond is just for scale or about the size of 100 microns, size of the with your hair. There's no bump ons because we couldn't have handled the physics of the capacitance of that wire bond. So we've got both physics and economics working for us enabled to unlock the performance. Here's an example of the technology and the electronics that went into the Model G with our first receiver where Black Forest Engineering and OptoGration came together. And if you look at those, you see a time to digital converter chip up in the front. And the reality is we were able to take all of those electronics and then put inside in Hydra 100 of those Tdc boards. All that technology came together and to get there Luminar was the first to deploy a custom ASIC for LiDAR back in 2016. And you can see the progression over time as more and more compute horsepower combined with a mixed signal ASICs go in. And in 2022, Luminar was the first to take a custom ASIC like this for automotive series production, and we've taken all this technology and made it auto grade. We've been able to take tens of thousands of dollars and turn it into single-digit dollars as we put these systems together. We're running the same playbook we did before on the receivers. You can see what we did with that big hunk from the Model G of the laser and moving it into Hydra and then now what you're seeing going into Iris and our next-gen systems. As the receivers became more and more sensitive, we were able to see further and we had lower noise in each generation, we were able to then scale our lasers smaller and smaller. Less and less components, which then allowed us to drive the efficiency and also drive the point resolution. We got, again, power consumption and the economics came along for the ride. We've also significantly reduced in the latest generation of the electronic components by moving more and more functionality from those electronic boards into Black Forest Engineering ASICs. You can see how the ASICs and the technology and the electronics come together. The key to proactive safety is both autonomy and range plus resolution, you need the range and the resolution. There has been almost no progress in laser diode technology for the last 2 decades. Freedom Photonics is the technology leader in scaled high-brightness laser diodes, as you can see from that hockey stick of growth that we announced just back in January at Photonics West and Photonic Integrated Circuits. The lasers and the BFE ASICs are supporting Iris today, and their innovations will drive the next-generation platforms and innovations on the receiver side that are coming for next generation combined by functionality on the laser side. I have a saying 1 plus 1 equals 11. Freedom, Optigration, Black Forest Engineering when combined together into an entity creates a lot of power and a lot of synergies. We internally, and you've heard them called Luminar Semiconductor. We'll be launching a unique brand for this company and a new name later in the year as a separate company owned by Luminar. I am personally exceptionally excited to be taking on a new role as Chairman of this new company. I'm also -- thank you. I'm also excited that we found an exceptional leader to go run this operations day-to-day. Mike McAuliffe will be joining as CEO of Luminar Semiconductor. Mike brings decades of experience of taking companies like this and building semiconductor and technology companies, including several start-ups to exit. Mike previously led Seeing Machines a publicly held Australian company, his focus there on computer vision and bringing market leadership in ADAS, driver monitoring systems and processing to some of the world's largest automotive OEM. We're now going to leverage that experience inside Luminar Semiconductor, as you heard Austin talk about to grow that company. Welcome, Mike.

Thank you, Jason. First of all, it's a privilege, really, when I joined the company last year to join such bright people on such a burning mission. I knew to be exciting, and I was not wrong. So Luminar Semiconductor, why do we exist? What's the difference between Luminar and the 3 component companies. Jason described the chip-up strategy, described why OptoGration, Freedom Photonics and Black Forest Engineering were acquired to be the best-in-class. But now as Luminar Semiconductor, our mission really is to take those into an integrated company and to take them to the next level, the next level of capabilities, the next level of products to drive the current and future Luminar road maps, but importantly, also to create a broader ambition to be a photonics player in the wider market. So at our core, what do we do? We build chip-scale components and chip scale optoelectronic engines to solve really impactful problems for the world, working with industry leaders. And number one, of course, is our parent. Number one is Luminar, our #1 customer, solving probably the most impactful problem of them all, but there are more. And today, I hope to share with you some information on what they can look like. So we've heard about stacks today. We have big stacks, and I guess that means that we're a substack maybe. But we have our own stack too. And our stack goes from, we call it, photons to decisions. And in some ways, it represents the signal processing flow from the generation of photons with the lasers from Freedom Photonics, the detection of photons at OptoGration and the processing of those photons from Black Forest Engineering. We then combine that with advanced packaging up to what's then the next stage of DSP processing and AI processing, as described by the team earlier on. And largely, that feeds the signals for LiDAR, LiDAR software, perception and mapping as we described earlier on. So now the key insight is that the same technologies and the same capabilities that make Luminar successful, they can be applied to other markets. Other customers have similar problems who are generating photons, detecting photons, processing photons. They recognize how difficult it is, what value we can create. So they're also interested and we can create broader opportunities across wider markets. So I think it's fundamental to our strategy to understand that this common needs can drive common platforms. We only focus on the hardest photon processing problems, generation, detection and processing. This is what LiDAR is. It's one of the biggest, most difficult physics problems today, but we have other customers solving similar, very difficult photonic problems. So we can now take the same platforms, the same products closely aligned and leverage that across multiple markets. We solve 1550. There is a general perception that 1550 is an expensive traditionally an expensive process, and it's maybe a drawback. That is not the case. We have solved that problem through, number 1, architecture, as Jason described, number 2, advanced packaging; and number 3, we can drive the economies of scale and the economics associated with that by leveraging both Luminar volume and volume to the wider market. So we're shifting gears. We're shifting gears to take what has been a subscale industry, subscale ecosystem and industrializing that at scale. And we need to do that. We need to do that both for the economics and for size, weight and cost in order to deliver LiDAR at millions of units. And as an example, the other key advantage of siliconization is this. People think that maybe the biggest advantage is you control your own supply chain. That's true. You can control the quality, that's true. You can control the performance. That's true. But the biggest advantage of siliconization in general is you can take a lot of complexity and cost out of the product, complexity in cabling, complexity in optics, complexity in electromechanical. So anything you can put in silicon, you can put in silicon. And I think we can see from other industry leaders like big fruit companies that owning that stack and internalizing and siliconizing anything you can delivers elegant architectures, breakthrough performance and breakthrough economics. So what's the basis of competition? How are we going to compete? How are we going to win? What gives me confidence Well, number one, we have what I regard as really the brightest team maybe in the industry. And by the way, that's not me saying that, that's some of our customers saying that. We have about 100 people across the 3 companies today. Over 85% of them are engineers and over 1/3 of them are PhDs in engineering. And they're spread across our sites, Santa Barbara's lasers, ASICs in Colorado, receivers in Massachusetts. But our platform is also very unique. We have -- I won't go into the acronyms and some of these look pretty dangerous materials, but these are pretty special materials. They're special for generating photons at different wavelengths and for detecting photons at different wavelengths. So we have a unique set of platforms and capabilities right down to the atomic level so that we can control, design and optimize semiconductor platforms for achieving this full stack of technologies. Also, it's a very important part of our model that even though we own the process, we have our own facilities, we control the design of the process, we are also scaling now to ship millions and millions of chips, scaling a new model of state-of-the-art fab partners, packaging partners, test partners across the world. And we don't have a legacy infrastructure to worry about. We're kind of the barbarians at the gate. We have a clean sheet of paper to design a brand-new supply chain, and we can control the control of the process, the process recipes, specialized manufacturing here, and we can choose world-class, high-volume, best-in-class fabs and foundry partners in the world. And that's key to our economics model as well. So as we execute this strategy, what is the broader opportunity, and again, I stress broader outside of Luminar, we believe will be clearly the market leader LiDAR. So I don't really have to do a lot of work to generate these numbers, and you don't either. You can go to any one of the photonics companies in the public domain and look at their TAMs, their SAMs, their market analysis. And the answer is it's rather large. Across communications, aerospace, precision manufacturing, broader LiDAR and 3D vision, emerging sectors such as optical sensing, quantum computing, quantum sensing, which are very heavy users of photonics and life sciences, biosensing, gene sequencing. These in total are over $50 billion. But that's interesting because it informs the potential runway. But what we're much more interested in and what I'm more interested in is, what's the ground market pull, what's the bottom-up opportunity? Who are we working with? What problems are we solving? How do we solve them, how do we turn this into a scalable business? So something you may not know and most people don't know, because I guess it's been under the radar or I guess, under the LiDAR in this case, we have over 60 current programs at leading customers across the world, and we've shipped over 0.5 million products. And these products, I can tell you, are going to some of the most demanding mission-critical applications, you can imagine. So we have a strong reputation. People come to us to solve very hard problems, and that's core to what our potential is going forward. So what's happening in the industry landscape? Well, number one, photonics in general and 3, 5 materials is rather immature, but we have the potential and we have the playbook to go and drive that industrialization, drive that scale, drive that economics. And that's something that's core to our strategy. And we think we have the -- with the ambition, the resources and the strategy of Luminar, we have a great chance to do that. Secondly, many, many new markets and applications are really driven by laser technology. Many of the examples I shared with you earlier on, the core of that technology is laser technology, narrow beam width, tunable lasers, high-power amplifiers and they create new markets. So we're going to use that to create a new beachhead, to land on that beachhead and then expand delivering a full solution of detectors and processors as well. So that's key to building our stack and building our moat. So in summary, I would leave you with this to-do list. This is our to-do list in the company: number one, drive productization, drive scaling and drive a chip-level road map for Luminar, number two, build foundations for the company and build beachheads, build beachheads where we believe is a very large opportunity to build a separate company. So build, disrupt, deliver and scale. Four small words, one very large impact, we believe, for Luminar and for the industry. And lastly, I would say in the words of that great Gen Z song, we've only just begun. So I think stay tuned. I think it will be an interesting journey. Thank you.

Please welcome, EVP and GM, Taner Ozcelik.

All right. Good afternoon, everybody. Yes, I've been in automotive for 2.5 decades. In fact, I came out of retirement and have built several billion dollars businesses from ground up. I've learned that the success in automotive relies on getting 5 things right. And that's what I want to talk to you about today. First, you have to innovate and you have to innovate very rapidly. This is true for any industry, but it's also true nowadays with automotive because automotive has become basically a tech industry. You have to design and build the products that can scale easily. You have to have an architecture of the product line that makes it so easy to leverage that massive R&D that you're pouring into every product every time and rinse and repeat and build that into the new products. Architecture happens to be also very key to competitive moat. I've learned this over my career working for some of the great. You have to make sure that whatever you design can last for decades. That's called being automotive grade. And it's important because that know-how incorporates the massive learnings of the industry that is 100 years old, it's one of the oldest industries, many scars. And last, but not least, and arguably the most important is the talent. People talked about it. It's important to have a talent base that is infused with both automotive experience and heavy technology bent. Austin shared that earlier. So let's dig into the details. As I mentioned, everything in tech relies these days heavily on innovating and innovating rapidly. You can see examples of this in many companies that either became incredibly successful or disappeared. NVIDIA, my -- our partner here, my Alma Matter, is a great example of a company that rapidly innovated and reestablished the company to new heights in almost every major generation of products. On the other hand, as we all know, our beloved BlackBerry, once everyone's indispensable technology fell by the wayside because of what, lack of rapid innovation. So innovating and innovating rapidly is incredibly important to cementing that leadership. That's what I'm after. It's important in tech, it's important in automotive as well. Well, I can now confidently tell you because I'm going to show it to you later, that Luminar is not only innovating fast and delivering on our milestones, whether it's our customers' milestones, our internal goals. We're also accelerating our pace of innovation and development. Here's the proof. Hydra, as Austin showed, was the first product introduced in 2020. It was a game changer product, many automotive companies, including Toyota, adopted Hydra for R&D and it was deployed in industry in thousands of units. But Hydra took 3 years to bring it to a much more functional level. The automotive industry terms that maturity sample as B sample. Hydra was Luminar's first attempt to build something super complex yet very effective, in my opinion. Hydra put Luminar on the map as a credible tech company. And arguably, it paved the way for the LiDAR industry as we know it today. Iris, on the other hand, took a little too -- over 2 years to get to that same stage, and it was launched in 2021. Won several customers as Austin showed, and establish itself as the first volume product designed for automotive and its slim design and also what Austin showed, basically and the performance coupled with it raised a lot of eyebrows in the industry. As you know, and we already talked about this, we launched our first vehicle with Shanghai Auto in China last year. And now we're proud to announce that with Iris+, we accelerated our development pace by more than 2 times. So we not only improved our execution, our rate of improvement doubled went from 28% from 36 months to 26 months from Hydra to Iris to 58%, 26 to 11 months. So we accelerated. We accelerated big time. Execution is everything in automotive. 100,000 parts in automotive come together to have one SOP. You cannot be late. Otherwise, all of that massive R&D basically gets wasted. So execution is critical. So with that, I think we're now in the top class of top tech companies in terms of velocity of innovation and execution. So let me show you how we did it. We kicked off Iris+ in March of last year. In 1 month, we had the mechanical design locked up by working very closely with our lead customer. Next month, the development picked up speed with the electrical design and thermal analysis. In June, we already had modeled it incredibly intensely and stress tested it with something called FEA, finite element analysis, basically a bunch of differential equations and complex math to solve. In August, we had already built the receivable modules. This is incredibly important. Receiver module in LiDAR is one of the critical subsystems that determine the performance of the product. Mike and Jason showed you some of the building blocks of that. We were able to build this incredibly complex, super critical system class because we had the entire semiconductor chip designed and developed in-house. We would not have built this fast if we didn't have this capability in LSI in-house. It typically takes at least a year to get to a new product from a -- to get a new product from a semiconductor provider. And when you get it, it's not exactly what you want. It's essentially the least common denominator of what the entire industry of customers want. I was on -- I know this because I was on the other side of the table. Now I get to sit on both sides, semiconductor and systems. And I think Luminar has a huge advantage because of that. In September, we had the first transceiver, which includes a very complex laser subsystem. In October, we had the first [ engineering ] sample in our hands. In 7 months, a record for Luminar. In November, we had the first point cloud out of Iris+. I remember that day really well. The feeling was basically similar to having your baby speak for the first time. I think Iris+ said Marco, our Vice President of Engineering. That took 8 months from inception, another record. At the beginning of January, we had the first fully assembled B sample of Iris+. It was built in Orlando in our automated pilot assembly line. And yes, we even automated our pilot line. Thank you, Debbie and Jeff. And here we are today, ranging in our state of the art test facility and producing a beautiful point cloud that I think you guys will see later today. Iris+ right now is not only speaking words, but it's now telling us a story, a story of what it sees around it, like a baby would do. It's quite a rewarding feeling all in 11 months from inception to today, I am incredibly proud of our team. So ladies and gentlemen, this is Iris+, world's highest performing, automotive-grade LiDAR with largest order book. It's built to scale, built to outperform and built to lead the industry, all in record time. But we're not stopping there. Austin mentioned this. Recently, we have acquired the LiDAR team of Seagate. Yes, Seagate. You might be thinking but what does Seagate have to do with LiDAR. Of course, Austin also mentioned some of that. But Seagate is the largest producer of hard drives. Hard drives are optical devices. They go into data centers that all have 24/7 incredible, and they have to operate incredibly reliably without major hiccups all the time. So reliability in data centers is one of the most critical things to achieve, and Seagate, I believe, nailed it. by building highly reliable optical systems in tens of millions of units every year. So does that sound familiar? Yes. You got it. LiDAR is also an incredibly complex optical system. It has to be built reliably. So you can perform the ultra-stringent automotive reliability requirements. You wouldn't have guessed. There's a lot of parallel. So when we first saw the opportunity at Seagate, I mean, literally, as a company, we pounced at it. I've never seen a company this excited on any acquisition. One unified company. And what I show here and as you can see here, there's a tremendous amount of commonalities between highly reliable and highly scalable optical hard drives and LiDAR. We believe we can leverage about 80% of the know-how in building hard drives to LiDAR. And our partnership with Seagate is not only that it expands to supply chain in manufacturing and leveraging all of their expertise in that regard as well. So this, in my opinion, is immense. We're incredibly delighted to welcome the Seagate team to our home and leverage their R&D and manufacturing expertise to accelerate our ambitions in LiDAR for the automotive industry. Welcome home Seagate team. So how do you build products that can scale fast and delight the customers at the same time. The answer, and I worked with a lot of companies, many people who worked in a lot of companies. The answer lies in platformization. If you look at any company that has scaled its products to tens or even hundreds of millions of units, whether it's Apple, NVIDIA or even Mercedes or Volvo, you'll see that their development is always based on platforms. Platforms allow for rapid development. Platforms allow for tailoring to all customers' needs without the full-scale customization. Platforms create immense leverage on R&D, which can be rinsed and repeated over generations of products. It's often how well a company has platformized this development is what determines its ultimate success. At the core of platformization is modularity. So we create modules of technology, each of which can be innovated on its own at its own pace. It's own constraints and often leveraging different parts of the industries. So take for instance, the receiver module. The receivers we use are all built in-house in LSI at a rapid pace, fitting exactly to our needs. In our history of innovation in APDs, the avalanche photodiodes, which are the detectors, is a testament to the power of modular detectors. Our APDs have never failed in the field, and we have supply to lots of mission-critical applications like military and aerospace. [ and LiDAR ] on the other hand, I believe is another subsystem that I think we haven't even scratched the surface in terms of innovation. We're on a rapid path to increase both the performance and the cost and lower the cost at the same time in lasers. So I would say you haven't seen anything yet. Stay tuned until you see the next generation of lasers that we're building at the moment. You will be blown away. Processor complex, we have is also highly specialized, and we build our own ASIC in-house again in LSI. We have, again, a massive runway in innovation in ASIC ahead of us. So modularity and platformization is absolutely key to scaling and winning multiple designs, and that's exactly what we're doing. So again, here is the proof of that. We have racked up 8 top-tier OEM design wins to date with the same Iris family spanning across more than 20 production vehicle models. That's because of its platform design. Because it was designed as a platform, same family of products, we're able to garner millions of units and design wins, which in dollar terms equates to billions of dollars of revenues over lifetime. You heard about our announcement with Mercedes last week. That announcement basically was about increasing the breadth of our product adoption across Mercedes to a wide range of vehicles amounting to, again, just in itself in billions of dollars in revenues from a single company from a single product line. That's the power of platformization. The other key point I want to make is this. Just as platformizing our own products is important, so is partnering with highly successful platform companies. And that's NVIDIA. That's Qualcomm, that's Mobileye. Each of these companies have platforms on their own, and we're delighted to be part of their platform. Say, for instance, NVIDIA we're thrilled to be part of their Hyperion platform as an exclusive partner to them in LiDAR. Mobileye is a clear leader in ADAS based on cameras and we're delighted to be partnering with them in their MaaS or Mobility as a Service initiative. Mobileye happens to benefit from LiDAR in urban and other ODD environments like operational design domains. And Qualcomm is our newest partner, as they have announced at CES as part of their Snapdragon Ride platform. So we're super excited to be working with the world's best of the best in computing platforms. And next, I'll turn it over to Gary, an old friend of mine from NVIDIA to say a few words about our partnership.

Hello. I'm Gary Hicok, Senior Vice President of Automotive at NVIDIA. Luminar is a highly valued automotive ecosystem partner for NVIDIA, and we've been working closely together since 2018. As safety is our highest priority in developing autonomous vehicles, NVIDIA has long recognized that LiDAR is an important component of a diverse and redundant sensor suite and helps us create a robust autonomous vehicle perception system. Our collaboration with Luminar enables us to integrate best-in-class technologies for autonomous driving functions. NVIDIA Drive is a high-performance, open platform designed for the entire transportation industry to build automated and autonomous vehicles from passenger cars to commercial trucks, to robo taxis and to shuttles. At the core, our fully programmable AI supercomputer gives OEMs and Tier 1s the flexibility to select the best and most cost-effective sensor solutions for their unique needs. In addition, NVIDIA drives them software tools to uniquely enable the physically accurate simulation of sensors, their positions on the vehicle and then allow run time, real-time perception algorithms to ensure this system will perform safely in the real world. In 2021, NVIDIA announced that we had selected Luminar's Iris long-range LiDAR solution as part of the sensor suite for our NVIDIA DRIVE Hyperion 8 autonomous vehicle development platform and reference architecture. This forward facing long-range LiDAR will be used in DRIVE Hyperion's Level 3 highway driving configuration. We believe Luminar offers a unique, sustainable solution that complements DRIVE Hyperion, which is accelerating the development of autonomous vehicles around the world. By offering automakers a qualified complete sensor suite, coupled with NVIDIA centralized, high-performance compute and AI software DRIVE Hyperion provides everything needed to develop production autonomous vehicles. We applaud Luminar's effort in bringing to market cutting-edge LiDAR solutions that meet the stringent performance, safety, security and automotive-grade requirements for the autonomous vehicle industry.

Thank you Well, thank you, Gary. So we're obviously absolutely thrilled to be partnering with you and I cannot wait to know with you together at Luminar at the speed of light, and he knows exactly what I mean. I've talked a lot about how we do things at Luminar, but not so much about the architecture of what we're building. As I mentioned, architecture is incredibly key to the success of the products that are built on that architecture. So good architectures allow for scale. Bad architectures waste a lot of valuable R&D dollars because they cause major misses in the market sometimes. Good architectures delight customers because they form the best, they perform the best. They are built for reliability and are easy to build. Good architectures are simple. Bad architectures are complex and cannot easily scale. Elegant architectures are the ones that are both simple and high-performing, simple and high performing. And that's exactly, I believe, what we have. That is my [ hats off ] Austin's brainchild. Someone that devoted all of his life to photonics so much so that he didn't even want to waste time at Stanford. So when you're building an elegant architecture, you first have to start with the ingredients or the building blocks. Ingredient selection is absolutely key because building a highly scalable, highly reliable architecture is absolutely important. And those are the -- these 5 are the core elements of building a world-class architecture. Everything you see here in blue is built internally at Luminar. Why is that important? Because with this, we can not only build the product the way we want, but we can build the building blocks of architecture the way we want at the pace we want. That's the power of vertical innovation. Now that power in ingredient selection results in this elegant architecture. On the right is our elegant -- wonderfully elegant architecture. It's simple yet hugely powerful. That's the reason we're winning. On the left is all of our competitors' architectures. Some of them already in the graveyard. Wrong architectures can be unfortunately very unforgiving. I asked that to all the LiDAR companies have folded in the last 12 months. And let me mention 1 more thing about this 1,550 versus 905 debate. I bought a company in my old company that built 905 technology. And we marketed and sold our products to practically everyone, but Luminar in the LiDAR industry, including Innovus and SI. I know what 905 can do and what 905 cannot do. In that case, I was focused on short range in mainland robotics. The nuance is the architecture. All 905 needs arrays. In fact, the detectors themselves are a bunch of detectors clustered together like SIPMs, silicon photomultipliers. Arrays mean cost. It's now getting basically in the order of hundreds of these to build a good 905 detector. In our architecture, we use a few, usually 1 or 2 detectors. So even if the 35 semiconductor material is a bit more costly in wafer terms, we use orders of magnitude less wafers. So the cost of 1,550 regardless of what they tell you, doesn't matter. As Jason mentioned before, I think, it's practically in the order of single-digit dollars. Single-digit dollars for us. So that's the fallacy of 905 argument. And what's more, 905 cannot give you the range that 1,550 can. Here's the proof. That's why people are buying our product. So both performance and cost scale better with 1550. And the icing on the cake, 1550 happens to be more eye-safe. So I want to wrap up my talk with arguably the most important thing about building products for automotive. And that has to do with building these products for an industry that has seen it all in terms of companies that came and went. Companies that struggle to build shareholder return because they struggle to build highly reliable products. We know them all. So we take being automotive grade, to the heart. We pour our hearts and souls into designing something that can last for the life span of vehicles. Products that can tolerate harsh environments from extreme temperatures to harsh vibration requirements and to all sorts of other severe operating modes. Some of you may not be familiar with what it means to be automotive grade. The acronyms I show here are only some of the things that you have to do to be called automotive grade. Just take a look at the soup of alphabets. That's why many start-ups cannot survive and cross the chasm as the great old author Jeffrey Moore put it once. We did it. As you can see, being automotive touches every single stage of development from concept to serious production and behind -- beyond. Many companies fall into the chasm somewhere in the middle and successful companies not only across the chasm, but understand the power of iteration to get it just perfect, especially from development to spec. Happens to be the area that people just follow. We did exactly that with Iris and Iris+ and crossed the chasm, iterated rapidly and brought the company to where it is right now. So building products and great products is just not easy. There are many facets of perfecting that art. But 1 last thing in that regard that I want to leave you with is testing. Testing is pivotal to any product success. The more you test, the better you test, the more success you have in the marketplace. It's that simple. So we built what we believe to be the world's largest most advanced facility for testing. It's over 300 meters long with up to 500 meters of total ranging capability. LiDAR is a sensor you measure and reconstruct the world in terms of photon by photon, which are elements of light that each travel at 300 million meters per second. So our testing facility had to be so advanced that we could measure these deviations from this insane level of precision needed to build a great product. It's truly mind blowing to see what we do and how we do it. And finally, I would like to thank our entire employees for building these great products, delighting our customers and building a great company along the way. Our talent, in my opinion, this is why I'm so excited about this company is unmatched in terms of skills, breadth of experience, level of passion and just sheer intellect. We have the best of both worlds, as Austin mentioned. We hear about half of our company come from pure tech background and the other half from automotive, a perfect blend for a great journey. I can't wait to build many, many great new products that will change the automotive industry, help save lives and bring more comfort to our travels. With that, I'd like to say thank you and hand it over to my colleague, Debbie Poppas, Vice President of Manufacturing Operations.

All right. Thank you, Tanner. Good afternoon, and hopefully, all of you are feeling the energy now. As has been talked about all afternoon, Luminar's LiDAR is recognized as industry-leading. So what does it take to constantly design and build premium quality LiDARs at scale. It's an approach to quality that touches everything we do. In this session, we'll take you through our efforts to successfully industrialize and scale our manufacturing capabilities to serve global automakers whose rising demands are placing pressure for near perfection in every aspect of our business. In the automotive industry, exceptional quality performance is one of those rising demands. There are 3 keys to our automotive quality approach at Luminar. Advancing industry standards, commitment to quality and implementing key quality initiatives. We are leading the industry to raise the bar for safety and autonomy and engaging with industry standards organizations. Luminar's business system aligns to proven automotive best practices, captured by automotive quality management standard, better known as IATF-16949. It was released by the International Automotive Task Force Group. I was a Board member of the AIAG organization, which is one of those members. It's a U.S. Association that's tied to that task force. Luminar has also been assessed by a high-end German OEM against the BDA Standard 6 Part 3 and received favorable feedback on addressing the observations that were noted during that assessment. Not only do these standards -- they're required by our customers, but compliance delivers improved business performance. Luminar is driven to secure an automotive certified business and supply chain. Our quality management system consists of 5 main components: design quality, as Tanner shared with you, ensures that we design for reliability, manufacturability and assembly. Product [ debt ] validation confirms that our LiDARs meet the demand of our customers' applications. Built-in quality is the commitment our advanced manufacturing engineers and contract manufacturers make to build consistently flawless LiDARs every day by leveraging technology, and error prevention. Supplier quality maintains alignment with Luminar's design requirements and our quality expectations. Continuous improvement is a cornerstone of our commitment to quality as we drive robust problem-solving to improve performance, also efficiency and productivity. Luminar is investing to scale our quality management systems and our processes while embracing 0 defect strategies and advancing an automotive quality mindset throughout the organization. These strategies are translated to our entire supply base. You will see shortly a video demonstration of built-in quality in Luminar's high volume assembly and manufacturing lines at our contract manufacturers. A tour later today of Luminar's long-range test facility illustrates our product validation approach to ensure LiDAR performance excellence. Luminar works closely with our supplier partners to ensure that our products are the highest quality and most efficient to manufacture. A successful Luminar supplier meets the criteria shown on this chart, Beginning at the top left, the willingness to invest their business for growth, quality and technology leadership; demonstrated quality leadership across industries, but in particular, automotive. Product and process innovation is key. A Luminar supplier is expected to deliver required quality product at the lowest cost. 75% of our purchases are covered by long-term partnerships that provide suppliers incentives to continually invest in their business. Our suppliers meet our global product requirements and serve our production and distribution needs locally. As of today, we have 89 suppliers in 16 countries that have demonstrated ability to scale and our automotive grade capable, which ensures our customers are receiving the best quality product at the right time and at the most competitive cost. Global Supplier management operation is deployed everywhere our supply base is located, which ensures quality and delivery expectations are maintained. Our contract manufacturing partners are an extension of Luminar, assisting us actively by managing much of the extended supply base. As you heard from Jason and Mike, the strategic acquisition of the LSI businesses assures we have the most critical LiDAR components secured. We are aggressively looking to expand our ability to directly control all aspects of our supply chain. Our automotive customers require ISO and IATF certifications that span business processes, including purchasing product design, validation and manufacturing, and that translates to our suppliers. As we have progressed from an R&D tech-focused company to an automotive Tier 1 supplier we have seen that transition reflected in the quality of our supplier partners. Over the last 5 years, we have progressed from a mix of 80% ISO only and 20% IATF certified suppliers to 2% ISO and 98% IATF ISO. The objective is to continue to drive that number to 100%. In addition to our suppliers, Luminar expects to be fully certified later this year. Our industrial journey has been underway since 2016, focused first on expanding our Orlando pilot line to keep up with customer demand for samples and at the same time, producing sensors for design validation. Low volume equipment is installed at our contract manufacturers, Celestica in Mexico and Fabrinet in Thailand. Since late last year, volumes have been ramping up. We completed all customer low-volume run at rate checks at are shared Celestica facility. And as been mentioned, we started production for SAIC's rising auto R7. We are focused on the next phase of industrialization with our dedicated partners as we prepare for scale and additional customer launches. Luminar is on track to have our dedicated high-volume facilities online in 2023. Jeff Jaisle, our VP of Manufacturing, will now take you on that scaling journey. Thank you.

Thank you, Deb, and good afternoon to everybody. Ever since I was a little kid, schlepping around my bucket of Lego, I have absolutely loved to build stuff. Some say that was just last week, but that's a lie. And here at Luminar with a management team and the founders that conceived of a vision and assembled the management team for a noble vision with a great best-in-class product and customers that keep wanting to add 0s to their volume expectations, it's a great place to build stuff. And so I'm super excited about what we have in front of us. And to start off with, I'd like to show our -- a little bit more detail on our automated line that we're pleased to report is coming in ahead of what the previous guidance was and is now expected to be online in the second quarter of this year. So here, you can see a small segment of the lines. The line is built on a modular flexible architecture. So it's easy to be able to change and adapt to the changing product requirements, changing volume requirements. And it actually comes for the Iris version in 5 sections. So there's 5 separate subassemblies that all culminate in the final build of that product. 4 out of those 5 lines are not only installed, but have passed in conjunction with our customer, all the requirements necessary to get qualified for the site acceptance. And by the end of March, the fifth segment will be also qualified in conjunction with our customer. And so by the second quarter, we will be able to complete all the remaining qualifications necessary and be able to produce our first sensor at this high-volume facility. It's highly automated. Each step in that has multiple types of quality checks, whether it be 2-dimensional camera vision systems, 3-dimensional camera vision systems, which you can see there and a variety of sensors in that almost every step is error-proofed to make sure that we have a reliable product at the end. That line, as you see it, is capable of 250,000 units per year. So at the end of 2023, there'll be -- after we make the [indiscernible] sensor, there will be some ramp-up curve as we get everyone trained and ramp up the line. But by year-end, 250,000 a year will be the run rate. And with a modest investment, it can be scaled to 0.5 million sensors per year. Obviously, we have to scale up the transceiver as well in Thailand to keep pace. So we have our partner in Celestica doing the final assembly and the final test and calibration being supplied by the transceiver and optical components out of Thailand. Here's some video of a brand-new dedicated facility, which we're in the process of launching now. So this facility is designed to be able to keep pace with the facility that we have done in Mexico, all the transceiver, the Polygon balancing is all coming out of the operation. So everything for the buildup. And this is really just the beginning. The next generation Iris+, as Tanner alluded to is going to set the stage for an even easier route to our scalability goals. In conjunction with engineering, we've worked hard to make the Iris+ even more efficient to manufacture. We put in a pilot line in Orlando just to test proof of concept for what we thought we could do from an assembly. And so we've already been able to prove out with that line that the next-generation will take almost 1/3 -- or will be 1/3 of, not 1/3 less, 1/3 of the floor space required for the Iris and roughly 1/3 the labor as well. And beyond the Iris, the work that we're setting the stage for with -- in conjunction with the Seagate acquisition will take us to another whole level of momentum as we start to -- our journey and scalability. Obviously, our customers are global, so we have to be global as well. The sites where we actually are related to the operation and the manufacturing. We've got our Minnesota and our Florida sites, this is where we'll do a lot of the development and basic process work that we have to do to prove out before we go to high volume, and we'll also do some prototype builds out of Minnesota and Florida. Mexico is concentrating on Iris, final assembly and the test. Thailand is, again, where we get most of our transceiver and optical components from -- and we're happy to say that the next footprint, the site that we're looking for now will be planned in Asia. And in Asia, we're looking to install another 0.5 million units per year. So in summary, we've got the process resources to be able to develop the core technologies that we need to scale in Minnesota and Florida. We're already in series production learning every day with the products that we have. We're already -- we're ahead of guidance in that on our first high-volume ramp-up, and we're well positioned with an increasingly efficient design architecture to further facilitate that scale and expansion into Asia. And with that, I think we'll pass it on to Tom.

Please welcome, CFO of Luminar Tom Fennimore.

All right, everyone. We're in the home stretch here. I'm the final presentation for the day. I plan to cover 3 topics. First, I'm going to talk about our China business and strategy. Second, I'm going to go into a bit more detail about our new insurance product and business. And then finally, I'll close with the discussion of the numbers, both our results in 2022, our 2023 guidance and milestones and our updated medium- and longer-term financial targets. All right. Let's start by discussing China. China is near and dear to my heart. I spent 3 years living in Beijing last decade covering the automotive industry. My wife and I started our family there. We raised our first son in his formative years there. And after I left China, I would go there multiple times a year until COVID. And so I know that industry relatively well for a partner, have a lot of the relationships there today. China is a very important market for Luminar. We want to continue to be a winner in China. We want to grow and be successful in China. We want to be a winner in China, but we need to have a thoughtful strategy to execute. Let me walk you through what we did in 2022 despite a lot of challenges, including travel restrictions preventing us from getting there. Last year, we made significant progress in building the foundation of our China business. We hired Jackie Chen to run our business. He used to run Harman's China business and Shaeffler's China business. He's a very dynamic leader who not only understands the local China industry very well, but he knows how to operate within a multinational company to deliver at China speed, which is very important. We have formed our own legal entity in China. We're building out our headquarters in Shanghai. We're building out a local team, primary engineers, to support our local customers and vehicle deployment. And finally, as we discussed last year, we reached our first SOP with the SAIC R7 brand at the tail end of last year, not bad for not being able to go over there. Now before I go into more details of what our strategy is, I'm going to spend a second taking a deeper dive into the China market. Last year, in 2022, there were over 20 million vehicles sold in China. Approximately half of them were foreign brand names. What do I mean by that? That's Volvo, Mercedes, Nissan, selling vehicles in China, primarily through their JV partners, but with their brand. The other half is local China brands -- and if you actually take a deeper dive into that, about 3.5 million vehicles, the fastest-growing segment are new electric vehicles. That number has increased 5x in 2 years. It's rapidly growing. It is important in that EV segment to be viewed as a technology leader. So we see those type of companies as the ones that want to adopt LiDAR technology the fastest. If you take a deeper dive into that 3.5 million, we also look at the price point where our technology, which is the best, we believe, in the China market, but maybe not necessarily the most -- the cheapest where it makes sense for us to focus on that. And if you actually look at vehicles priced RMB 250,000 or above, which is about low to mid $40,000 depending upon where the exchange rate is, that's about 1/3 of that 3.5 million, and that is our focus area. So when you take the foreign-branded vehicles, which you're going to have natural exposure to from our global relationships and our global progress, plus that, call it, top tier of that new EV brand that's nearly 60% of the market that we believe is going to be very addressable for us. Now let me go in a little bit more detail on our China strategy. First, we talked about the focused customer strategy. The first 2 I discussed on the previous slide. We're also focused on the commercial trucking space and are making some progress there as well with that customer base. Second, we want to continue to expand our China footprint. We're going to accelerate our growth of the local engineering team there. You need to build out the right software and data structure in China to comply with local regulations, and we're starting to do that. And as Jeff and Debbie talked about, we're looking to establish our Asia manufacturing capabilities there to help serve the China market. Finally, we can't be successful in China on our own or as successful. We want to work with the right strategic partners I'm going to go into more detail in a bit on 2 of them: Pony.ai, who's an autonomous leader in the China and the global landscape and then ECARX, which really helps the [indiscernible] family of vehicles develop new technology. So we have strategic partnerships in place with them. We're going to explore other potential partnerships as well to help us grow in China and do so in a smart way. Now I'm going to go into a little bit more detail on our relationship with Pony. So last year, we expanded our partnership with Pony. As I mentioned, they're a leader in the China autonomous vehicle market as well as globally. We will be their long-range LiDAR supplier for their next generation of vehicles with an expected SOP somewhere around 2025. In addition, we extended our partnership to include the commercial vehicle market, as well as additional software and development support. Now I'm going to have -- here's a video from Pony CEO, James Peng to talk more about Pony in the relationship with Luminar

Hello, everyone. I'm James Peng, Founder and the CEO of Pony,ai. The vision of Pony.ai is to build a safe and reliable autonomous driving solutions and delivering it at a global scale. Since very early on, we view LIDAR as an essential sensor for our autonomous driving vehicles because LiDAR has the capability to view the surrounding work in 3D with great precision. So from very early on, we have worked with Luminar to use Luminar LiDARs as essential part of our autonomous driving solutions. Over the years, we have really enjoyed a tremendous growth in all aspects of our business. We have launched full driverless autonomous driving vehicles in both Beijing and Guangzhou. This is the first time we have seen fully autonomous drive vehicles, navigating the complex traffic scenarios in major cities in China. Besides the robo taxi -- we are also working on robo truck and licensing our technology to the OEMs to really make the mobility to be much safer. And we were fortunate to have Luminar as our partner over the years, and we have worked together on many aspects of our solution. And we really view the scalable and reliable LiDAR solution that Lumina provided to us can help us to make the system even safer. So our hope is that with an even deepened collaboration between Pony and Luminar, we'll be able to build our next-generation solution that really empowers very large-scaled autonomous driving trucks, and I hope that our partnership can result in even greater things down the road. Thank you.

Thank you, James. Last year, we also entered into a strategic partnership with ECARX, a key technology provider to the [indiscernible] ecosystem in China. We are working together to jointly develop an ADAS and autonomous turnkey solution specifically designed for the local China market that utilizes our LiDAR and other technologies. Here is Ziyu Shen, ECARX CEO, to talk more about ECARX and his partnership with Luminar.

I'm Ziyu Shen, Chairman and CEO of ECARX. We are a global mobility tech provider that partners with OEMs and technology leaders to reshape automotive landscape as the industry transactions to all and all electrical future. We are developing a full stack solution, central computer, system on chip and software to help continuously improve the in-car user currents and advance the development of new connected automatic cars. The ecosystem of partners we work with is extremely important to the delivery of our end goal, transforming cars into fully integrated information, communication and transportation devices. ECARX and Luminar partnership began in May last year when Lumina made a strategic investment into ECARX as part of our NASDAQ listing. At the time, we committed to work together on automotive [ great ] technologies with aim of enabling advanced safety and automated driving capabilities. Luminar is a true innovator bring new technology vital for the deployment of safe, automated highway driving to the global industry. By partnering with Luminar and other tech leaders, ECARX wants to develop a cooperative ecosystem that accelerates the transition to smart mobility. By integrating Luminar's long-range LiDAR and software with ECARX's suite of automotive [indiscernible] driving products, automakers in China and internationally will have a clear path to deploy advanced safety technologies and driving capabilities on serious production vehicles.

Great. Thank you to you, Ziyu. We're excited to work with you and develop this product together for the local China market. I will now transition to talk a little bit more about our Luminar insurance product and strategy. Austin talked at the beginning of his presentation about how our technology is going to significantly improve vehicle safety and how that is going to result in what we believe are going to be very meaningful insurance savings. As you can see from here, 80% of forward collisions fall into 3 categories that we think our technology is going to directly reduce, mitigate or even eliminate these type of insurance -- these type of collision scenarios. We also anticipate that the traditional insurance companies are going to be very slow to adjust their pricing to reflect the improvements that our technology brings. So as a result, we want to be in a position after studying this problem and looking at what others have done in the industry, how we think the traditional insurance companies are going to behave, of putting our money where our mouth is and building our insurance business to capture and underwrite these savings to help subsidize the cost of our technology, including driving standardization across vehicle lines as opposed to having it as an option. If we're not willing to bet on our technology, why should anybody else and our technology is going to create these insurance savings. And if we don't do anything in our OEM partners, the traditional insurance companies are going to capture those savings instead of the people providing the value. Let me try to frame what the opportunity here. As Austin mentioned, the average vehicle insurance premium in the U.S. is about $1,750 a year. If we're able to get 20% insurance savings, which we believe could be a conservative number, that's about $350 a year or almost $2,000 over 6 years. That's almost twice the price of our current Iris LiDAR. Now how are we going to do this? Our plan is to build a scalable and asset-light insurance product that we can partner with our OEMs with and make it scalable to work with multiple OEM partners. We also want to build the flexibility to go directly to the consumer and the purchaser of our vehicle. This insurance product, we want to offer significant discounts to consumers who purchase vehicles with our technology to encourage the purchase if it's standard or the option selection if it's an option. We plan to do this by building our own digital insurance MGA to operate the business and our own insurance captive to manage the risk. We intend to obtain reinsurance protection to mitigate the risk. We will use reputable third-party operating partners who -- and there's about several industry leaders and a technology platform, this has been done before to minimize the investment, enhance scalability and allow us to build this quickly without applying a lot of capital to it. Our initial target markets are the passenger vehicle markets and commercial truck markets in the U.S. but we will hope to scale this over time globally. We expect to have our initial insurance product ready to go in 2024 in select U.S. states. Now why are we confident we can do this? We are -- we hired Alex T. who did this at Tesla. He's the person who builds Tesla's insurance platform, he's going to oversee the construction and industrialization and launch of our insurance product. He knows how to do this well. He knows how to do this quickly. He knows how to do this digitally. He knows how to leverage technology. and third-party partners to do this in a very asset-light and scalable way. As Austin mentioned earlier, we are also partnering with Swiss Re, a leader in the global reinsurance industry. Swiss Re is working with us to both test our technology on real-life vehicles to help quantify the safety improvements and insurance savings and helping to bring our insurance business to life. We also intend to work with Swiss Re to build our ensuring pricing algorithms and to participate in the insurance risk. Now here is Russell Higginbotham, the CEO of Swiss Re Solutions, to share more about our partnership.

Hi, everyone. I'm Russell Higginbotham. I'm the CEO of Swiss Re's Solutions division. So Swiss Re, you may or may not have heard of us. We're a 160-year-old company. We're active in all of the lines of business of insurance around the world and in all the major markets in the world. And our Solutions division is actually where we try to bring all the knowledge we've gained over the many years to our market. So let me tell you, we're excited about our partnership with Luminar. We're excited to partner with them and to work with their life-saving technology to bring safer driving to the world. So mobility is clearly becoming more and more autonomous. And that means that some risks associated with mobility go away. But others, new ones, they appear. And these new risks need to be understood and assessed. So this is what we do. We capture real data around the advancements and around driving behavior and around the conditions of driving and actually, then we translate those into risk assets and risk insights. And then ultimately, what we do is help insurers around the world understand the risks and help them price the risks, which is ultimately what insurance is all about. So what we're specifically going to be doing is assessing the safety capabilities of Luminar's LiDAR sensor technology and actually working out how it works in reality, how it prevents and how it mitigates risk. So then as I said before, we translate that into risk scores. So this offers as well as helping the insurers, this offers a feedback loop to the engineers at Luminar as it's a continuous learning process around driver behavior and the conditions around driving. So Swiss Re, we're excited about this partnership with Luminar. I think it makes their excellent technology more viable. It makes it more commercial in the real world. It clearly enhances road safety and it makes insurance more affordable. And I think ultimately that makes the world more resilient. And I think if you bring all those things together, we would all say that, that can't be bad.

Great. All right. Now on to the final topic, the financial section. Let's start off. We talked about this at CES. We met or beat our 4 key public milestones last year. We reached series production with Iris at SAIC. We made great progress on our Sentinel software suite. We ended the year, as Austin mentioned with over 20 awarded program lines. We talked about Polestar earlier this month. We talked about Mercedes last week. We introduced scale this week. And we ended the year with an order book size of $3.4 billion. I'll talk more about that in a second. Here's a summary of our results for Q4 and full year in 2022. We recorded $11.1 million in revenue during the quarter and $40.7 million for the year. Both of these were in the range but at the lower end of our guidance. And the reason for that was primarily due to the timing of our NRE revenue recognition. Of the $40.7 million in fiscal year 2022, a lot of that was program development revenue, which is NRE related, which was a portion that kind of drove that timing risk. We had GAAP EPS loss of $0.40 and a non-GAAP loss of $0.26 in Q4. This was slightly higher than expectations as we accelerated investments in industrialization and future product development. As you can see, we've got a lot going on at Luminar. We're not only launching Volvo with Iris, now Mercedes with Iris+, we have more business. We're investing in our future product line. The growth is there and we are focused on the long-term value creation even if it's at the expense of spending a little bit more in the near term. We ended the year with $489 million of cash. Cash spend was higher sequentially primarily for the same reasons our EPS was higher and we expect our quarterly cash spend to decline as our new dedicated Mexico facility starts to ramp up in the second half of this year. Now let's talk about our 2023 milestones and financial guidance. We are laser focused on execution in 2023. And while we have a lot of milestones here at Luminar for the year, there are 3 that are the most critical. First, we need to scale up our series production. That means bringing our high-volume automated facility in Mexico online and meeting Volvo's SOP requirements. Our second milestone is to keep advancing our technology and product road map. Specifically, we expect by the end of this year to enter the C-phase for Iris+. We want to develop a prototype of our next-generation LiDAR that the team spoke about earlier. And we want to complete the necessary based software for the Volvo and Mercedes SOP. Our third milestone is to continue to grow our business, and we want to grow our order book by at least $1 billion this year. Now let's turn to 2023 financial guidance. We expect to at least double our revenue this year. We expect the revenue to ramp as we go through the year and really step up once we get our new Mexico line up and running. So it will be a little back-end weighted. We expect Q1 revenue to be in the range of, call it, $11 million to $13 million and EPS for the quarter to be roughly in line with what we saw in the last quarter or 2. We expect to be non-GAAP gross margin positive by Q4 as we launch our new facility, increase production there and lower sensor contribution cost. We expect to have at least $300 million in cash and liquidity at the end of the year. We expect that 2022 was our peak cash spend year and our quarterly cash spend rate should begin to improve in the second half of this year as our new facility ramps up and our launch costs associated with it ramp down. Finally, I would expect our share count at the end of the year to be in the 395 million to 400 million range. Now let's talk a little bit in more detail about our forward-looking order book. As a reminder, we tend to be conservative on how we calculate this. Not only do we only include customers that we won in here as defined by them giving us a serious production program or equivalent and having that memorialized in a signed agreement. But for those won customers, we only include the vehicle lines that they have officially awarded to us. While we're talking to them about their future product road maps and are confident we're going to be on more and more business. Look at what we've done with Volvo, Polestar and Mercedes over the last year or so as a track record of our ability to grow with our customers. We don't include any vehicle lines in this order book until it is officially awarded to us. When it -- for those awarded vehicle lines, we use IHS assumptions for volume when available, contracted pricing and conservative management judgments if they're not available. So based upon that conservative calculations, we ended 2022, as I talked about, with $3.4 billion order book. We expect it to grow at least $1 billion this year. Now I'm going to transition now to talk about our updated midterm and long-term guidance. And I apologize for the financial modeling experts in this room, this may be a little too basic, but I think it's a good overview of how we updated our longer-term model and help everyone think about how our business is going to grow. So the foundation of our longer-term model and our growth is the number of Luminar equipped vehicles sold. That translates roughly into sensors sold, maybe a little bit off if there's multiple sensors on the vehicle, but I think you get the point. So there's 3 components of that. 1 are the awarded programs. That's our order book. The order book goes from potential revenue to actual revenue when each of those vehicle lines reaches SOP. The second bucket is the additional programs at existing customers. As I said, we have a track record of growing with our existing customers. If the vehicle line is officially awarded, it's in the previous box. If we're in there talking to them and have good visibility on when they're going to put our technology on their next vehicle programs, we put them in there, probability weighted adjusted in terms of we take a haircut to it because it's not won yet. Finally, there's new customers. So these are the customers that we haven't officially won yet. We're not done winning. We're talking to a lot of other people out there. When we have something ready to announce, we'll announce it. But we have a portion of non-won customers yet in our volume projection, we'll try to quantify that there in a second. Needless to say, we're talking to a lot of people, a lot of great conversations there. Okay. So that's volume. Pricing, what's my revenue per Luminar equipped vehicle? Well, I got call, the hardware and base software ASP. So we're moving to an integrated approach. I talked before about how the base software in Volvo, Mercedes needs to be ready to go. What we're seeing now is our customers are asking us to put more software capabilities into our base LiDAR. We're kind of including that for the time being in all the hardware category. But there is a good element of software-enabled capabilities that are in the base hardware that we sell. Then I want to talk about the software and solutions content per vehicle. So what is that? That is additional functionality from our Sentinel software suite. We talked about HD mapping. We talked about insurance. So I kind of think about that as the net insurance profit. So there's going to be revenue risk. I just said let's just kind of put the profit in there. This is take rate adjusted. As much as I would like this to be on every vehicle, the content per vehicle here is measured in the thousands of dollars. But what we do is we take rate adjust it and we'll kind of talk about what that take rate adjusted ASP is going to be. This stuff, really, if you step back and think what we're doing, it's represented by the insurance business we're getting into as well as mapping but our LiDAR creates an ecosystem of value. You see it on the insurance and there, we're putting our money where our mouth is to help our customers and the consumer capture that value -- mapping -- we're going to be collecting all this rich data. We talked about what we're doing with scale. That's an acquisition we made last year. We see that as an opportunity. If our technology standard on each vehicle, when you upgrade the consumer to highway autonomy, that creates a lot of incremental profit. So some of this software and solutions content per vehicle will be upfront, some of it will be subscription and some of it based upon the conversations will be revenue sharing. And so our LiDAR can create a lot of value in the ecosystem and we want to work collaboratively with our OEM partners to try to capture that value to help accelerate the adoption. All right. That gets me to revenue. Let's, can we go back. I accidently hit the button. Okay. Unit contribution costs. Okay, primarily looking at the cost of the center. There's the BOM. There's the manufacturing conversion cost, what Celestica, Fabrinet and our future partners charge us. And then there's going to be other variable costs, warranty, logistics. In the case of when we start selling the software and solution, there are going to be some cost of services associated with that as well. And then let's look at the other key items. There's going to be my OpEx and other fixed costs. There's going to be CapEx and the investments I need to make to continue to grow. And then we have this great new business that we're forming, Luminar semiconductor -- Mike's been appointed to run that. These provide the core components to our LiDAR, but separately, we've asked Mike to go and grow that business. And we see great opportunities there, but the lion's share of the growth in this model is coming from our core LiDAR and associated business. All right. Next year, I'm going to spend a few minutes to talk about how our forward-looking order book starts to translate into annual volume and revenue. And a key point I want to leave you with is that we have the awarded business in place and the customers in place to support this massive scaling in our business. Sometime next year, once that Volvo business is up and running as well as the other business we have, we're going to reach that [ 100,000 ] plus rate, right? So we're starting to ramp up. Now let's -- when are we going to get to that million mark, right? Because [ 100,000 ], won let's go to order [indiscernible], that [ 1 million ]. Well, 70% of it -- over 70% of that volume to get to [ 1 million ], we already have awarded. Once again, I'm not talking about from customers -- all of the customers want just the specific awarded programs, we're 70% of the way there. And while we're going to win new business between now and that time frame, I don't need any new customers to get there. If I just take the programs we're actively talking to our existing customers with. We're going to easily exceed that 1 million. And so the key to get there is going to be to continue to ramp up with Volvo and Polestar, win some incremental business there -- launch this massive Mercedes win that we have and then really move from the development stage with Nissan to the series production stage. Now I say that's going to happen in '26, '27, why am I being so imprecise on the time? Well, as Austin said, there's about 20 vehicle lines in there, and there's 20 SOPs. And while Luminar can control us being ready, there's a lot of other things that could cause some uncertainty in the timing. Remember what converts our order book from potential revenue to actual revenue is getting to SOP. And so this isn't easy building this stuff. Let alone the LiDAR, let alone the other systems on the car to do that. And so I don't want to be too precise in the timing, but give you guys a general sense of when we're going to get there, but we're absolutely confident we're going to get there. Now let's go to 2030. 2030 we expect to be north of 5 million units. Just with our existing customers, the awarded business plus the stuff we're talking to them about probability weighted, we're about 75% of the way there. I'm very confident we're going to win additional customers to get us to that number, if not higher. Growth isn't going to be done in 2030. That 5 million units is about 5% of the global vehicle build. When you step back and look at that, it's not that high of a number. I think aside put in their IPO document that they expect close to 50 million vehicles sold in 2030 with a LiDAR on it. I'd be very disappointed if we only had a 10% market share in 2030. We expect our order book by 2030 to give you a sense of how much growth is left to be close to $60 billion. All right. Let's look at the unit economics road map. So first, that revenue per Luminar equipped vehicle, 2025 target, okay. So this is when the initial launch of Volvo should be fully up and running, Polestar there as well. We expect that average revenue per sensor sold to be about $1,000 and it's going to consist primarily of that hardware and base software ASP. As we get to 2030, this is when we want to start to capture that additional value in the ecosystem, right? So I still think we're going to be at roughly that $1,000 revenue per sensor, take rate adjusted, the actual content per vehicle is much higher there. But what we're going to try to do is capture those software and solutions, monetize it and use that to bring the cost of our LiDAR down to increase penetration. Now let's look at the unit contribution cost. Our 2025 target, which is primarily the Iris family, we expect to be about $650 that's BOM plus manufacturing costs plus other COGS. By the time you get to 2030, we're going to be in our next-gen sensor and it's going to primarily be the BOM and the manufacturing cost for that being sold, plus a little bit of cost of service for the software and solutions, and we expect the BOM to be in that $350 range. All right. Let's look at the OpEx and CapEx. All right. We ended 2022 with a non-GAAP OpEx of about $221 million. A lot of the infrastructure we need to put in place to grow our business is there today. And in fact, there's a lot of, what I would say, launch costs and other stuff associated in that number now. We expect probably another 20% growth in that number in 2023 as we start to bring this stuff online. And then we expect that number to tail off a little bit in 2024 and be roughly flat and then from there, grow roughly 10% to 20% per annum. CapEx, we spend about $50 million last year. We're bringing up this new Mexico line, investing in the automation equipment, building it out. This year, we expect that number to be cut in about half as we kind of finally bring it up and do the remaining investments necessary to do that. Next year is when you're going to see the lion's share of the investment necessary for this new Asian manufacturing facility. We learned a lot doing our first plant. We think that there's going to be a lot of improvement for the second one. And so we expect the necessary CapEx investment to be in the $15 million, $20 million range and that's primarily going to be in 2024. After that, we're going to learn even more. We're continuing to design our product not only for capability, but for manufacturability. And a rough thumb is each million unit of incremental capacity we need to add would be about an incremental $10 million of CapEx. All right. Profitability road map. This year, by the end of this year, gross margin profitable. By the end of next year, core business breakeven, what do I mean by that? That's our core LiDAR and components business. If I wasn't making any investments in software insurance solutions, et cetera, we believe that we can be profitable by then. We don't think that's the right thing to do. We're focused on longer-term shareholder value creation. And as a result, we're going to get to profitability by the end of 2025. If you look at 2030, take the unit economics that I talked about on the previous page, the fact that a lot of our OpEx and CapEx is in there, we expect to have operating margins in the 35% to 40% as we get to the end of this decade. All right. I want to spend a couple of moments talking about our M&A strategy. We talked about the Seagate LiDAR acquisition we did earlier this year. We did Freedom Photonics last year as well as Civil Maps and Optogration before. We want to be positioned in this environment because we believe we're not only going to be a survivor, but a thriver, and we want to be able to act opportunistically. There are smaller companies that are struggling with their balance sheets and there are big corporations who no longer wanted to invest in the autonomous landscape. And so our phone is ringing a lot with the opportunities. We're going to say no to the vast majority of opportunities that present us, but there are going to be some interesting things that come through. Anything we do is going to be small to midsize and a strong strategic fit. What do I mean by a strong strategic fit. It helps build out the software and solutions ecosystem I talked about, anything that helps us vertically integrate in the core components, anything that will accelerate our R&D efforts, acqui-hires of great engineering teams. If you actually look at the M&A deals we've done in the last 2 years, it looks a lot like that. One of the things we're going to do is, while we have enough cash to get to breakeven plus a cushion, I want to protect that cash. I want our stock to be the primary acquisition currency of choice. And one of the things you're going to see us do here in the next few days is file some equity registration statements that will allow us to move quickly for those compelling opportunities that we see. These are going to be modest in size and we'll use them if we see a compelling M&A opportunity. And if we don't, we won't use them. All right. I'm going to leave you with 4 concluding remarks. First, we have the customer wins in place today to exponentially scale our sensor and revenue growth. I walked you through the credible path to get to 100,000, 1 million and 5 million unit run rate and we expect to at least double our revenue each year for the next several years on our journey there. We have a credible path to profitability, gross margin by the end of this year, core business by the end of next year, company by 2025. At the same time, we're going to be investing in this product road map to drive growth, allow us to capture this value in the ecosystem that our LiDAR creates and achieve superior margins. And then finally, you met the great team in place here that Austin built today. This is a team that can get this done. There's an even better team, no offense to the folks in this room, below them. We -- as I mentioned before, we have the cash on hand plus a cushion to get us to the profitability and make these investments. So with that, I'm going to wrap up this presentation. I'm going to invite Trey up here to help MC the Q&A. And Austin, if I could ask you to join me to help me with the Q&A, that would be great as well.

All right. Thanks, everybody, for joining us for what's been a wonderful day. I want to remind everybody who's watching via live stream on the webcast. You can e-mail us at investors.luminartech.com, if you have questions. And then here in the room, we've got a couple of mic runners, Kara and Tushar here. So if you can just raise your hand, they'll find you, and we'll get the questions that way. Right up here. Yes, let's go ahead Emmanuel.

Emmanuel Rosner from Deutsche Bank. First question maybe on some of these content per vehicle sort of opportunity and when I'm looking at end-of-decade target. So you have an essentially stable around sort of like $1000 opportunity sort of like mix adjusted, take-rate adjusted. Can you maybe speak a little bit about what goes inside as to sort of be able to maintain it at such a high level as your software content going up. I guess, can you just maybe give a little bit of a breakdown on how you would basically maintain such a high revenue per vehicle per sensor despite obviously maybe competition and other dynamics?

Yes. Sure. So here's what's going to happen with that Emmanuel. So as we start to commercialize our software Sentinel suite, there's going to be additional software products that we can sell the perception and the full stack capabilities for the proactive safety and the highway autonomy product that we're offering. And so the content per vehicle on those can raise anywhere from like call it mid-to-high tens of dollars for something like perception to a lot higher for the proactive safety and the highway autonomy. We're also having conversations with our customers to be creative in terms of how we price those. Instead of doing it just upfront, which is a traditional OEM model, you can do subscription base, which is where a lot of our customers are wanting to do. But we're also willing, as you see with the insurance, to bet on ourselves. And so if we want to -- if you want to upgrade the consumer to the highway autonomy or the next-generation ADAS safety, you need the hardware on the vehicle when it's sold in the standardization. And so -- and then that can create a very profitable upgrade revenue stream for the OEM as the consumer either upgrades that at the time of the purchase or any time over there via the over-the-air update. And so could we, at some point in the future move where maybe we take a little bit less on the LiDAR upfront. But then a little bit more of that upgrade fee as well and kind of partner with our OEM customers and once again, bet on ourselves and what our technology can do.

And I think this is actually a really, really interesting one and something that's overlooked -- actually a really good question in terms of total content per vehicle. And I think it actually would be interesting, we can even -- and maybe at the next earnings or something you do a more detailed breakdown because what I think is relevant, there's multiple different factors that affect these cars that basically involve very counterintuitive dynamics like when it comes to the economics. Like, for example, the whole point is the least amount of value, you will get on a Luminar equipped vehicle will be on the first day. The most amount of value is at the end state. It can already improve vehicle safety significantly off the bat, but then only continues to improve that over time as the software develops, as the systems develop, as the systems are deployed, and so the value of that system on the vehicle ultimately continues to accelerate via over-the-air updates, which generally I mean, I don't know if there's a single model that we're on that isn't equipped with that capability. So that's relevant. The other part that's interesting that I think is just when it comes to these other systems and services, it's really -- like we -- I think we were very, very conservative when it comes to modeling these things because we don't need any of this ultimately to be incredibly success. Like even with just the LiDAR, like we can be incredibly successful. But you're probably not going to build a multi-hundred billion dollar company doing just that. So that's where -- at the end of the day. So that's where I think like expanding with that ecosystem makes sense. Like for example, just the insurance alone from what I said in my presentation, that's like people spend on average in the U.S., $20,000 per vehicle over its lifespan there, too. That's like I mean, 20x the content value of the LiDAR. If we did actually save a material portion of that by safety improvements and other stuff on that. But the point is that then that comes into the take rate. So what's the take rate on that. So that's where you can end up with a dramatic variance into the thousands. And the same thing for like if you have automakers that are charging $5,000 for highway autonomy system, it's the same kind of discussion there. So all of that is accretive. And I think ultimately, we'll transform from something that is like a onetime $6,000 thing into x dollars on an annualized basis in a subscription model. Like I would be surprised if by the end of this decade, that whole software and services capability and division into products that we're building aren't on subscription as opposed to like one-time upfront sales. And think about it, like, for example, insurance already is the ultimate subscription, when you think about it. So -- that's just a different way to look at it.

Itay?

Great. Itay Michaeli from Citi. Thanks again for hosting the event. Also, you answered just before around the software opportunities for LiDAR to get better. Maybe talk about your customers' appetite to -- you've already seen to standardize your LiDAR initially, maybe out of the 20 programs or production models you're on maybe what portion do you think what will be standard and how long? And maybe Tom, on the 2030 projections, what are you assuming in terms of the success of the insurance in terms of [ center]. I know it's early days, but in your model, kind of what are you assuming in terms of where that is in 2030?

Do you want me to go first because -- when you look at the insurance, we're assuming very, very, very low initial penetration rates. I've seen enough from our conversations with multiple customers to make the decision to invest in this now. The amount of investment here isn't as big as you would think. As I said, Alex, walked me through how we built it to Tesla, we're talking tens of millions of dollars, not hundreds of millions of dollars to get this out and build at least a product ready to go for that. What's in there now is very, very, very low, I would say, single-digit penetration until we prove this out. We have a good sense of like what the penetration rate is for Tesla, where the OEMs there. But I got to see some real-life data before this really starts to reduce the model. But there's a lot of potential there. But it goes back to your point, and I don't want to steal your thunder on the standardization, but if you're able to partner with the OEMs on this and either -- and underwrite those insurance savings together because once again, those insurance savings are going to be there. And we go to our OEMs and say, if we don't do anything, then the traditional insurance companies are going to capture that value, so let's go capture this value together. We build out this insurance infrastructure. It's scalable. You don't do the work. We did it, and let's go collect these checks together. Like that's our pitch. And takes time for that to resonate, but that has advanced to the stage where we're willing to make that investment.

And I think that's absolutely 100% the right approach. But I think having the right very conservative attitude on these things to do is super important. Like, for example, even in the revenue build, when you take a look at like for 2030, it doesn't have any like negligible stuff for insurance and things of the like because listen, the reality is that if we build a business that can get to $5 million or $5 billion run rate there, too, at that period of time with the $60 billion forward-looking order book, I think quarter like -- that's already such dramatic growth where it puts you more into the category of like beyond what the Mobileye of today is or even what NVIDIA was a handful of years ago. So like we don't need to promise that to make it super successful, but it doesn't mean that we're not investing -- that's actually going to help accelerate the rest of that standardization vision, as Tom was mentioning, because it basically -- any it will help speed up the -- like there's already a business case today to have a LiDAR on every vehicle there too independent of even if there was no safety savings at all. But once you add that into the equation and factor it in, it really gives -- there is no reason why every automaker shouldn't have this product on every car that's produced. It just -- it makes -- it rounds out the math to make it incredibly compelling. And even that 5 million vehicles, that's only a 5% market penetration, right, by that time. So there's a lot left to be able to cover and a lot of upside with the kinds of ASPs and hopefully, content value should substantially accelerate then beyond that.

Thanks for the question, Itay. Let me go to one that we got from the live stream, and then we'll go back to the room. So I think this one is going to be for you, Austin. Can you share more details around the recent Mercedes announcement? What's the primary differentiation that drove the expansion of the win without any on-road experience validation of the technology with another series production launch?

Yes. No, it's a good question. I mean listen, these are very, very big bets. And I think probably as you guys saw today, it's extremely uncommon, maybe even unheard of to have a like automotive supplier like at this stage, like be able to get the level of focus and attention, everything from the heads of these companies that are employing hundreds of thousands of people with millions of things going on that probably have a lot of stuff to do during to ensure a successful execution of the business holistically. But the reason why is because I think the leadership of these companies strongly believe this is core to the overall road map and strategy of the automotive industry and their businesses generally going forward. And I think we are seeing adoption of this kind of technology at an unprecedented breakneck rate relative to what we see for other kinds of technologies historically in the automotive industry, where normally it would -- by all means, you would wait through successful execution of this. But that said, there is more behind the scenes. We absolutely have to prove ourselves at every step of the way. Like we're literally managed like day-to-day, week-to-week schedules for every automotive program. Like it is a very, very intensive execution, and that's what we're constantly heads down doing here. So that as we successfully proved out that we could execute to that stage and met the first mile -- and met all the milestones for the program ultimately to get that confidence to be able to scale successfully. So that's what enabled that major decision to be made. And then obviously, I mean it goes without saying from the other step, but the technology sort of speaks for itself. I mean it's differentiated. There's nothing else that could enable that capability with them. In this case, it's not even just a matter of like, "Oh, is it Luminar or someone else?" It's like, is it Luminar or are we going to wait x number of years to put it on the vehicles when it's derisked because I mean these guys are basically betting the futures of their company on us or -- and if we don't ship, their car is going to have a hole in the roof. So yes.

Thanks, Austin.

Yes, let's come back here to Jason.

Yes. This is a personal investor. I had a question about the business in Japan. Are you guys seeing an uptick in interest from OEMs? And what year will the Nissan deal possibly bear fruits?

Yes. So the answer to your first part of the question, Jason, is absolutely I think everybody noticed what Nissan did. And to remind everyone, Nissan came out last year and said they're developing this next-generation safety system where our LiDAR is going to be a key part of it. You saw some of the commercials and videos that they've released about that. And we're working actively with them on that development stage. A lot of the other Japanese OEMs noticed. And as you can say, that has resulted in activity. And we'll talk about that when there's something to talk about. Nissan has said publicly that they plan to start deploying that technology on their vehicles starting in the middle part of this decade. And by the end of this decade, 2030, have it on virtually every vehicle that they make. When you go back to that million unit year run rate that we talked about in '26 and '27, that first phase of Nissan, we assume happens roughly in that time frame. But it's actually -- what we're assuming is a very small percentage of it. Just to highlight, we talked about the conservatism in our order book. We're in the development stage now with Nissan. We've got to get the system to work with Nissan, and then they're going to start deploying it on their production vehicles. And so we're working with them in terms of what that rollout schedule is going to be and then getting to the formal nomination process. But right now, in my order book, I have 0 in there for Nissan, right? And if we're successful developing it and rolling it out and we get to that virtually every vehicle they make in 2030, Nissan makes 4 million vehicles a year, right? You saw that 5 million unit number I put up there when 3/4 of that is kind of from our existing customers. I have nowhere near that full 4 million in our model as of now. There's some but nowhere near that 4 million. And so we're very excited about that. That is all progressing on the right thing. And when we get formal vehicle line awards, then we'll start moving that into our order book.

One more question. Go ahead.

This is Joshua Buchalter from Cowen. Very informative day. I think it was clear throughout you guys are working on a lot. How should we rank order, in particular, the opportunities within that other non-core LiDAR hardware software bucket, which ones are you most excited about? What are the -- and how much are those contributing to the forward-looking order book and your future projections.

Great question. So when I talked about our public milestones for this year, I put top 3 up there for a reason, and that's really execute, execute, execute, grow our business and continue to invest in this product road map. In order for any of this to be successful, we got to make these LiDARs. We got to make these LiDARs in scales, and we got to get them on a lot of vehicles out there. And that's the top focus for this company this year and what everybody is focusing on. We're going to continue to make investments in the other stuff, the software and solution. Right now, remember, what's in my order book is only awarded business, the business we have awarded today is primarily that hardware plus the base software. So it's a rounding error for that other stuff in terms of what's in our order book. We're very confident that, that's ultimately going to start to result in real scalable revenue in the second half of this decade, and that's why we're making those investments. But there's a lot of upside in that order book and even what's in our longer term, midterm and longer-term financial projections. Yes.

And I think that's the important part is that no matter what, like we have the fundamental must-dos, cannot fail and what we need to execute on across the board. And that in and of itself can and will build a massively valuable business when it comes down to it. Like I think the other thing just is upside. And what we fortunately are in a position of luxury where we have the ability to make those investments and get to that upside and still have the business, the cash liquidity, the revenue and everything else to support you know what we need to do initially. And I mean, you take a look, it's like the perfect example of this, Mobileye is actually one that just recently went public again here, right, is that I think when they had, what, $350 million revenue, they were on the order of like $10 billion, ended up getting bought up for $15 billion. Now what -- it's -- they have over $1 billion, and it's like what, $30 billion plus as a company in terms of value, like that goes to show -- even like without all these autonomous too, even without like just being able to have like a great fundamental business with that core growth, underlying like the equivalent of what we're doing for LiDAR side is great. Obviously, the distinction here and I think what's relevant is that we are -- even from a LIDAR level, we're providing systems with huge content value on a vehicle, like it's basically our market opportunity is probably like what the fundamentals of like a basic ADAS system is or a chip for that. And I think that's why everybody is trying to go after that ultimately for that additional upside. But I think that's the way to characterize it.

Thanks so much. And just a couple of announcements before we kind of do a final video. I want to thank everybody who joined on the live stream. Thanks for spending the afternoon or morning with us. We really appreciate it. We're super passionate about the business. Hope you are too. For the folks who joined us in Orlando, huge thanks. It's not over yet. Now we're going to do all those great things that you heard us talk about, we're going to go let you experience at our long-range test facility where we're going to have demonstrations. And we've got our Head of software out there, CJ. Moore and his team who've been doing tons of work to get ready for that. So I look forward to you being able to meet him out there. The buses are going to be directly behind you out these doors this way. And then we'll go out there for some demonstrations and some adult beverages. With the live stream, we're going to have 1 last video, and this will be the end of the show. So thanks, everyone.

Thank you.