Call Start: 12:00 Call End: 12:33 January 1, 4539 ET
Luminar Technologies, Inc. (NASDAQ:LAZR)
Evercore ISI 2nd Annual Technology, Media & Telcom Conference Call
September 08, 2022, 12:00 ET
Company Participants
Thomas Fennimore – CFO
Conference Call Participants
Unidentified Analyst
Chris McNally and Doug Dutton from the automotive and mobility team, and we’re really just happy to welcome Tom Fennimore, CFO of Luminar, and for our next presentation for a really quick background. So Luminar is a company dedicated to developing LiDAR technology and software, and we’ll get into that for autonomous vehicles and advanced ADAS.
Founded by Austin Russell in 2012, Luminar trades under the ticker LAZR, L-A-Z-R, and has a market cap of about $3 billion. Before coming to Luminar, Tom was the Global Head of Automotive and Co-Head of Industrials Group at Jefferies. Prior to that at Goldman Sachs, where he was on the advisory side, and sort of seeing everything that was going on in autotech years before the public markets.
Tom, first, welcome, thanks so much for taking the trip here. Maybe because you can convey the story of Luminar sort of even better than we can. Could — particularly for the guys who don’t know a lot about Luminar and the LiDAR industry. Can you just give a brief overview for a little bit of the history of the company.
Thomas Fennimore
Sure. So Luminar was founded a little over 10 years ago by our CEO and Co-Founder, Austin Russell. It’s one of those technology stories where a young teenager really developed this technology in his garage, and then over the years has really come to commercialize that technology.
So the core of what Luminar does is what we call LiDAR. As I described to my two young boys, we effectively develop products that shoot lasers out of cars so that they can drive themselves. And so effectively, what we’re doing is — our LiDAR sees 250 meters out. It sees — has about a 120-degree horizontal field of view and then a 30-degree vertical field of view. Why 250 meters? That is the distance you need to bring a vehicle to a full stop traveling at highway level speeds. And so in that cone, 250 meters by 120 degrees by 30 degrees, we effectively see every object out there, and we operate at a wavelength of 1,550, which effectively allows us to put 17x the photons into that cone relative to a 905 wavelength is kind of where the telecom industry grew up.
So that is the core of what we do. Our products are going to start to go on vehicles being sold to consumers around the end of this year. Our first wave of customers include Volvo, Polestar, Mobileye, Shanghai Auto. And then our second wave of customers will include names like Mercedes-Benz, Nissan, Daimler Truck and a few others. And so we are on the verge of start of production, which will see a significant increase in our revenue here in short order.
Question-and-Answer Session
Unidentified Analyst
And Tom, if we had Austin here, we know you would start to get into the advantages of 1,550 and quickly would get over my head in terms of some of the physics. Could you just also break down because there’s so many LIDAR companies we get this technical question. There are some competitors who have done well in 905. There’s also FMCW. In really layman’s term, some of the advantage you talked about range, but just how you’re going to customers and sort of maybe about 1,550 being a sort of Gen 2 technology.
Thomas Fennimore
Yes. And so I mentioned earlier that 1,550 allows you to put 17x more photons out there relative to 905. The reason for that is primarily because 905 wavelength is right outside the visible spectrum of light. And if you put too much power through a 905 laser, it will actually do damage to the human eye, including blindness. And so if you think about autonomous vehicles, you can’t have autonomous vehicles with 905 lasers driving around, putting too much power through them because of the unsafeness to the human eye.
1,550 doesn’t have those limitations. Historically, the issue with 1,550 has been twofold. One is you need a very specialized receiver that can process the light at that wavelength, and that is an InGaAs material. In gas is more expensive. And so that presents the second challenge is not only the special material, but the cost. What we’ve been able to do, and what Austin really figured out, is how to take an InGaAs chip and just put a very small dab of that on to receiver and being able to make that process the light at that wavelength. And so we’ve been able to do that at a few dollar cost. And so that what’s gives us the robustness.
And so the reason you’re able to get to 17x more photons in there is because you can have a more powerful laser on that at that wavelength and still have it be eye safe.
Unidentified Analyst
And then Tom, Luminar clearly has one of the largest book of business. You talked about the sort of 2 levels of clients. Can we dive into that and talk about some of the most important sizable contracts, particularly? Obviously, Volvo where the Luminar offering is going to be standard fit as opposed to an option. So if you could just go through — again, we don’t have to use hard numbers, but just kind of rank order some of the size of the programs that are going to really matter over the next couple of years?
Thomas Fennimore
Yes. So we’re very proud of all the customers that we’ve won so far at Luminar. Volvo was our first customer. And I believe how that relationship has progressed is very instructive in terms of how generally the automotive industry is looking to use our technology.
So Volvo, initially when they selected us a couple of years ago, first, we had a multiple year relationship with Volvo. It took a while. This is very important technology that an OEM is putting on our vehicle. And so — they’re going to want to test it. They’re going to want to validate, they’re going to want to do their due diligence before they make the conscious decision to select that technology to put on vehicles that they’re going to sell to the consumer.
Initially, when Volvo made that selection with us, we were going to be an option on the successor to their flagship SUV, their XC90, which will be coming out here shortly. About this time last year, they decided that instead of it just being an option where the consumer had to make the conscious decision to select that trim package when they bought the vehicle, we’re going to be standard. And the benefit of that is it’s going to give us more volume of certainty where you don’t have to take the take rate risk and the consumer risk. And so as part of that, that is allowing us to ramp up our volumes with more confidence, and I’ll come back and talk about our capacity planning in a sec in terms of how that impacts us.
Volvo has also said publicly that their plan is to ultimately put this technology on every vehicle that they make. And as long as we continue to execute and as long as we’re reasonable on pricing, we’re very confident that we’re going to get there, even though the only vehicle that is currently in our backlog is that successor to the XC90 and not all the other vehicles that Volvo makes.
And so Volvo has a sister company, Polestar. And so we’re — Polestar is another important customer. The other major customer we have in terms of that initial launches that I mentioned earlier is Shanghai Auto. Shanghai Auto is the largest automotive company in China. They are developing their — a new brand that they want to compete directly against Tesla in China, particularly the Model 3. That is what they call their Rtec. One of the key ways that they wanted to differentiate that vehicle platform relative to Tesla was to actually put a LiDAR on it. And so they reached out to us. They wanted us to move quickly to get launch for later this year.
If you live in China, you can actually put in a reservation now to buy an R-Tech with our technology on it. And in some of the marketing videos that Shanghai Auto was done in China, our LiDAR is actually at the forefront in terms of how they’re positioning that to the consumer.
And so those are going to be 3 of our initial customers and very excited about there. And once again, we’re getting a toehold in with these customers. They tend to start at their premium vehicles, but they’re really developing their autonomous systems around our technology, and it creates a lot of stickiness. And as I mentioned earlier, as long as we execute there, as long as we keep pricing reasonable, we will get more and more business at these customers.
Unidentified Analyst
And Tom, broad — broad time line for some of those platforms. So XC90 end of this year and then obviously ramping Polestar. What is the first start?
Thomas Fennimore
Polestar is going to probably be a little behind that as well. And then…
Unidentified Analyst
We know the model, I know Polestar is going to be very long.
Thomas Fennimore
That’s going to be the Polestar 3.
Unidentified Analyst
Polestar 3.
Thomas Fennimore
And then we’re going to be on the Shanghai Auto one. I think is the — that’s going to be imminently here. That, I think we’re — the Shanghai Auto, their plan is to have that vehicle out on the road with our technology by the end of the year.
Unidentified Analyst
And in the past, you’ve given a pipeline number of OEMs that you’re working with sort of RFPs that you’re participating. Any numbers that you can update around, sort of — it seems like those numbers are quite…
Thomas Fennimore
We have approximately 10 major commercial wins in total so far. We’ve announced 2 major ones this year, one with Mercedes-Benz. The other is with Nissan. The Nissan, both of those are very interesting. We’re very proud of those. The Nissan one is interesting because that’s the first mass market brand.
And what Nissan has done is they’re developing their next-gen safety system, and they’re incorporating LiDAR into it, our LiDAR. And so we’re working with them to design the system. Nissan has said publicly that this system is going to be ready to go around the middle of this decade where they’re going to start putting it on consumer vehicles. More importantly, they said their goal is, by the end of this decade, 2030, to put this safety system on every vehicle that they make, and Nissan makes approximately 4 million vehicles a year.
When you’re talking about the cost of the LiDAR, it’s one thing to take that cost of a system, and our initial ASP is going to be around $1,000 plus or minus, a little bit lower in high volume, a little bit higher in low volume.
But if you’re selling a $70,000 SUV, it’s much easier to absorb the $1,000 cost into that. When you think about the mass market, the value proposition is a little bit more challenging just given the lower price point. And so to see a mass market brand like Nissan, say, we want to take this technology and put it on every vehicle we make by the end of the decade, that’s a very powerful statement.
Unidentified Analyst
And is Nissan do you think that potential customer — I mean that first one that you could start to look at lower price points into the end of the day?
Thomas Fennimore
Yes. Look, we clearly have product for [indiscernible] to take our costs from $1,000 to a lower price point to help increase the penetration rate on the mass market brand. And that’s going to play out over the next few years. But absolutely, we think that the value proposition for putting LiDAR not only makes sense for the luxury vehicles, but it can make sense for every vehicle.
And it’s really — Austin is — one of his goals and he kind of laid out his strategy earlier this year is he wants to save over the next 100 years, 100 million lives with our technology. And in order to do that, you really need to democratize the safety of the vehicle, right? Not only should the high-end luxury vehicles be safe, but we think every vehicle on the road in every country of the world needs to be safe.
Unidentified Analyst
And obviously, it’s going to be customer dependent. But can you remind people some of the feature sets that your advanced LiDAR is allowing? And again, I know Austin hates the Level 2, Level 3 plus [indiscernible] and Volvo doesn’t like that. But there’s clearly different amps — Volvo is going to be probably on the high end of the feature set, Nissan somewhere in the middle.
Thomas Fennimore
Yes. So a couple of responses to that. I think there’s been these levels that are out there, I think that sometimes they get lost in these definitions, right? You now have Level 2, then Level 2+, and I’ve even seen Level 2++. We kind of look at it. We try to simplify it at Luminar. And at the end of the day, you’re either assisting the driver or you’re autonomous. And those are the only 2 things that matter, assisting or autonomous. And then you got to look at the domain because that’s important as well.
And what we’re on, particularly with our customers and the autonomy that we’re enabling initially is highway autonomy. Well, why highway autonomy? That’s a little bit of a less complicated domain set to solve for. If you actually think about highway, look, I’m dramatically oversimplifying this. But the 2 rules of highway autonomy are: Stay in your lane; don’t hit anything in front of you. The challenge to developing the highway autonomy is to have a sensor set that enables you to see out 250 meters that — because you’re driving at higher level speeds, you need to see further. That is what our LiDAR enables.
And so when we speak initially of autonomy on the passenger vehicle set, the domain that we’re focusing on initially is the highway autonomy, and we have a unique sensor that is able to see that far to enable it. We think — at Luminar, we have a point of view that full Level 5 autonomy, where you’re doing not only highway autonomy but robo taxis and autonomous [indiscernible], we think that we’re still years away to having a high penetration rate for that. And it is probably the end of the decade at the earliest, but highway autonomy is something that you can solve here in the near term and that we’re working with Volvo to solve.
Now our LiDAR not only enables that highway autonomy, but it enables that next generation of what we call ADAS or what we call Proactive Safety. And so if you know what your environment is 250 meters out in front of you, you’re actually able to get significant improvements in your ADAS systems. A lot of the ADAS systems that are on the road today are vision based. And they’ll have RADAR and maybe some other sensors to help them.
The problem with cameras is that it’s a 2D technology. And you’re ultimately doing a bunch of calculations to come up with an estimate of what the objects are in front of you. With a LIDAR, you’re able to get a very precise ground truth understanding where every object is in front of you. So if I took a picture of that wall there and gave to you that picture, you wouldn’t be able to know precisely where it is. But if you start — you kind of like adjusted your distance from the wall, took a bunch of pictures, you can probably see outside, you can just do it. If I gave you a laser pointer, you could know exactly where that is.
And so our LiDAR system on there. And I think you’ve probably seen some of the videos that we’ve done of our Proactive Safety demos. Some of them have gone viral there. But it allows you to do this ADAS system and significantly improve them and you don’t have any false positives.
Unidentified Analyst
And particularly in low-light situations where camera becomes…
Thomas Fennimore
And then you even take it at night where your visibility goes down. And so that’s what we’re enabling. Now the good news is that our LiDAR enables both that Proactive Safety as well as the Highway Autonomy. So it’s the same LiDAR.
The difference is the software. And the software you can either put on the car at the time the consumer buys it or you can do it afterwards with an over-the-air OTA update. And so our value proposition to our customers is safety drives the standardization for putting our LiDAR on every vehicle. But once you have that hardware on it, you can upsell the consumer to the autonomous package either at the time the consumer buys the vehicle or any time thereafter with an OTA update or unlocking the software. And it’s a business model that really resonates with our customer base.
Unidentified Analyst
And just another real quick from me, and I know Doug has an industry question. But you mentioned the full Level 5, you are working with Intel on their deployment. But there are a lot of internal solutions at some of the robo taxi companies. Do you see that as an untapped opportunity that right now, it’s an R&D project for them? But when it’s commercialized, they’re going to look to Luminar to be sort of right on the road.
Thomas Fennimore
And we’re working with a lot of the robotaxi companies. You mentioned Intel, Mobileye, they’re actually using our LiDAR in their system. We’re working with Pony.ai. But in terms of what we prioritize, it’s this passenger vehicle opportunity, because the volumes are going to be there in the near term. And so we are working with those robo taxi companies.
But our strategy is, the passenger vehicle in that near-term deployment is going to allow us to bring our cost down significantly because of the high scale there. We’re actually going to learn a lot, right? We’re not perfect today. We realize that there’s a lot that we’re going to learn over the next 5 to 10 years as we get our products on several hundred thousand and then millions of vehicles that are going to be driving around the world. We’re going to get a lot of data that will allow us to continuously improve our products.
So that as we get those learnings over the next few years, as we get the scale to bring our costs down. And then as the robotaxi market ramps up, it’s — we’re going to be there with a very low-cost solution that works that has been tried and true tested and ready to go. And so that’s part of our strategy there.
Unidentified Analyst
Sure. So thanks again for being here, Tom. You got into this a little bit with your example of the laser pointer versus stereo cameras. But what would you say is the most misunderstood part of Luminar’s business, LiDAR industry in general? Is it really that customer education is needed, the benefits and how it makes computer vision possible in a way that cameras don’t? Or is there something else that your customers have had trouble grasping?
Thomas Fennimore
I would say, I don’t know if it’s necessarily our customers, but there is a lot of questions we get, particularly with Tesla trying to do a camera only. And look, Elon Musk he’s been very successful in life with Tesla and SpaceX, and so he’s a tough person to bet against from a technology perspective. And he’s been out there with a very public view on LiDAR, which is do you need it? And his argument is we humans drive with eyes, right? We don’t have laser shooting out of our heads to help us drive. And so if humans are able to do it, I can develop a system that can do it as well.
And there is some — our reply to that, I would say, 3 typical replies to that. One is we’re actually wanting to design a system that is much better than a human driver. 95% of the accidents on the road today are caused by human error. And so — we’re not benching marking ourselves to the human driver. We want to create a significantly better system. And so giving the car an additional sensing system to do that is one way to enable it.
Number 2 is the computing power of the human brain is immense relative to like the best supercomputers out in the world, let alone the computing power that you can actually put on a chip to put on the car. And so there are computing power limitations.
And then the third reply is if there is any other company out there that is motivated to make a camera-based system work without LiDAR, it’s Mobileye. They understand vision better than anybody. That’s their core bread and butter business. Their robotaxi system, their autonomous system that they’ve developed has a LiDAR on it. It actually has multiple LiDARs on it. There are LiDARs. And even they admit that you need this additional sensor out there to have full autonomy work.
Unidentified Analyst
Makes sense. I’ll ask one more on financials, too, while we have you here back to the order book. And if maybe you can tell us where you sort of see that financial inflection point coming at a high level? I know there’s not much you can give us, but where breakeven happens, what margins sort of look like in the steady state going forward?
Thomas Fennimore
Yes. So we’re going to update a lot of our longer-term financial outlook at our Luminar Day we’re expecting to have later this year. What I would say to this is we’re really ramping up next year with serious production. We have a new dedicated facility, which is about 250,000 units that we’re going to be bringing online in the middle part of next year.
We’re really going to start to see the significant growth in our business over the next couple of years, and that’s ramping up the volumes, this first wave of customers, the second wave of customers happening shortly thereafter in terms of giving a little bit more precision on what that volume ramp-up is going to look like in some of the margin targets. We’ll update that over — at the end of the year. What I would say is we had a set of projections when we went public a couple of years ago.
We’ve generally been online and meeting or beating those. And I think the longer-term opportunities when you get into the second half of the decade, we’re actually seeing — those are coming together at a much better rate than what we’re thinking about initially, right? We’ve talked about the Nissan opportunity, when we’re successful there. That in and of itself could be a significant part of our business.
Unidentified Analyst
And Tom, that 250,000 unit, is there a capacity limitation through the mid-decade that we should think about? Meaning that 250,000 sort of puts you into sort of a capacity level for, let’s call it, I don’t know, 2025 or so?
Thomas Fennimore
Well, that’s one line, right? So it’s one line at our dedicated plan. At some point in the near term, we’re going to reach the limits of that. And that leaves us 2 options. One is we can add a second line or more in Mexico. And with this dedicated facility, we’ve — we’re constructing that facility in a way to make it very scalable. So that is even literally, we just add a second line, right? It’s — it’s very doable. Or at some point, does it make sense to open up a second facility somewhere.
Unidentified Analyst
Maybe talking a little bit to the business model in sort of the mid to long term. We’ve been mostly talking about hardware, right? But it’s been clear with Volvo and your Sentinel product that you are moving much further into software. So it’s really not only about detection, but object validation, classification. And even it sounds like in the Sentinel product, some forms of driving policy. Make — can you talk about that, because that’s clearly more ambitious than the other sort of hardware companies you’re talking about.
Thomas Fennimore
So there — I would say there’s 3 pillars to our software business, and then I think there’s some additional service — additional service opportunities are available on that as well. So there is an embedded level of software functionality that’s going to be part of any LiDAR sensor that we sell. And I would say the level of functionality that our customers want us to include on the base center like they’re asking for us to do more and more, which is good.
The second level, and that software functionality is going to be part of any unit we sell — perception. And then the second level is perception which is then taking the point cloud and then translating that into, okay, this group of dots is a person, 150 meters out, this group of dots is a tire on the road. You can do some object tracking as part of that as well. So that’s perception. And I think we’ll get our fair share of that business. Because it kind of makes more sense for us to do it once and sell it to the OEMs and every OEM.
Then you’re going to have the full stack, which takes the output of our perception, fuses it with the output of the other sensor and then does the decision-making and path planning, actually telling the vehicle what to do. That is something where I think some OEMs are going to try to do themselves. They’re going to be probably the small or medium-sized OEMs, which are going to need a partner to do that with them. And I think that, that’s going to be the full stack portion of it, that we’re also developing the capabilities to do that. But it’s — I think getting customers to adopt that is probably going to take a little longer than the first two.
Unidentified Analyst
And what about that collaboration when I think about full stack working with the other sensor companies where some of the other modalities, we particularly envision, it’s been more black box. Do you hit — is that going to be sort of touch and go? And are those going to be hard commercial agreements to get data that’s comparable from the vision companies from RADAR so that you can be the full stack?
Thomas Fennimore
We are getting data set from a lot of our initial customers because we need that to continuously our product. You need every data that’s collected, no. But — what’s most important is to get access to those edge cases, right? When is the system failing? Why is it failing? How can you get access to that, bring it back, make the improvements to your products so that the system continually involves itself.
I think what we’re learning on the autonomous driving problem that we’re trying to solve, solving for the first 99% is relatively easy. It’s that last 1%, which is very difficult to solve for. And you need to have a system that works 99.999% of the time. Think about it, right? Having a 1% on the vehicle is unacceptable, you can’t — you can’t not hit 99 out of the 100 pedestrians that you see out there on the vehicle.
And so you really need to get access to those unpredictable events to improve your system. And so we’re getting access to a lot of the data from our customers to continue to involve and improve our systems on.
Unidentified Analyst
And just on the software side, new markets like mapping, I know that’s been discussed. Any concrete sort of progress and, obviously, a universal mapper with — once you have the LiDAR on all these vehicles?
Thomas Fennimore
Nothing to really discuss too much about mapping now. What I would say is a lot of the mapping that has been done today for the automotive market is 2D. You — to do 3D mapping, you need a LiDAR to gather that information, right? You need the 3D data to do a 3D mapping.
And so it is something of interest to us. It is something that we’ve started to invest in. And it’s something that I would say, hopefully, we’ll talk more about here in the future.
Unidentified Analyst
Okay. Perfect. I want to really quickly just open it up to quick — questions if anyone has something specific. If not, I think this is a — very recently, yourself other members of management, it’s very clear that there is a high portion of stock compensation. What caught our eye, as a team as well, as I think in Austin’s specific compensation. He’s almost — I think he’s taking $1 salary. And there’s some provisions that there is very high milestones associated with things like the stock going up 6x.
It reminds us of Elon’s compensation many, many years ago. It’s a very big bold statement. And I know there’s a lot of confidence on the team. Can you talk a little bit about that other than the state? Because it’s — I think it aligns investors with the company.
Thomas Fennimore
Yes. So since we’ve been a public company, Austin really hasn’t taken a dollar of compensation. We might — to be honest, you forget if we pay them $1 or not, it’s — that’s the — line item. But what Austin really wanted to do and I think what Elon Musk did at Tesla, is kind of the foundation for it. But Austin is going to earn $1 of compensation until the stock gets to $50 a share and we reached start production.
And so he really wanted to align his personal compensation package with the long-term success of the company.
Unidentified Analyst
Fantastic. Maybe in the last couple of minutes, what’s next? We talked about this. We’re going to get some of the strategic sort of framework in the December Analyst Day. But any other big milestones?
Thomas Fennimore
Start of production is the big one, right? That’s — making these products to the standards and the robustness that the product needs to go on vehicles being sold to consumers, that consumers are going to drive on the road in real life conditions over a long period of time. It’s not easy, right?
Our team has done a great job of getting us to the — on the verge of getting to that start of production. And when you think about the risks at Luminar, there’s the technology risk, can you design a product that meets all the different specs. I think we talk — we’ve done that. We’ve proven that out.
There’s then the balance sheet risk. Do you have enough cash to kind of monetize or commercialize your product? I think we tick that box. We have the best balance sheet in the industry. There’s then can you attract the team to actually take the technology and the cash and do everything? We’ve built a great team here, and I think you can look at some more recent hires.
Then it’s the execution. Can you bring all of this together? And we haven’t ticked that box yet here at Luminar, but we’re on the verge of ticking that box. And once we actually get these things produced and going on consumer vehicles, that’s a huge milestone for us, because then it really changes the conversations that you have with the customers and you really have with the investors, which is can you guys actually make this stuff? And I can sit here and say, yes, we have a plan, walk you through it. Here’s everything we’re doing.
And — but ultimately, the right way to address that question is say, yes, we’ve done it here, look at it.
Unidentified Analyst
And look at the product on the road.
Thomas Fennimore
And look at the product on the road. And so being able to do that in size and scale, we’re on the verge of doing that. And I think that, that really is going to change it’s going to elevate our game to the next level, both with our customers, their ecosystem partners as well as hopefully investors.
Unidentified Analyst
Okay. Well, fantastic. Well, Tom, I think we’re about running up on time, we really look forward to December. I think that’s going to be exciting and thank you so much for coming.
Thomas Fennimore
All right. Thank you, Chris. Good to see you guys. Thank you for having us.
Be the first to comment