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

Great. Good afternoon, everybody, and thank you for joining us for our next session. I'm very excited to host Luminar, again, at our Global Tech Conference. I'm Itay Michaeli, Citi's U.S. auto mobility analyst, I'll be your host this afternoon. I want to keep it to a pretty informal fireside chat session with management. I've got a bunch of questions. A lot happening, of course, in the lidar ADAS autonomy space. But most importantly want to make sure that your questions are answered. So please feel free to e-mail me questions directly, and I'll definitely lead them into the conversation. Of course, if you need of our legal disclosures, we can make those available to you as well. So from Luminar, we are delighted to have Austin Russell, company's Founder and CEO, calling in from Germany after the IAA conference, so I'm sure we'll talk about the latest there; as well as of course, Tom Fennimore, the company's CFO. Austin, Tom, great to see you. Thank you both for being here, and welcome.

Thanks for having us.

Yes. No, it's been a definitely a great, great show here. And no shortage of things going on, but always -- I know you always ask great questions and stuff. So look forward.

Absolutely. Let's get right into it. So I thought maybe we can kick off the first part of the session just for investors who might be newer to both lidar as well as Luminar. So we'll kind of just start with some of the basics, and we'll kind of get into some of the other key debates out there. So maybe just -- again, at a high level, we can just start with thinking about what lidar brings to the table. And I think we like to think about the approach in sort of two buckets. One is safety and the other is autonomous. Just love to get a sense from you kind of how you see the value proposition on both for just the lidar industry in general?

Yes, yes. So I would say there's definitely a couple of different dimensions to this. As you mentioned, there's the safety side and then the autonomous side. I think up until just very recently, everyone was thinking really to be all in on -- there's only autonomous application on this time of technology and primarily focused on urban environments. I think what we were able to prove out by having kind of first of its kind deals with automakers to see this in the series production for an autonomous use case is to take that highway focus for autonomy to have a much more solvable problem in a more constrained environment and a more realistic time frame. And the other side of it is the safety side. And that's something that we're also showing off here at IAA. Just an opportunity to be able to utilize the same hardware that's capable of in front of me, but now we're actually just moving towards the vision of creating the uncrashable car of something that they can actually be able to take safety improvements on vehicles completely to the next level because they're actually surprisingly bad to that. I mean, you would think that programming your vehicles to just not run into the thing right in front of you is actually straightforward, turns out that is actually really hard. And that's why people get in accidents all the time, 1.5 million lives are lost on the road every year, 15 million serious injuries. This is a real problem. And there is a direct opportunity to do that where it also doesn't have to be a fully driverless robo-taxi everywhere all the time to be able to solve that problem. We need to start with the production vehicles in the existing $4 trillion a year industry that's out there. And that's what we're working on. And we're focused on passenger vehicles, we're focused on trucks, and we do work with robo-taxis as well. But it's good. So just for reference, lidar sensor for light detection and ranging. So you just kind of go into the fundamentals here. It's basically the eyes of the autonomous car that allows it to see and understand everything around it. We're sending out millions of these laser pulses measuring the exact distance to objects down a centimeter level precision to know exactly where they are. And this is the enabling system that allows the car to know where everything is, such that, it can safely navigate and ultimately come to a safe top when it needs to. So that's a really important part of this. And from a value proposition standpoint, we developed fundamentals and really the first and only lidar sensing system that can meet the very stringent requirements from a performance standpoint as well as a scalability and economic standpoint, actually getting into series production to enable this level of capability. But at the same time, we've also developed the software stack that helps better enable these systems because really hardware alone by itself doesn't really let you do anything, you have to interpret the data, you have to make sense of the data, you have to act on the data, and that's what's really important. And that's where having a tightly integrated hardware software solution is incredibly valuable, particularly in these kinds of systems. So working with OEMs and that has been great. Part of our -- again, our claim to theme is from a performance and technology perspective. But also from a commercial market perspective, we've been working -- we're already now with 8 of the top 10 largest automakers to be able to see this technology through, many of which have already announced series of production deals with us to be able to enable this as an option on their vehicles. And now most recently, we had a breakthrough announcement with Volvo where they're not actually going to be standardizing this on their new vehicles, starting with the XC90 platform that they have. So that's something that is going to be certainly a great, bright future ahead. Why is it starting to be implemented on vehicles, not just as an option but also as a standard? It all comes down to safety, it comes down to the fundamentals of what you can do and this is really the biggest step function in the safety capability that we've seen in a decade since the advent of seatbelts and airbags and kind of those basic capabilities. So, I see, that's the quick summary. A lot more going on, a lot more to unpack, and a lot of crazy stuff and good stuff here at IAA. But that's the overview. Hopefully, those could -- are kind of what you're looking for, Itay.

Yes, it's great. Thank you, Austin. Yes. When we think about sticking on the topic of safety, if we just look at the crash logistics and they seem to be going directionally in the right direction, but they're still very, very high. And one could say that even with ADAS penetration now being pretty meaningful that it may be -- by now, we would have expected to see maybe even better stats. And of course, that brings in lidar and the discussion, particularly with being able to detect not even -- under more adverse conditions. But one question I get from investors on the safety side for now is, is there a way to quantify kind of from a safety performance perspective, what adding a lidar, let's say, camera radar configuration incrementally provides in terms of incremental safety?

Yes, absolutely. And I think the most significant thing is just -- one of the things that we were actually showing here at IAA is a benchmark system against like a modern vehicle, in this case, we had like a modern Tesla here as a computer vision example of about what the state-of-the-art is. And what we did is we have a static pedestrian scenario of where you had a child dummy that was out there. You may be familiar with these kinds of dummies from like different end cap and safety tests and everything in terms of how they do collision testing. It turns out even in the most basic scenario of all, like you put a child right in front of the path in a straight line with the Tesla just like going straight at it, the car will not come to a stop in time to actually prevent a collision. Why is that? Okay. Well, you would think that like, okay, that's the most basic detection modality and mechanism. And how do you change that? Well, why do you not do it? It's not because you don't see it. And actually, most -- it's really easy to see things and detect things like this. The golden question is, how do you have enough confidence in that detection to be able to really understand that there is something there when you have a bunch of false positives, false negatives, all this other stuff in the environment, you never really know what's there and what's not. And as a result, you will not generally take controls over from the human until the absolute last second and by then it's too late. So on the other hand, we have our system that we also test along the same track in 3 different scenarios. We had one where it was just a static pedestrian which is able to go and come to a safe stop where you can actually stop to give you about a meter worth of room left over to in that situation and showing that it is possible to come to a safe stop when you have high competence detections. The next scenario is we showed actually a kid running out, if -- I'm trying to show the -- from the perspective of the car, you have a parked car here and the kid running out or walking out behind it or it's just a last second view. Again, some systems were successful at slowing down before the collision. In the other example, none of the systems were successful in that scenario. And that's where, again, the lidar equipment system and our software were able to come to a safe stop. The last scenario we had is where actually you had a kid walking behind the car, but they stopped and were in the collision path of trajectory. And as a result, the car shouldn't go on fully slam on the brakes, because that can also be a situation that is own right if you're slamming on the brakes all the time for no reason. So as a result, we were able to show that in without the car stopping. So that was one thing we were doing in private demos here at IAA, but we're actually going to be showing this off to the public at CES and potentially even letting individuals drive the cars themselves to really experience what it's like to be in the middle of these kinds of safety tests. And the shocking nature of where we are today and where we need to go. Again, there's -- despite all -- and that's obviously, there haven't been advancements in safety, there have been, but they've been relatively incremental. I mean, the thing is that with this, you're talking about I mean the way that Volvo described it is eradication of accidents. Obviously, that will take some time. But the theoretical potential from even our models is about up to 7x increase in safety. This is like a 2%, 5%, even 10% increase, this is 7x. So that's the opportunity that we have, and that's what we're going to make happen.

And so Itay, you said that the vehicle test and the vehicle safety have actually improved. And if you actually look, and this is surprising to me, since the start of the pandemic, it's actually gotten worse, not only on a per miles driven number, but actually on an absolute basis as well. And so NHTSA had just released data for the first half this year. And in that data, there's already been an increase of 7%, and that's a double-digit percentage increase last year. And if you actually do it on miles driven because miles driven has gone down because of pandemic, it's actually a lot higher. And so I haven't heard a good reason yet for what that dynamic is. I mean, we've had the distracted driving around there. It's like we're driving less. So we're losing some of our skills to -- there's less traffic out there, so people are being aggressive drivers. But while there's been a lot of technology to make the systems more safer, they really haven't worked that well over the last couple of years.

Yes. I'll turn it quickly. We have the last couple of updates here.

It's what stuns everybody.

Yes. Absolutely, absolutely. That's a great point. And honestly, there's a lot of good reason to improve safety on the vehicle. But for short regulations, right, it's always been -- I'll talk about Volvo of course, the standardization announcement a bit later, but most of it has been sort of a take rate sale historically in autos where you sort of get only a portion of the volume attached with the more advanced systems. But two things that like seem to be changing that are number one, the potential, and you mentioned also 7x improvement potentially in safety to significantly reduce insurance costs and it doesn't take a whole lot to sort of argue you can pay for the entire system as well, including some of the other sensors in compute. And the other kind of thing that's sort of changing is that all these connected electrical architectures are allowing your customers to sell these features as over-the-air update, as subscriptions and now have a lifetime of shots on goal from revenue generation. So my question is, is that real? Austin, you just spent a lot of time with customers at IAA. When you talk to them, are they excited about these types of opportunities, insurance and kind of lifetime revenue? Just talk us a little what's actually happening in the ground when you talk to your customers in light of these opportunities?

Yes, yes. No, that's great. And then we've had the opportunity to be able to take some of the various CEOs or the OEMs out here all on these exact demos. They too have been super impressed and blown away by what we've been able to achieve so far across each of these different iterations of samples of our product as we've gone through over the past few years and just perfecting, industrializing and getting to the stage. But I think what's also interesting is just -- I think was surprising more than anything is just -- it's always funny like you talked to like the heads of various automakers. It's always their own vehicles and they're like, well, Mike, my car, like of course, you wouldn't do that. And then you go and test it and it was like, well, it's not actually that different than the standard. The bar is just super low right now for safety. And yes, these cars pass. Yes, some of them even get 5-star crash safety ratings. Does that mean that it's not going -- like it's not still easy to get into an accident? No. It still is. And that's the shocking part more than anything, I think, that's come about. So -- but when it comes down to it, I think there's really four factors that are influencing the course of ultimately standardization of this technology over time and particularly the safe factor. So one is consumer demand. Obviously, there's a huge consumer push for increased safety on vehicles. It's one of the top things that continue to look for in vehicles. And that's even for just incremental improvements. The reality is that it will be either plus or minus, they're not that dramatically different in terms of it. I mean, maybe on the high -- the super high end or the super low end. But on balance, there hasn't been a step function and valuing that step function is key. For something like Volvo, I mean, they built our entire brand off of making those kinds of innovations and then standardizing it throughout the industry previously. The other one is just the OEM [indiscernible] and drive. OEMs are obviously always pushing for safer vehicles by all means. That's part of the consumer value proposition. Three is the insurance side of it, there's definitely a strong insurance play, part of which we've been involved with. We previously announced a partnership with Swiss Re, which is a largest reinsurer in the world to be able to enable this and things are only accelerating when it comes to these things. And there's absolutely strong interest because when you take a look at the total cost of ownership, if you do this right with the right hardware and the software, you are going to effectively have these systems pay for themselves. And not only that, an additional revenue stream to us and an additional revenue stream to the OEM. And then lastly, the last component is regulatory. The reality is there is a strong driver that's taking place for this. Ultimately, you have things like -- we mentioned seatbelts and airbags and those other things before they have been effectively mandated across not just the U.S. or Europe but internationally at this stage. And at the same time, when it comes to other kinds of advanced safety systems, if you want to be able to get a decent crash safety rating or even just have been known as a safe brand, you want to be a leader in that space as a tech leader and as a safety leader. So I'd say all those four drivers are what's enabling this to converge. I think a lot of people have thought that maybe it will be like 10 to 20 years by the time we see this. Well, I think some of the original predictions were, by the time you see this kind of tech in vehicles, it will be on robo-taxis on the road in like 2 or 3 years. And then by the time it's on consumer vehicles and standardize it will be 20 years. It turns out the equation has completely flipped where we're actually already starting to see standardization of this technology on vehicles. In fact, just one year from when Volvo originally laid that stake in the ground and saying, "We're going to put this on our vehicles as an option as the first of its kind to enable on consumer vehicle highway autonomy capabilities." Just a year later, it's now already starting to be standardized. That differential is pretty mind-blowing when you think about it. So I'd say that's the disruption that we're already seeing. And this trend is only going to continue for sure. But it all comes down to having the right capabilities. I mean, like I said, we've been fortunate enough to pioneer this technology, the system and everything that's gone into it to really have this very unique value proposition. And just -- I mean, I don't think we could possibly be in a better position to realize the rest of this growth curve of this industry. So it's super exciting to see everyone on board with it at IAA.

Awesome. That's great to hear. Yes, I promised myself, I wouldn't spend too much time on lidar architecture because I think -- in the past, but I do have many questions from investors around, why is it that there isn't sort of a consensus among the lidar players around sort of a right architecture? Everybody seems to have their own view of the lidar approach. Maybe I'll talk a little bit about that. And of course, your differentiation and kind of how you see the architecture debate evolving over the next 5 years? How important should it be for investors in terms of the consideration of investing in that?

Yes. So I think there's one simple answer to that at the end of the day. These companies wouldn't exist if they were to claim that there was one right architecture. And that's the hard part. People build their names, their brands around a certain theoretical technology approach in terms of what they wanted to do that spun out of some lab or whatever it was originally. And I mean, I think all these approaches are interesting. There's 2,000 different ways that you can go and configure a lidar of different combinations of permutations the way that you can assemble the components. The challenge is that for all of those except one small branch of those approaches, we're not aware of any way that can actually meet all of the specifications while also creating much less a scalable product, auto grade device, et cetera. So I think when it comes down to it, a part of it is the problem of -- from a technical analysis perspective, there's a lot of interesting theoretics around this stuff. But the question is that how do you have all of these performance parameters and achieve them simultaneously. That's the really hard part. And ultimately, OEMs are happy to you to be the judge of this. And then the other -- but the most important question most of all is what problem are you trying to solve? What are you trying to achieve at the end of the day? It's easy to build a lidar and be able to build it cheaply and I mean there's been $50 golf rangefinder lidars that you can use -- that you can put into series production, but that's not going to be super useful because it's a single point. So the question is that what is it being used for? And if you want to achieve a new level of autonomy, if you want to actually get the driver out of the loop and you want to be able to make a dramatically safer vehicle, you need a completely different level of performance. And that's the most significant thing that a lot of times, I think people -- there was this thought initially, maybe like 5 years ago or so, that do you really need a huge level of performance because autonomy is going to be an easy problem, everyone is going to have it solved and we're going to welcome our robot overloads a couple of years from now. Obviously, that didn't happen. We made a call that, that was going to happen and that you were going to need this extreme level of performance to get a ground review of what's going on. And maybe surprisingly or unsurprisingly or in recent history, everyone has converged on the spec. So that's part of the significance in terms of what it is. Now again, that's not to say that you couldn't endure or use another -- like another lidar architecture under a car. So just the golden question is, what will it do? You need a certain level of capability to achieve autonomy and ensure safety. Otherwise, some people are thinking about, okay, well, maybe you can make a better L2 system, you can improve on that, you can improve on [indiscernible] system improvement and it's true, you can do those things. But I mean, listen, this is even coming from a lidar guy. Like I'm not sure you really even need the cost of the lidar to be able to do that. Like, the reality is that folks like Mobileye have been in the space like super successfully for the better part of a decade-plus now. So we're not here to try and compete with our business. We're not here to try and compete with that. I think that's a dead end. I think you need a step function capability. But speaking of folks like Mobileye, I mean, that's also where you have folks like that going all in on our system of the actually -- the CEO of Intel, Gelsinger, just build our new -- their new car powered by Luminar as well upon stage at IAA, which was awesome to see. So we also had a couple of other partnerships, Webasto Inalfa Roof integrators that we had that were showcasing Luminar Tech on their vehicles. But the Mobileye one was really interesting. But that goes to show how you can have complementary technologies. You're not trying to compete with one another on this. But now that's not to say that people are not going to try and compete to us. So I think that's absolutely the case on a number of fronts. But like I said, we're not aware of any other architectures that would be a little easier to do so.

Awesome. And that's super helpful. Thanks for lot of details. And definitely -- I'll circle back a little bit to that as well. Maybe one more kind of overview question. I definitely want to dig into Volvo because that was a very significant news not too long ago. I think you've shown a $150 billion TAM in 2030. Just curious if you could just give investors a bit of a high-level assumptions behind that TAM. And maybe how -- roughly how many lidar companies do you think will attain that TAM in 2030?

Sure. So I'll try to make the math relatively simple. So the way that we look at it, and we're focusing primarily on the passenger vehicle market and the commercial truck market. And then we have 2 primary product offerings, which we call our proactive safety, which we spend a lot of time talking about, which we just unveiled and then we have the highway autonomy. Volvo, the first customer that kind of standardize it on one of their vehicle programs, not going to be the last. We think, as you mentioned before, Itay, that there -- the vehicle safety improvements and the resulting insurance savings will effectively pay for the cost of that technology to put proactive safety on there. And once you get the cost right and once it starts working its way through the system, we believe there ultimately will be regulatory pool or push to have that become standardized. And so we look at it as there's 100 million cars on the road we think globally that are sold each year. We think the value proposition is there to ultimately put this on every vehicle. And the hardware and software, you add those two together to kind of a solution there, and it's about $1,000 content per vehicle, roughly equally split between the hardware and the software. So that gets you to $100 billion. That doesn't include the highway autonomy. We spoke in the past that the content per vehicle for that is about $2,500 per vehicle. So if you kind of take the incremental debt, figure a 10% to 20% penetration rate in 2030, that's probably another $20 billion or so. And then when you look at the commercial truck market, there's about 5 million commercial trucks sold annually. The value proposition once again for both proactive safety and highway autonomy is very powerful in that segment. And then more importantly, for the commercial truck and if you look actually on my background there, you can actually see that there's multiple sensors per vehicle there, call it, 3 or 4 in our conversations with our customers. So you add that up and you get to about another $30 billion. So when you think about the $150 billion in 2030, about $100 billion of that is proactive safety for the passenger vehicle side, about $20 billion of that is the highway autonomy upgrade and the incremental revenue from that, and then the $30 billion is from the commercial truck space.

Awesome. Super helpful. Thanks for that breakdown, Tom. So let's talk about Volvo, right? Because I think it's come up a couple of times in terms of technology paying for itself for this potential for a domino-effect from other automakers adopting standardized lidar systems or full AV systems. So maybe talk about what you're seeing thus far. It's now been, I think, a couple of months that announcement which was made public. Are you seeing any other automakers, maybe premium automakers begin to think about this and say, wow, if they're doing it, we need to also do it ourselves. When do you think this could maybe even domino into the mass market? Just talk about kind of how significant this potentially could be for lidar adoption going forward?

Yes. I think unquestionably, it had a substantial effect on this. And again, it's -- I think it of course, extends beyond the lidar, but you have to have the corresponding software to enable it. And by the way, that is part of the reason why our TAM is what multiple -- a multiple of what's in the other TAMs you might see for another -- someone that wants to be a lidar provider for these systems, just because you have that full stack and you have more like an AV type company for that approach. But when it comes to standardization for OEMs, there's no question the dialogue at the CEO and executive or Board level is completely changing. And it really -- I mean, that's where the strategy is set for a lot of these different types of opportunities. And I think being able to -- it's just as much about education as anything else. But I will say this is that I've heard the comment that people had to -- had their 15-year road map planned. And what happened with Volvo just totally disrupted that and threw it out the window in terms of how they're thinking about the business. So that's where there's absolutely an opportunity to be able to change the conversation. It's already happening with a lot of these guys. And in their words, it's more a matter of when than if, at this stage.

Awesome. Best way to hear it. And then maybe for Luminar specifically with the standardization, I imagine it should help reduce hardware costs, give you more software collection as well. Maybe talk a little bit about that, particularly in the software side since it is an important part of the future TAM. Maybe talk to us maybe at almost a technical level around how getting additional data helps you develop better perception software, higher confidence levels to the extent it does, of course, just maybe talk about both the hardware and software implications of kind of the standardization and just the acceleration that you're seeing?

Yes. Getting the data in series production is critical, and that's something, I think we're the only company to do so at this stage for these kind of systems. And that's really important to be able to actually train your software, train your systems. Of course, you have to be able to have the lidar data, that's the most important thing of all. People have already had treasury troves and camera data, and that's great and all. But it doesn't really get you super far. It kind of asymptotically trails off in terms of the usefulness of it after a certain period of -- certain point, a certain amount of data. So when someone says they have 1 million miles versus 1 billion miles versus whatever it is, it's not actually that dramatically different. The thing is that when it comes to these kinds of autonomous capabilities with lidar, it really does make a dramatic difference. And being able to see all these different -- you see all these different use cases of the lidar across the world, is everything. And that's where having these initial deployments with folks like Volvo, SAIC, Daimler Truck and other key players across the industry, it makes all the difference. So I would say -- yes, so on a software standpoint too, there's no question that's going to be an increasing part of our business. It's already a nontrivial part of our revenue streams today, but expand -- going to be expanding as things go from series development into series production. But that said, I think the long-term value of this is unquestionably going to be unlocked as you continue to get over-the-air updates to these vehicles that are already home equipped with Luminar and people are offered upgrade options as well where you can flip a software switch to be able to enable autonomous capabilities, for example. The other thing that we're really starting to see, and I think we're pioneering subscription models as well for these different approaches with automakers, I don't know if anyone has successfully done this to date, but this is where we're changing the conversation as well when it comes down to this for these kinds of engagement models and revenue models with the customer as they're also thinking about it from an end consumer standpoint in terms of how they really want to charge for these things. But there's a pretty strong business case that's to be had on all sides. We're not squeezing them to a point of there's nothing and vice-versa as well. It's just a question of how much value you can create. When you create a lot of value and a lot of recurring value, you can do a lot with that.

Absolutely. Also, how should we think about the company's like future product pipeline? When you talk to automakers, obviously, you're bringing Iris to production, a lot of focus on the current programs you're supporting. You mentioned you met the specs for Level 4 autonomy. But what are customers asking for now in terms of future lidar specs, whether it's range, frame rate resolution, costs? And kind of how should we think about the pipeline for Luminar in the next few years? Because I think every lidar company is talking about new generation of products and kind of new requirements. Curious what you're seeing and how you're kind of positioning for that?

Yes, yes. Honestly, the blunt answer is, they don't care. They ruthlessly don't care what you have to say about x number of years down the road. They're tired of hearing all the VS that they've been hearing for years about all these different possible approaches and what's going to happen on what time line. They want to know what do you have, what do you have now, what are you actually going to deliver at the end of the day here and what is it going to do. And I think people have made enough wrong decisions in this industry as well as they've had to back out of that it's -- not necessarily wrongful league in the sense that people are learning on these things in terms of what works, what doesn't, what it's really hard, what's not. But I would say no. Okay. That's the big picture. In an ideal world, what now -- there is one thing that obviously people will always try and push on and like, literally, automakers have whole apartments just dedicated to this, which is cost. It's not -- we've been able to get our cost to like a pretty decent level here. But it doesn't change the fact that -- when an OEM knows you're making a good margin on something, they don't like it. They like their suppliers making like 4% margin, not 40% margin, plus or whatever it may be. So that's the difference. Now that said, I think that there is a natural cost curve with that. But I don't actually think that the whole notion of having like a $100 lidar that you have to charge for it at one point -- like I think that there's just so much value to be had here that -- if you have the right level of capabilities on it, then the value creation is over an order of magnitude greater, there's really no reason why you have to do that. So I would say the biggest thing is just, frankly, delivering an auto grade product that actually meets all these facts that can solve the problem and it is like this is -- obviously, that's part of what we've been aligned with in doing with these automakers, but it's not done yet. There's still a year left before really it starts to launch into series production. So there's a lot of testing or validation that needs to be done. Over the long term, for this, I would say, from a performance level, they are pretty satisfied with this. Maybe the trucking guys wanted a little bit more range of why we did kind of a software unlock on that to be able to give the max range it sends out to 500, 600 meters from 250 for the car guys. Cost of focus. Integration is one of those things as well. How do you successfully integrate this into a vehicle. And that's why part of the point of why we launched with these partners, Webasto and Inalfa to be able to see the successful integration into the roof of -- right now, I mean, you control what I think is the majority of the market for the roof business that's flying to different OEMs. So now they can actually offer this as an option that's integrated into the roof line that just goes straight into the OEMs or the clean design. So there's some different aspect to that. In terms of what else is ahead, I think the game is going to be -- the biggest thing is going to be the economies of scale, improving yourself out, like again, less about the actual product itself, but more just proving that you can produce these and that you actually get to SOP and that you could make them and make lots of them. And then more than anything, I'd say, the thing is software. It's all about software. I mean -- and we started with the foundation from the hardware. There's a lot of work left to do on the software to ultimately get to where we want to go. And but that's also part of what over-the-year upgrades are for. But having that basic level of safety capability as well as enabling highway autonomy is going to be key. Yes. So I think that about covers it. From an architectural perspective, we're basically pushing the boundaries and the limits of physics for what's possible. So the thing is that you can always try and do more, but there's really a set of trade-offs. There's like a Moore's Law type asymtotic factor of performance improvement, you can have it, like yes, you can, but there is a cost to everything, either metaphorically or literally, it's just more cost. So I mean, we have the foundation to make any spec of a lidar system we want to at the lowest possible cost, but that doesn't mean that there isn't still some trade-off. So I think that's where we found the sweet spot. And I think there is going to be execution, delivery through a scale and just the software is going to catch up as well.

Itay, the one thing I would add to that is, when we think about what our product road map and what we want Luminar to look like 3, 5, 10 years in the future. We spend a lot of time on hardware and how can we extend our lead today, build out the competitive mode and kind of anticipate what the customer requirements are going to be. But where we spend more time thinking strategically what I would say is on the software and the services side. And so we don't want Luminar to just be the best lidar company out there. We want Luminar to be one of the leading autonomous solution providers there. And so what we can do on the software side, I mean, you've seen that some of what we released on proactive safety that we kind of developed with our software team, you saw what we did last year with the Samsung autonomous software team, the acquisition we did in Germany and this [indiscernible] critical to helping us develop the proactive safety solution. You saw what we did this year with partnership with Zenseact to get into the full stack space. We've talked about what we're doing with Swiss Re on the insurance side where we see an opportunity to be very creative strategically, how do we kind of take our leadership position in lidar and become the leader in the autonomous solution provider. And I think that that's where you're hopefully going to see some good progress from us there.

And that's a great point. I mean, just for the sake of clarity, we absolutely do have a 10-year road map laid out for the lidar that we do have a number of iterations when it comes to additional performance capabilities, and size cost, all that other stuff that goes along with it. But that's not going to be the gating factor whether this industry happens or not. As Tom was saying, it's going to be the software capability, how this happens after we already have the lidar proven out. But I think that's where we've proven ourselves as the only one that can actually build product for the serious production applications, both on the hardware and on the software side. And Tom says one of the leading, I'm all in on making us, at the end of the day, the leading provider of these holistic autonomous solutions and being the partner to every major OEM out there at the end of the day. When it comes to these capabilities, and again, that's not to say that no one will ever work with anyone else for different types of capabilities. There are other players to be had. There's other ecosystem providers. There's great folks out there like the Mobileyes and the NVIDIAs of this world that we were fortunate to count as partners in this, but that's -- those are well set.

Absolutely. Maybe I'll sneak one last question here about the business pipeline. I think, you're targeting couple of additional wins by year-end. But as you consider all these opportunities we talked about today, but also your available resources, are you kind of prioritizing or targeting certain end markets or verticals over others? Like, is there a wish list that we want to do more auto, more robo-taxi, more trucking, more EV toll. How are you prioritizing internally, which business you really want to, kind of, win in the near future?

Totally good question, and there is a lot of prioritization happening. And it's tough, right? It's -- we're having to say, no, as many -- to as many folks as we say yes to. And sometimes just because we think they have no idea what they're doing, sometimes because we just don't have the time and energy to actually put into it at this point because, I mean, our dance card is already effectively close to full for between the 2022 to 2025 starter production window. I mean, it takes a lot of work and a lot of effort to support an OEM for like a real program on these things. And our goal is work with the top players in each of their respective markets we have, like when we talk about our different vertical leads and we have our lead in Europe and for safety Volvo, our lead series partner in China is a large automaker in China, SAIC. They are the lead trucking partners, the largest manufacturer commercial vehicles and 40% market share in any aspects for long-haul trucking is Daimler trucks. We have our lead ecosystem partner, robo-taxis and ultimately provide consumer systems, which is Mobileye. So we have folks like that, that are really just pushing the envelope in terms of what's possible with these systems. And I think there's an opportunity for a couple more to be able to take on that we're really thinking through who we want for this next iteration. And -- but what we're also seeing is standardization of these systems. Right now, a lot of folks are all over the place in terms of, like, the different vehicle requirements, what they're putting in. What we're trying to do is really standardize it all to the same product, the same technology, the same platform, and that's what we've successfully done with all these guys, because, frankly, it just makes absolutely no sense to take 20 different approaches to what ultimately everyone is solving for, which is the same problem, from -- even if it actually works in all those 20 different approaches, which it won't. But even if it did, it makes no economical sense to do so. So that's part of the point of the standardization of hardware and standardization of the software, not just across a single OEM across their vehicles but across multiple OEMs. And that's what we're really starting to see.

Terrific. That was super helpful. We could go on forever, but we are over our time. So Austin, Tom, and the entire team at Luminar, thank you so much again. Really great conversation. I appreciate your participation in today's conference. It's great to have you.

Thanks, Itay.

Thanks for having us.

Anytime. Absolutely talk later. And thank you, everybody, for joining us for this session. Have a great rest of your day. And with that, we can go ahead and conclude. Thanks again, everybody.

One talk. Bye now.