Luminar Technologies Inc (LAZR) 2021 Q3 法說會逐字稿

Well, I hope that gave everyone time to watch that video. Welcome to our third quarter business update call. I'm Trey Campbell, Luminar's VP of Investor Relations. With me today are Austin Russell, Founder and Chief Executive Officer; and Tom Fennimore, Chief Financial Officer. As a quick reminder, this call is being recorded, and you can find the earnings release and slides that accompany this call at luminartech.com/quarterly review. In a moment, you'll hear brief remarks followed by Q&A.

Before we begin, let me remind everyone that during the call, we may refer to GAAP and non-GAAP measures in our remarks. 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.

With that, I'd like to introduce Luminar Founder and CEO, Austin Russell.

All right. Well, thanks, everyone, for tuning in. Great to be here with all of you guys. Glad we could do this and kick off this quarterly update. Certainly, no shortage of exciting things going on, but certainly a quarter to remember. So good video. Hopefully you guys enjoyed that as well, too, as we continue that series. So we're live from our HQ in Orlando. And with that, I hope you guys also enjoy kind of the video format that we have here and some of the interactive nature.

So just jumping right in, I think this quarter has certainly been one to remember following the relentless execution across the board and a host of new commercial wins and partnerships that are helping drive this industry forward. So it would be good to take a few minutes to talk through some of the recent events in a little bit more detail before I hand it off to Tom here for an update on our key business milestones and subsequent Q&A.

So it could be beneficial if maybe we share a couple of slides. These are also publicly available on our website for a quarterly review. So let me bring this up. You see that okay?

Thumbs up.

All right. So yes, first off is with regards to NVIDIA. We've got a lot of good questions for that. We're really coming hot off the heels of the announcement with them. So that's been great, and happy to share a little bit of more context about the significance of this. And their selection of our lidar for their autonomous vehicle platform known as DRIVE Hyperion. So Luminar's Iris will be the only lidar designed into the platform, which is targeted at delivering safe, highly assisted and autonomous capabilities. And NVIDIA has really been leading the charge on the modern compute revolution. And the way that we see it is that it's really going to be no different when it comes to the auto industry as cars move more from machines to effectively being centers and computers on wheels.

And of course, the emerging trend and comments read between our companies is more that we're becoming the de facto solution and standard for major automakers to power the next generation of vehicle platforms. We've already begun showing up alongside each other on a number of different OEM and AV partnerships. So in the case of NVIDIA's solution with us, those specifically and the full stack they're building out, the planned production is starting in 2024. Obviously, a lot of work that's been going on between that, but that would be the launch. So with that, it hopefully gives a little bit of context. And I think yes, you could see on the slide there, too, in terms of the position for lidar.

So just going straight to it. What's next? And what's next is -- let me see -- so from a passenger vehicle standpoint. So moving on to wins and momentum for passenger vehicles, I think it was important to highlight a couple of things. One is -- and I think some of this may be overlooked. But an important one is another major win, was publicly disclosed during Polestar's announcement of their planned public market debut, which they named Luminar as their partner for long-range sensing to enable autonomous driving capabilities. And super excited for them and the opportunity ahead. So we'll be looking forward to that.

The other side of it is Mobileye. From an execution perspective, they successfully built and publicly debuted their first fully integrated Luminar-equipped vehicle this quarter at IAA during Intel CEO, Pat Gelsinger's keynote address. So it's a big step forward and shows the progression of not only in new wins, but rapid development on the existing programs that we have. And I think we actually even went out public with Mobileye less than a year ago. So it's pretty good to see this all come to fruition.

So next up from a commercial trucking perspective. There's -- things have only been accelerating when it comes down to it. And we've announced actually 2 partnerships this quarter with 2 of the leaders in the AV trucking space for their next-generation platforms, which only reinforces our holistic foray into trucking. And our tech is going to help power and accelerate Embark's autonomous commercial trucking service that they'll be launching in 2024. They've got a number of great partners along the way from a vehicle fleet and operations standpoint that will be fantastic for them, and to see that happen and materialize is great. I know they also actually, I think just today, they closed their merger and now a public company as well. So it's great to see.

And then Kodiak, we're enabling some increased capabilities and maturity as well for the next generation of trucks. So it's good stuff.

Next up -- and this is all in Q3 so there's a lot going on. The vehicle integration partnerships. Roofline positioning is absolutely key to enabling a safe and a successful program, and that's why having that almost halo effect up on the top of the roof there is certainly important from getting that perspective across the board and making the most of the sensing systems. And so we've successfully announced 2 new partnerships in Q3 with top global automotive roof companies, Webasto and Inalfa. So these companies directly manufacture roofs for many of the biggest global automotive OEMs and customers, and now they're integrating Iris into their offerings such that OEMs can seamlessly adopt our optimized solution without having to do much extensive integration work themselves. And this is an important part towards the ultimate vision of how we can see this technology being standardized, brought to the broader industry.

So that's kind of the quick view there. One of the things that you may have seen in Trey's background that he had was actually be -- something that we call the Indy Autonomous Challenge, and I'll get to that in a sec. But one important model that we have that's kind of informal in the company is show not tell and we've had that since our inception. And one thing this quarter that we're very proud of is to be able to show off the true capabilities of our tech. One of the examples of this was this Indy Autonomous Challenge. It's effectively evolving itself into what could be seen as the modern DARPA brand challenge that kicked off this autonomous vehicle industry. It's a high-speed race around the Indy 500 track and with new drivers. It's pretty wild seeing cars that get going 155 miles an hour without anyone behind the wheel and certainly, the fastest our tech has ever gone. But it hopefully serves as an inspiration for what's possible. And I think what's really cool about it is seeing that it's the same tech and the same kind of subtech that's in these high-powered Indy cars that actually was making its way into consumer cars that you could buy. And that's what makes it, I think, really inspirational and again, showing on talent in terms of capabilities so -- raising that awareness. Now that was an inaugural event, but hopefully it becomes a recurring thing.

So at the same time, we also did conduct and conclude our European live road show this past quarter. The team had over 60 different customer or third-party engagements out there and had vehicles out in Europe to be able to do these live demonstrations. And what we heard loud and clear is that proactive safety and highway autonomy are going to have the opportunity to change the industry as we know it and excited to be at forefront of it.

All right. So all this is great in all. New major customer wins, expanding ecosystem and certainly, exceeded our expectations. But execution is everything when it comes down to the product. And at the beginning of the year, we outlined 5 key milestones in the business and how we measure success, and that's from both a commercial perspective and an execution perspective. And Tom will be getting into greater detail, but sneak preview on that is that we've been meeting or beating all the different 5 key milestones and part of why we increased guidance on some of those milestones in the past quarter. But for one of those, we've already accomplished that increased guidance as well.

So Tom will be getting into greater detail on that. But I'd say for the past quarter, we've had a particular focus on supply chain and industrialization as we progress on the path to series production and do have some significant updates as it relates to Iris. So both engineering and tooling are now predominantly complete. And we've secured our supply chain for next year in an environment where that's not always a given and are now entering the C-phase of our product cycle for Iris. So a huge leap forward. And in parallel, our contract manufacturing partners, Celestica and Fabrinet, are also continuing to execute alongside us as they help optimize process and build out capacity in preparation for series production. So it's obviously important to note that all this is a first for the industry. This infrastructure largely doesn't exist. We're creating it. It's really hard stuff.

Just like our technology, we're building out our supply chain from scratch. We're building out our tooling, all our process, all this stuff. So it's a very intensive period. We've been heads down just trying to execute to a tee on, but we believe that all of that will pay off big time over the mid and long term. So that's kind of the update as it relates to Iris.

From a software standpoint -- and the big things at the last update was the development of the Sentinel alpha. So that's a big step forward, and we actually were able to do a debut of our proactive safety with live demonstrations to customers at IAA in Munich. People were pretty blown away by what was possible and just appreciating how much we can improve safety today when it comes to active collision prevention. So great to see that all materialize, come together and make that happen. And it's all sowing the seeds from what we've had earlier on in the year and the investments that we make. It's all kind of materializing at this point in substantive capacity.

So before I hand it over to Tom, I just want to acknowledge and thank the Luminar team for all the hard work and just everything that goes into this. We know how intense things are, which over the past year, though, the team has actually nearly doubled in size and is only getting stronger by the day. So we're really able to punch above our weight class by attracting top talent from across both the tech and the automotive industries. And great to see people that are big believers in the vision as we pioneer this industry.

One last note as well just for all those that are on the call is with regards to the upcoming CES. You guys may be familiar, it's the large tech or now as much tech as automotive actually show in Vegas that obviously didn't happen the prior year. But as the world's kind of opening up, it's great to be there. We're going to have a blowout CES, a lot of stuff going on. So I would encourage everyone to -- if you're able to make it, it'll be fun.

So with that, I'll hand over to Tom to talk to the milestones and financials. Thanks.

Great. Well, thank you, Austin. So earlier this year, we outlined 5 2021 milestones to measure our progress against Iris and Sentinel industrialization, major commercial wins and cash spend. We raised guidance on a couple of these milestones last quarter, and I'm happy to say we continue to remain on track to meet or exceed each of these 5 milestones. And in fact, we've already achieved a couple of them.

Let me review now in more detail. Our first milestone is to enter the C-sample phase for Iris. We have now predominantly completed both our engineering and production tooling for Iris and secured our supply chain for the subsequent year. We are entering the C-phase as we speak and continue to work closely with both Celestica and Fabrinet to get ready for series production. Manufacturing complicated optoelectronic devices in series production scale at automotive-grade quality is not an easy task, and our team has done an amazing job working around the clock to make this possible.

Separately, our integration of OptoGration is going very well. This secures a key element of our supply chain for series production for the foreseeable future. And that's very important, especially amidst this global chip shortage we all experience.

Our second mile is delivering an alpha release of our Sentinel product. In our last update call, we showed a video demonstrating proactive safety on the track. We followed that up with a debut in a more challenging test scenario at IAA in September. This included automatic emergency braking functionality at speed scenarios that existing ADAS systems just can't match. We have successfully developed our alpha version of Sentinel and will unveil it at CES in early January.

The third milestone was to achieve 6 major commercial wins this year, a target we doubled during our last earnings call. I'm proud to say we have already achieved this goal with the recent NVIDIA and Polestar wins. These 2 major wins bring our publicly announced total to 6 for the year. Two other wins this quarter we are excited about are the autonomous trucking companies, Embark and Kodiak. Combined with our Daimler Truck partnership announced last year, these wins extend our leadership position in the autonomous trucking space.

Our fourth milestone is to grow our forward-looking order book this year by at least 16%, an increase from our original 40% target. We're going to do a formal tally of where our forward-looking order book stands at the end of the year. But based upon our success so far this year, I'm confident we're going to meet or exceed that target.

Our final milestone is to maintain strong liquidity, with the target to end this year with a greater cash balance than we started. We remain on track to meet this and ended Q3 with $545 million in cash.

Before we move to Q&A, I'd like to share some quick highlights from our Q3 financials and outlook for the year. Revenue for the quarter was $8 million, up almost 90% year-over-year and 26% from last quarter. The increase in revenue was driven once again by increased program revenue as we start to launch and execute more series production wins.

Our non-GAAP gross loss for the quarter was $1 million. Our contribution margins for our lidar sensors continue to remain very healthy. Of the $9 million of non-GAAP comps we had in last quarter, approximately $8 million of that was nonrecurring engineering expenses and fixed manufacturing costs associated with our advanced manufacturing facility here in Orlando. As we discussed previously, our gross margin private series production is not a good indicator of our future unit profitability. Instead, we are focused on optimizing our unit contribution margin for when we reach series production and scale. We plan to discuss this in more detail as we get closer to series production, including the progress to achieving our mid- and long-term BOM targets, and we continue to be on track to achieving those.

Our cash spend during the quarter was a little over $37 million, incrementally higher than the first half of the year as we accelerate our investments to fund the successful launches of our multiple wins. For the year, we still expect revenue to be in the $30 million to $33 million range as program revenue from recent wins continue to come online. As a reminder, this guidance excludes the impact of any accounting contra revenue charges. During Q4, we expect to incur approximately $1 million of contra revenue associated with warrants provided to a customer before we went public in exchange for helping provide resources for product development and validation.

In closing, I'd like to continue to thank the Luminar team for another amazing job and strong quarter. With that, Trey, let's move on to Q&A.

Thanks, Tom. Thanks, Austin. With that, we'll start the Q&A, and we'll try to get everybody on the call. (Operator Instructions)

First question is going to come from Tristan Gerra at Baird.

You mentioned a couple of trucking design wins. Could you give us some color as to when you expect those wins to start ramping? And also, will it be for L3 or L4 capabilities right off the bat? I mean given that L4 is basically transferring the liability from the driver to the vehicle and trucks, obviously, are very heavy, do you think that it's going to take some time? And will your lidars be more for proactive safety into trucking?

Yes. No, it's a good question. And I think there will be more to come ultimately when it comes to holistic trucking strategy at CES, but this is as good of a lead-up as we could hope for. And the focus, I'd say, for these deployments is generally around driver out-of-the-loop L4 capability for this. So that's what people have been going for. And basically, they're targeting a 2024 time line in the case of Embark for the commercial deployments. So exciting to see that happen.

And obviously, part of the missing puzzle piece has been the level of performance and capability and safety of what can be enabled with the sensing aspect. And from a perception perspective, that's incredibly valuable. They did a whole analysis on that and published that as well. So that's been good.

Kodiak is at a little bit earlier stage, but they've been -- they just raised their next round and excited to see them at full speed ahead. So it's -- from a commercial standpoint, I continue to think that there's a massive opportunity when it comes to autonomous trucks. And in the wake as well of the global supply chain shortages and everything going on, I think that need and demand is only ever more clear. So that makes sense.

Great. And then for my follow-up question, as you reiterated the full year revenue guidance, your implied Q4 revenue is basically 50% sequential increase and more than doubling from your Q1 revenue level. So could you give us a little bit of color as to what's driving that inflection point? And also at this point, how much NREs are embedded in your revenue currently?

So Tristan, most of our revenue now is the program revenue or as you said, the NRE. As we win more production wins and we start doing the development work to get that ready for series production, that revenue is going to grow. And so if you go back to Q1, we really had only, I would say, one win that was the predominant contributor to that. As we get to Q3 and then also to Q4, there's actually going to be several. And so the more wins we get and the more work we do, the higher that revenue goes.

Note that, obviously, when it comes down to it, the revenue that you're talking about in this context is all leading up to series production that is a different kind of revenue altogether. So as much as anything, it's -- I'd say it's an indicator. It's not a precise indicator at all when it comes down to it. But the question is that, what kind of revenue is it with this? And at the end of the day, obviously, the most significant indicator is the actual wins themselves that you can see. I mean that's the factor that matters most that leads to how you can build an order book and how you can build fundamental value for the business.

The next question will come from Emmanuel Rosner at Deutsche Bank.

Can you hear me?

Yes, we can hear you, Emmanuel.

Yes. Yes, we can hear you, Emmanuel.

Awesome. First question is on the NVIDIA opportunity. And I would love to hear you speak about what you find exciting about it, maybe if there's any thinking around time line or sizing of the opportunity. Obviously, this is a partnership now. How many automakers or other type of players do you expect to source system directly like this as opposed to sort of different parts of the system separately? Just any color around how you think about this opportunity.

Yes. Yes. So no, I mean, I think this is -- it's really interesting for sure. It's very different. And I'd say it's, in some ways, even much more significant than any given single OEM win because this is a platform play, right? And being designed in where this is super meaningful. I mean I imagine in the case for some of you guys, you may cover NVIDIA as well or have some context with regards to how well they've done taking over the -- leading this modern compute revolution as they're describing it. And I think in the auto industry, they've been doing incredibly well for themselves. So it's awesome to be working with -- I mean, there's really kind of 2 key players that dominate the space. There's Mobileye and NVIDIA, and now we got both.

So yes, I think -- and would certainly look forward to a lot of success on their part. Obviously, some of it's in our control, but some of it's in their hands. At the end of the day, our goal with all of this is to create scenarios where it's heads, we win, tails, we win type scenario regardless of who ends up -- who the OEMs work with at the end of the day.

Just in terms of sizing up the opportunity or for the time line, any details you can give on this?

Yes. So they've stated publicly the plan -- the first planned production is in 2024.

And then second question. So obviously, you've hit a lot of your goals already for this year, especially in terms of commercial wins. Where do we go from here? What should we be looking forward to?

It's a very good question, Emmanuel. We've already hit our 6. I don't think we're done winning yet this year. I don't know what that means that we'll necessarily get to a point where the customers are going to be in a position to make public announcements. But the pipeline of the opportunities that we're engaged on, I would say, is not only the biggest it's ever been, but it's the highest quality it's ever been, too.

Yes.

The next question comes from Gus Richard at Northland.

For the most part, you guys are vertically integrated. The only exception is the compute engine and your software is written to NVIDIA. And I'm just wondering if you can talk a little bit about the strategy of having everything vertical except for that piece of the solution.

Right. Right. Well, I think it may be tough to buy NVIDIA but -- at this stage, but maybe next time. So I'd say that a lot of the kinds of vertical integration that we're doing really comes down to extremely specialized work that we do when it comes to the fundamentals of the lidar, when it comes to certain aspects of the software on that. I would say that like from a compute perspective, it is much a platform play as anything. And I don't actually think we need to own that. This is also where -- there are great solutions that exist today. Like Mobileye has a great solution, NVIDIA has a great solution with that. I don't think we have to reinvent the wheel on those. On the question -- in the sense that I don't think that those are the fundamental bottlenecks to seeing this kind of technology and system and value standardized across the broader industry. I think that we're moving full speed ahead on that.

So the question is that how do we continue to accelerate things from a value perspective as it relates to the lidar, as it relates to the software, as it relates to all of this stuff. And we have to make very smart and deliberate decisions when it comes to make-buy partner and -- for the broader ecosystem of these things. And in this case, when it comes to compute, this is one that we've specifically chosen to partner. So if that helps answer the question, makes sense.

I think it does. And then a lot of the auto OEMs are developing their own silicon for autonomous driving. How difficult if some of the OEMs develop their own ASICs? Will it be to port your software over to their ASICs? I mean how do you see that playing out over the next couple of years?

Yes. Yes. It's a good question. I think -- well, for the most part, I'd say for the vast majority of OEMs and folks, we see them sticking to NVIDIA and Mobileye-based platforms as well that get designed in. But that said, we are able to be compatible with other compute architectures as well, and we're actually looking into that in a couple of cases. But there is incremental -- I mean, as with each different compute architecture, there's incremental work that is required, but we're obviously trying to build our software in a scalable of a way as possible when it comes down to it. So it depends totally on the type of compute architecture and just how exotic it necessarily is. But for standard architectures, it's not exceedingly difficult but internally doable.

The next question comes from Michael Filatov at Berenberg.

So the first one around the NVIDIA win. I guess I'm curious, is the NVIDIA platform sort of an open-source platform, which you guys were selected as sort of the primary offering or the one with the biggest stamp of approval? Because it seems like you go to their website and there's 3 other lidar suppliers on there and 8 different sensors from other suppliers on the NVIDIA website that are being offered. So I guess how does that compare to what you guys have won here?

Yes. Yes. No, it's -- and I think it's actually not an unreasonable question. It's actually kind of confusing. So this is where -- just for sake of clarification. The distinction here, and this is really important, is that -- so NVIDIA obviously does have a broader ecosystem of companies that use their chips. The key distinction here is that -- I mean, using their chips is great and all. We've been doing that for a few years and along with other companies as well and -- as part of the DRIVE ecosystem and everything they have, which is cool.

But the significant thing here and the really newsworthy thing is not that at all, which, I would say, probably half the companies in the industry has. And to their credit, NVIDIA has done a good job of collaborating their chips and technology. But when it comes down to this, this is for their autonomous driving solution. This is for their full stack solution, and we're the only lidar company for that. And that's a very important and critical distinction. This is what they're building their software stack on. This is what actually gets designed in. That's the important part. And again, that's not to say that no one else will ever work with NVIDIA by default for autonomous vehicle companies. Like I said, most -- or lidar companies. I mean most do in some capacity. But this -- specifically for Hyperion, this is their play.

I guess, ultimately, will the OEMs who are working with the Hyperion platform, will they still individually source the sensors? Or will they take sort of this one-stop platform with all the sensors on their website that are listed, including yours?

Well, I think it would be virtually impossible to try and hot swap a different sensor for a software that's not built on it, even if you tried necessarily. But -- and obviously, the only reason why you'd do that in the first place is if, a, it was -- had better performance; or b, it was more cost effective and could actually do that. But that said, that's -- this is one of those things where it's obviously very high barrier to entry, but equivalently very high barrier to exit with this. So it's part of the design of this.

And again, not to say that someone couldn't work with other components. Or actually, I mean, like one configuration is you could also -- I mean, people work with NVIDIA where they don't have a lidar at all. You can do that. It's just a question of if you want the autonomous driving capabilities and you want to be able to enable that and you want to use NVIDIA's solution, then you would use this. An automaker can still develop their own stack and use an NVIDIA chip, but then that defeats part of the whole purpose of what's trying to be done, is to see standardization of functionality, standardization of technology as part of this play.

Okay. Got it. And just one quick follow-up around the vertical integration and the acquisition of OptoGration. It seems like you're vertically integrating a very -- around a very specific design. And I'm curious, does that mean you're really confident in this design sort of in perpetuity? I mean specifically the in-gas photodiode, right? Is there any road map to change away from that? Because if you're vertically integrating around it, I guess what's the strategy in mind?

We have absolute 100% extreme confidence that this is not just a path for us, but actually, I don't know if we're aware of any other viable path to get to this level of stack and economics that are ultimately needed to be able to meet kind of these long-term trajectories of building this kind of business in this value and to scale Luminar 10x, 100x over the long term for what we want to do. So yes, it does absolutely reflect that conviction.

Now part of this, though, is -- that doesn't mean that we're like done in terms of our tech development for -- by any means. And there's 2 aspects of that. One is obviously going beyond the hardware and into the software side, which is really starting to see a lot of results on. But the other aspect of it is, from a lidar standpoint, too, part of the reason of bringing these guys in-house, I mean, is to actually build out an automotive supply chain. Like most -- if you try and go into series production with a lidar as a steady state, like the supply chains don't exist to support a series production product for a lidar as it stands right now. Like they're not there. They're not automotive qualified. They're not IATF certified, none of that stuff. You have to be able to actually establish this established process and have that ownership to make this happen. Otherwise, you just don't see it even coming to fruition.

So we see all the supply chain challenges that people are having today. The risk is 10x higher when you get to this level of specialization with all this. This has already kind of been the case for this. And that's why we've gone -- in some of these cases, like I said, part of the reason of the significance of how we were able to do this is that for these kinds of companies -- and the same thing for Black Force Engineering, for example, that we acquired previously.

We signed up these guys with exclusivity agreements years ago such that it would enable us the opportunity to work closely with them, collaborate on this and ultimately have the opportunity to potentially buy them. And that's -- and basically, that puts it in our favor. But yes, it's going to continue to be an important part. But again, for next-generation component development for this, these are the right guys to do that. It's iterating the same architecture, but we -- there are improvements, and we can move at an even much, much faster pace and take it out of niche world and in automotive world.

The next question comes from Aileen Smith at Bank of America.

I wanted to follow up on Michael's question on how automakers can develop their own hardware stack and spec components accordingly while still using NVIDIA's chip. Is there any precedent with NVIDIA or other players and not just for automotive but possibly other industries as well that you guys look to assess how components and hardware on a reference platform ultimately make their way into being quoted on production platforms by the ultimate customers of NVIDIA? I think we can assume that being on a reference platform makes for somewhat of a sticky position, but are there examples that you can point to where that clearly demonstrates that is the case?

Yes, there are examples. I'm not sure -- I don't think we could probably speak on behalf of NVIDIA for them since they're not all public. But I think when it comes down to it, the opportunity here for what it is -- and you have to separate out into 2 different things, is one is like the compute platform itself or the chip people using video chips, I mean originally, they actually -- I think they got their name in gaming, right, of all things and then kind of moved over to automotive and all the stuff. So people use it across the board.

The value here is the autonomous vehicle platform. It's the full stack. It's the software. It's the solution. And if you're using an NVIDIA solution as part of that, then that's where we get embedded into that. That's what I'm saying. You're not going to say that it's impossible to use an NVIDIA chip without a Luminar lidar. Like people are doing this today on cars, and they don't have that. But again -- but part of the whole point of the benefit of what they're doing is to enable that solution.

So for other automakers that have developed their own full-stack solutions or are developing it, you may have seen some of our announcements already with them in terms of what it comes out or which partner directly regardless of whether NVIDIA is in it or not with their full-stack solution.

One example, for example, with this is that -- and here's like a hybrid, is you have Volvo, which we've obviously been very -- Volvo has been very public with us. They're using an NVIDIA chip. They're using the full-stack solution from Zenseact and us and then also using our lidar. So that's like one example there, too, of where it's not that, but it is still an NVIDIA platform. But we still got the win. And that's where it kind of goes back to the comment of heads, we win, tails, we win. It's -- that's the kind of play and setup that we want to have.

Okay. Great. That's helpful. And then I want to make sure I understand the 6 commercial wins secured during the year and the 2 additional ones in the quarter, which I think you cited as Kodiak and Embark. I would have considered the NVIDIA one to be included in those commercial wins. So the question is, was NVIDIA already included in your cited commercial wins? I think you had one that was undisclosed as of earlier this year. Is it not counted because it technically closed post-quarter? Or is being selected for inclusion on a reference platform not considered a commercial win until it's for production with an automaker?

Yes. So Aileen, we've kind of laid out in the past our definition of major commercial win...

Sorry, but just to clarify, I think there's a confusion on that. Kodiak and Embark are not the major commercial wins.

Right.

It's NVIDIA and Polestar. So sorry if this is confusing.

Yes. That's where I was going to go, which is the 6 that we've had this year that are publicly announced that kind of meet our threshold for major commercial wins were SAIC or Shanghai Auto. Pony.ai, Airbus, the Volvo standardization win, Polestar and then NVIDIA.

The criteria that we have for major commercial win is that there's a written agreement that kind of memorializes that the customer has deliberately selected our technology. It needs to be with a major industry player that is going to lead to a series production program or an equivalent major commercial platform, right? So when you think about Airbus, that's not series production, but they're a major industry player. And what we plan to do with them, we expect to be a significant commercial opportunity.

That's the bar to get to a major commercial win. Kodiak and Embark are great customer wins. We're optimistic that their -- those partnerships are eventually going to evolve into major commercial wins. We're not counting them in that category today.

In the past, we kind of said we had one undisclosed. That was Polestar. And so look, there's sometime a time line between when we actually achieve the win and then when the customer is going to publicly disclose that. We work with our customers on very sensitive projects, and we don't publicly disclose them until they're ready.

The other thing that I would say is in our forward-looking order book, which once again is our estimate of the expected revenue over a lifetime of a program that we've won, we only have our major commercial wins in our order book. And when we update it at the end of the year, we're only going to include the major commercial wins we have on December 31, 2021, in that calculation. Now look, it's -- we're going to include those major commercial wins, whether they're publicly disclosed or not, but they will be in that calculation.

Got it. Perfect. That clears up some questions I had around the commentary on Slide 12, but that makes sense.

The next question comes from Itay Michaeli at Citi.

Awesome. Great. Tom, you mentioned to an earlier question around the pipeline and the funnel being really strong and of high quality. I was hoping you could give us a little bit more high-level detail about what you're seeing there maybe by geography, premium versus maybe even mass-market automaker interest as well as the idea of sort of option versus standard equipment as you've seen with your other customer recently. So just curious if you can maybe go a little bit more into your earlier comments about the funnel and pipeline.

Yes. Itay, it's all of the above. If you -- I always think of the automotive industry as having 3 major regions: China, North America and Europe. And then you also kind of have the luxury brands and then the mass-market brands. And then right now, I would say, because we have finite resources, we're really prioritizing where we're going to devote serious resources for the potential programs that we're engaged in.

And so from this perspective, I would say the pipeline is across the 3 major regions, across premium versus mass market. It includes standardization opportunities. It includes both hardware as well as hardware plus either Sentinel software aspects of it. And I would say it's almost exclusively companies that are in that passenger vehicle or commercial trucking landscape.

Perfect. Good. That's very helpful. And then just as a quick follow-up on the 2 commercial truck partnership wins, is there any potential software contemplated in those wins? Or is it still a bit too early for you to know and finalize that?

I'd say, for the most part, for some of these providers, they're developing their own software stack when it comes down to it. But there are different things that we contribute certain components of Sentinel to. It's just the levels and varying degrees depending on the OEM or, in this case, AV player. I would say there's a general tendency for greater opportunity with the OEMs themselves, which is our focus, as opposed to providing software for independents like tech AV companies when it comes down to it and because that's where we think the value is, say, like the Daimler Trucks of this world, for example, which something that we had the prior year. And that was a good example of like a major win. So yes, hopefully, that makes sense.

Yes. Maybe a quick follow-up, Austin. Would you get data back from these partners as well that you can leverage internally? Or is that to be determined as well? Even if you're not selling into software, is there any leveraging of their data that you can leverage yourself?

Yes, yes. There's absolutely significant data benefits when it comes to these kinds of things and particularly when there are software components to all of this, like in the case of some of the other relationships or the mobiles of this world or whatever it may be.

Yes, I mean I'd say that the intent ultimately is to be able to have -- you actually don't want all the data. You just -- you want useful bits of information as these cars are driving around to have that global presence, have those miles and everything. And I mean just as an example, any single OEM deployment and data from that is orders of magnitude more than every robo-taxi company and autonomous vehicle company on the planet combined like times 100. So that's the value that you're talking about here. And we don't have to pay $1 billion a year to maintain a fleet. We actually get paid to do it.

So of course, what that means is that you also were able to have product improvement, safety improvement and other stuff along the way. So it makes it easier -- it's not just like a -- it's not a one-sided thing. It's a win-win for everyone.

The next question comes from Jaime Perez at Lafferty.

Great quarter. So I'm looking at Slide 7 at the NVIDIA platform. But what I see is lidar that you used to perform redundancy, and so you're basically ever checking the 3 radars that are providing in front -- forward sensing. Is that correct? And so has that -- how does that -- can you give me a little bit more detail? Is that ever correcting?

I think that's sort of the semantic of how you described it. I mean effectively, all sensors have to be redundant. You have to have some kind of redundant mechanism. For example, if you had a functionally critical system that was dependent only on just a single Luminar lidar, like nobody is deploying, that's -- there's a reason why people still have to take a camera on their vehicles or other stuff or why they're implementing radars.

Obviously, the core of the data that you actually get to understand where the positions of the objects are and everything that's going around and seeing comes from the lidar, but it is redundant in the respect of the architecture of a vehicle such that if any single thing were to fail, the car would be able to come to a safe stop on the side of the road in autonomous operation or be able to still have some basic capability.

All right. And my follow-up, any initiative in the nonautomotive side, like aerospace and robotics and...

Yes. I mean...

Look, I think we announced Airbus earlier this year. But I would say where we're spending predominantly most of our time is on the automotive as well as commercial trucking landscape. I think that that's an area where our long-range lidar can make the biggest impact. And quite frankly, when we look at our TAM, it is by far and away the biggest part of the opportunities we see out there.

The next question comes from Dan Levy at Crédit Suisse.

Tom, I want to follow up on one of the prior responses you gave, and this is just on some of the -- some color on the order book. These discussions that you're having on the automotive side, are these all for series production programs? And maybe you could give us a sense of how much your agreements with Volvo on XC90 from earlier in the year has led to other dialogues or discussions with other automakers for series production programs.

Sure. Yes. Look, Dan, we're focused only on engagements that are going to lead to series production or for the area or equivalent for areas outside of passenger vehicle programs. We're not -- I think we're past the point where we're going to be engaging with customers and selling them a lidar so that they can test it on the bench. This is, how do you take our technology and how do you deploy it in the series production vehicles?

The announcement we made over the summer with Volvo, where we -- our original deal with them was an option leading to standardization. A lot of customers took notice for 2 reasons. One is it demonstrates Volvo's confidence in our continued execution that they're willing to give us more and more business. They've said publicly that their goal is to ultimately put our technology on every vehicle that they made. They said in their IPO perspective that they plan to put a lidar on all of their next-generation electric cars, not just the successor program to the XC90s. And so that's a big vote of confidence in us.

But more importantly also, for those competitors to Volvo, they see their competitor deploying this technology very soon. And they want to make sure that they're not going to be in a position where they're going to fall behind from a technological perspective.

Great. So it sounds like a lot of the other programs that you're looking at are probably similar in terms of the type of functionality and capability that it's enabling, sort of highway autopilot, this Level 2 plus Level 3-type functionality?

Yes. So here's the way that I would say it, which is -- and you kind of see it with Volvo, which is the highway autonomy is the sizzle that everybody loves to talk about and get in the door. And when you look at some of the pricing that folks like Tesla are able to get for that type of functionality and for an even better system, it's a great profit generator for these OEMs. So that is the sizzle that I would say drives a lot of the attraction.

What drives standardization is that proactive safety element of it where you can substantially improve the safety capabilities of existing ADAS systems, right? And so when you look at some of these proactive safety videos that we've been putting up there and what these systems are able to detect, it's just -- no other ADAS system on the road today can really approach some of the capabilities and some of the speeds that we've been demonstrating.

And it's important to point out that the kind of proactive safety functionality that like that's something that really doesn't even fit on the traditional level of scale, as you're talking about. This is -- it's an entirely different dimension of capability that also has definitely gotten people excited. And -- but yes, I agree with everything Tom said. And yes, no, we just got to keep digging into the next level.

And then my follow-up is I want to go back to the NVIDIA agreement. Maybe you could give us a sense of what the process was to get sourced. Was this a deep RFQ process in which they were looking at a number of different lidar providers, and after this process, they chose you? And how much is their exclusivity versus potential for NVIDIA to add other lidar providers to the platform?

Yes, yes. So 2 different things. I'll answer the first part of it. One is when you -- the whole point of when you build a stack out, you have to be able to have an architecture that you're building that with. And in theory, someone could build different stacks that are -- try and build off different types of sensing systems. But this isn't like a hot-swappable commodity with the same kind of architecture by any means, is that they're actually generally dramatically different. And when you do that, you really have to be able to build your stack on a given architecture.

And that's why it's been such an important focus of ours to make sure that we can get -- successfully get designed in to these key defining platforms because the decisions are being made today that can have effects for 10 years, maybe 20 years in some cases, on -- from a platform cycle perspective. So that's a really critical part, and -- when it comes down to it. So I would say -- and what was the first -- what was actually the first part of the question there, too, it was -- that you had?

What was the process...

The RFP process.

Oh, the process...

Correct, the process.

Yes, yes, yes.

Where you stacked up against other providers.

Yes. I'd say, man, like this stuff is intensive. It's -- and not in the sense that -- there's a lot of work that goes into this. I don't want to speak on -- too much on behalf of any -- NVIDIA or any OEM or anything when it comes into play, but these are multiyear processes that extensive amounts of diligence is done and a lot of work that goes into this. Let's just say, it doesn't exactly happen overnight.

And with that, that's going to end our Q&A session. I want to thank everybody for joining our Q3 earnings call. We look forward to talking to you for our Q4 earnings call and potentially at CES if you can make it there. Thank you, everyone.

Awesome.

Thank you, everyone.

Yes. [Good call]. Thanks, everyone.

All right. Kevin, you can close this out.