AEye, Inc. (LIDR) Earnings Call Transcript & Summary

Hi. Good afternoon, everyone. Thanks for joining us today. Very pleased to have with us AEye's Chairman and Chief Executive Officer, Matt Fisch, here with us today. Thanks a lot for joining us, Matt. We've seen exciting developments from AEye over the past few quarters and just in the past couple of weeks. And I'd like to start by giving you the opportunity to frame the conversation. I know you have a presentation that you'd like to run us through. So over to you, Matt, and thanks for being here.

Great. Thank you. Good to be here. And a little bit of legal speech here, but quite a bit happening since we visited last year as well. It's really been a momentum surge for us. There's a little bit of history of the company here at the top of this, but really I wanted to draw attention to what happened is really a key decision and a key pivot that we made around the end of '23, early '24. We're seeing a lot of signals from the automotive industry about taking more time for autonomy and integration of LiDAR and we went our separate ways with Conti at that time. And we really just made a difficult decision to kind of back off for a bit and take all those learnings that we had from the customer engagement through Continental and through direct, and really take the time to build a product to new product that we aim to be a leapfrog product in the industry. And that means performance, that means cost, that means size. And that was one of the things [Audio Gap] I'll just shout a little bit the friction grading the LiDAR into the vehicle. [Audio Gap] A new product called Apollo. [Audio Gap] Apollo In March of this year. So we aligned with a new cost-driven manufacturing partner called LITEON. They're a very seasoned provider of automotive. They ship more than 40 million components a year to the automotive industry and their area of expertise is aligned with OPTIS. We had a very good strategic alignment. Just a little over a month, 1.5 months ago, we announced our first major OEM win in the transportation space. And then most recently, within the last couple of weeks, integrated our Apollo LiDAR into the NVIDIA DRIVE AGX platform. One of the things that we've learned about working with NVIDIA is this sort of objective view on performance. And look, these guys are experts in when it comes to vision processing. So they know all the tricks that can be played in the specsmanship and they really aim to provide an objective view of the capability. And I'd invite anybody to go take a look at their LiDAR page, and it's pretty clear where we show up on that list of objective measurements. So all the partners on the right, you'll see we took out the ones that have gone in other directions. And these guys are absolutely giants in the financial, automotive and tech industry, and they've really helped propel us forward here in the last year since we've met. Let's talk about market. So we'll talk about market and then what we're doing about it. I'm feeling, and I think you saw it from some of the other talks as well is that there's growing recognition that LiDAR is essential, okay? I have a quote or a headline there from a recent talk by Jim Farley. And then I always love showing a fun picture about LiDAR versus camera processing. And I can assure you, in this picture here, which was shown at Google I/O recently by the CEO, Waymo, no amount of software is going to make up for the fact that you have a physics problem. You have a dust storm blocking a pedestrian on a road here in Phoenix. And the LiDAR picture on below, you can see the pedestrian clearly, in the LiDAR picture, and it's completely [indiscernible], there's no signal processing or comparison with history that camera image could do to identify that pedestrian. And by the way, I have a file with hundreds of such test cases in it. And look, China is a good signal for us. It's a leading indicator. It's really LiDARs reach mass production in that market. We'll probably maybe ship 2 million to 3 million sensors this year according to [ Yole ]. And it's true that the Western U.S. and Europe OEMs are lagging, but we do see -- we are seeing firsthand that there are on-ramps coming up in the next 12 months. And in the meantime, there's a lot of replacement cycles starting to happen in adjacent industries and also smart infrastructure, smart cities that are bringing a lot of inbound in the LiDAR space. And let me give you an example of what I mean by replacement cycles. There's a lot of pull for LiDAR right now in tolling applications. The way a toll works today -- and if you're an E-ZPass or fast-track user, there's a metal loop in the pavement that tells a camera. When a car is coming by and it tells them take a picture of your license plate and then the magic happens and you get billed for the toll. The problem is, is those metal loops wear out, you have to rip up the pavement, you have to pour new concrete or new asphalt to put those in. LiDAR does a job for a much, much cheaper cost and much cheaper recurring maintenance as well. So LiDAR is the trigger for the camera, and you need to see far because cars could be moving fast. And we've really built a strong pipeline in the last 6 months. If you go back to the previous slide, we went from product launch to first OEM win in 6 months. This is driven by another aspect of our product and our architecture, which is it goes right along with the concept of a software-defined vehicle. More often than not, I would say more than 80% of the time, if we need to change something, it's in the software. And this product that you see Apollo is no different. And this allows us to quickly respond to this adjacent sector demand. I also want to point out something that's very relevant in the automotive space. Veterans here will know that there's an aftermarket in automotive that moves very quickly, a lot more quickly than integration into the assembly line. And we're definitely seeing some pull for LiDAR there in a serious way. Think about instead of LiDAR steering the vehicle or making decisions about the vehicle speed or trajectory, you're driving a truck and it's too tall for the bridge in front of you and alarm sounds, you better hit the brakes and stop. So we're seeing this path where there are -- especially in the commercial vehicle space out there that there's a movement to retrofit LiDARs into vehicles and the volumes are very interesting in that space. It's the case where typically, you want to create an emergency situation, notify the operator who will then bring the vehicle to a stop, for example. Last but not least, we have a very strong cash position now. And I don't think we've actually talked about burn rates specifically in our latest earnings call, but we're over $80 million in the bank now. We've got the lowest burn rate in the industry. This gives us cash through '28 at this point. So while automotive is heating up, we've got what we need and the working capital we need now to take advantage of these nearer-term revenue opportunities in adjacent markets. Just an illustration of that. We changed the names to protect the innocent, but up to 10x lower than pure burn rate. We took 2 lenses here, free cash flow and OpEx. And we really think this paints a potentially promising future for shareholders is that it doesn't take nearly as much revenue for us to give significant returns. And secondly -- look, we've been working -- the company has been around since 2013. You saw it on the first slide. One of the major effort areas for us is to make the supply chain easy, and that's both from the OEM perspective and the final assembly. And with LITEON, it's very easy for them to absorb capacity, repurpose it and the investment that we have to make and the cost that we have to bear in terms of CapEx is incredibly low. And it's not going to go through the roof. It's going to stay low because we have that strategic alignment with LITEON. They have skin in the game and they have that flexibility will give us a longer-term advantage there in terms of cost structure. When I sat through the last couple of presentations, there was a lot more talk about specs and whatnot than I expected. But look, I am going to claim that we do have some advantages here that will be incredibly difficult to match. At least, look, time cures many things. But I feel like we have a very solid advantage here. Let me explain it this way. Everybody knows, you've ever seen a picture from the Hubble Telescope. It's got a big fat lens and people who are amateur astronomer, they have the little telescopes and they say, "Hey, maybe I can take a picture like Hubble. Now they have signal processing software that make it look better. It's never going to match the quality of the Hubble telescope. Why? Because the Hubble built into its design gathers a lot of light photons back. And we've made a decision architecturally that allows us to gather more than about 4 to 5x photons than even competitive 1550 LiDAR design. So we are a 1550 solution, but we have built into our architecture an inherent advantage, and that's what gives us a 1-kilometer range. And I'm not talking about just picking up a dot somewhere at a kilometer. We can see a car, we can identify it as a car at that distance, and we have a huge photon budget. So we can do all kinds of things with that to integrate into different markets. One of the things you see here in the upper right is we plugged Apollo. This is our test needle vehicle. It happens to be a Volvo XC40, but in the roof liner without moving anything in the car. So we didn't have to take anything hardware out of the car to put it up there, it sits right above the rearview mirror. We've been driving this. You saw the first picture around the country and into Europe to show that we perform at 300 meters through glass, okay? That's a hard thing to do to punch a light source through glass because we have so much headroom in our performance. We're passing what's being called now 140-kilometer per hour or 85-mile per hour testing. So the bar is going up on operating speeds. And that's due to this inherent architectural advantage that we have. Our scan mechanism is very unique. We don't have a spinner on top. We use a fairly off-the-shelf MEMS device for scanning, which allows for the infinite programmability. And look, we have a world-class manufacturing partner. These guys know how to do automotive all day. I just want to share a little bit about some of these other applications. it's pretty darn exciting. And some of the pictures on this page are propelled by our -- what's happening with Edge AI. And I'm just going to underscore the importance of NVIDIA here for a second. They're making -- through their platform, they're making Edge AI more reachable and easy to integrate. And because we are a partner there, we're getting a ton of support from them. I mean boots on the ground, making connections for us, connecting software, hardware and the ecosystem to enable these new use cases. As I mentioned, the upper right here is a tolling example. We're able to track speed, size and type of cars. The left one on the side, so a lot of collisions happening in the marine space. This is 400 meters away. And you can -- it's almost impossible to see in the picture, you have that big barge in the background. There's actually kind of a speedboat. Sailing in front of that, we're seeing that 400 meters away. That's exactly what the LiDAR sees. You can actually see as a ship 400 meters away. It really showcases the performance of Hollow LiDAR. And there's a 3-centimeter object. I put greater than 100 meters. You need to see like golf balls on an airport runway so don't get sucked into jet engines and things like this. This is crashed the concord, by the way, some years ago. So we're performing in environments we're seeing golf ball size objects at 3 centimeters up to 200 meters away. And all of this is being propelled by NVIDIA's support for Edge AI and the software platform to support and scale that. It's opening up a whole new dimension for us. Last but not least, and we can jump into the Q&A. Just a quick recap from our earnings call last week. I think we covered a number of these topics. OPTIS is the platform we'll be launching at ITS World Congress at the end of this month. That includes NVIDIA's Jetson Orin ecosystem. We now have revenue on the books. We've got 6 wins and growing -- 6 contracts and growing. We have 30 more in the pipe in deep negotiation. So we really feel like we have line of sight. Next year, ramp, we'll get revenue this year for some of these and the thousands of units next year. And I also want to emphasize that we do have a very large OEM win. We've called out the potential opportunity there. We will be saying more about who that is and what it is this year. And look, last but not least, we still are driving hard on the automotive side. This GM-sponsored initiative called WinTOR. I think somebody mentioned earlier about robotaxis in Sunbelt states. WinTOR is about enabling inclement weather driving in snow and rain, and it's being implemented at the University of Toronto under GM supervision. Look, this is a significant advantage for us for OEM sourcing because Apollo has a feature buried inside this guy that the WinTOR program loves because it provides us a feature in a way to filter out rain and snow. Now we got development work to do there, but the thought process here is that as it plays out, this is going to become a minimum requirement for OEMs. We have this feature. We don't believe anybody else has it. That will give us a leg up as we start to see autonomous driving being enabled in snow and rain.

That was great, Matt. Very insightful and touched on a lot of the topics that I wanted to address as well in the Q&A. But maybe just taking a step back and talking about the adaptive capability of Apollo sensors, being able to dynamically adjust resolution, scan patterns and frame rates. These translate into tangible advantages for customers across different verticals. Could you talk about which customer sees the most value across some of these software-defined functionalities of your sensor? And how does that tie in with maintaining a capital-light business model?

Right. So I think there's 2 pieces to this. One is the value that's provided to the company and our shareholders itself. The fact that we can be working on a given day in a large OEM and another one sitting on an airport runway looking for golf balls or similar size object with only a software update that allows us to use a very modest-sized team of support engineers and field engineers to make that all work. So there's the shareholder and the company advantage piece. It's very, very scalable. I'll give you another example of something that happened very recently from a customer advantage. So we had -- we were approached by a customer that was representing a government entity, like a city level government entity. And they are working on an efficiency initiative where they're hauling a certain payload. It happens to be a grain truck, but -- and they're trying to load balance across the fleet. They have to go through a way station procedure, trucks are backing up at waste stations, so on and so forth. They have hundreds of these across the country. And they tried to deploy LiDAR and cameras to send information back about how many trucks are waiting in the queue, how many trucks are stuck in the parking lot day or night, the nighttime part is really important here, and they just couldn't make it work. We brought in LiDAR. We're able to show them we're able to count things in the dark, just sort of shorthand of where the camera and the radar fell short. And lo and behold, maybe 6 weeks later, they said, "Hey, we have another use case now where we think LiDAR can help. It's not the same as the Graintruck example, and we've got a software update. We're on the ground. And so the customer is going, wow, when I talk to somebody else about this, they said, well, they'll come back to us with a proposal in 6 months. You guys came back in a week with a proposal and a software update, and we're hitting the ground running. So the development cycles and the ability for customers to work in different areas with a software update has been very powerful and will drive scale on the product and revenue side as well.

I think that was a great segue into my next question, which was just around how to maintain financial discipline, and how are you going about allocating resources, given the traction that you have seen over 100 potential customer engagements today. How do you ensure that this funnel converts into long-term recurring revenue rather than one-off projects while also maintaining that capital-light business model?

Right. Well, so you'll notice that Conor is not here today. He's covering something else, but it helps to have a CFO that's absolutely maniacal about this point. Look, one of the things we had to do is put a little bit of structure into this. We have -- our Head of Operations is a seasoned veteran in building products and creating customer pipelines. But we have a methodical evaluation process. It's a 3-phase process where we evaluate the velocity of the customer at a given point and the tax or the burden that it's creating on the company. And we don't let things to the next gate where it consumes more resources potentially unless there's a good financial argument to do so. So Conor and I run what's called a deal desk, and we're meeting now multiple times a week. Thankfully, it's a good problem to have. We're checkpointing this process along the way to make sure that we're not -- that the return on investment of each of these prospects that are in the pipe are paying off for us. So we've kind of flipped that discipline switch over pretty quickly.

Great. That's very helpful. You also have 2 distinct platforms, Apollo and OPTIS, both with NVIDIA integration. How do these platforms compliment each other? And how does this dual platform approach expand your addressable market?

Right. So one is actually a superset of the other. So OPTIS actually incorporates Apollo for the sensing side along with NVIDIA's Jetson Orin platform to run the AI itself, the thinking part of the AI. And then we also have another fourth slot, if you will -- sorry, there's 2 software slots beyond that. One is the perception and object classification software that runs on the Jetson platform and say, okay, this is a person, this is a car. Here's how fast it's going, here's the direction. And then we have a set of plug-ins that sit alongside of it. So for example, interface to a signal controller box on the side of the road. We have a software partner that does that. The announced and will be coming soon. So view OPTIS as a ready-made solution to provide sensing and thinking, which includes Apollo as part of that as well as a host of plug-ins to plug into different infrastructure pieces. Look, we found that, that having that ready-made solution with both sensing and thinking is just really necessary to do business outside of automotive. We just -- we don't have a platform like ADAS, where everybody kind of rallies around that. So Apollo will continue to drive and improve as we need it, but it's integrated as part of the OPTIS platform. Think about seeing thinking. Apollo will always be the same part.

No, that's a great way to put it. Maybe just talking about the NVIDIA integration. I think that was like a key milestone for the company recently. What tangible impacts are you seeing from that integration in terms of your discussion with OEMs, proof-of-concept trials or just acceleration or [indiscernible]

Yes. So look, a few things is once you become integrated into that ecosystem, even though somebody mentioned earlier, I think that OEMs make decisions at the end of the day, look, being part of NVIDIA's portfolio allows them to offer you as a solution proactively. So you're getting that outbound connection with every OEM that NVIDIA goes to. This is amazing for us. We're a smaller company. We can't be at 10 OEMs at once, but NVIDIA can. So we get that outbound contact. And we've definitely got inbound back from OEMs as a result of NVIDIA's outreach program. That's been tremendous. And also just the validation part itself. NVIDIA is a giant. OEMs are just learning to deal with visual bound computing where they're processing images and making decisions based on it. NVIDIA is a rock star in this space, right? So working with them, especially intensely over the last year made us very fit and very robust for ADAS and automotive applications. So you have that outbound marketing channel, first and foremost. And then second, we get true tried and tested by their software validation and testing to make sure that we're very suitable and very capable in the ADAS space, and it actually just help us build a better product at the end of the day.

Why is AEye at the top of that list? Like just your thoughts on that. Is it just performance? Is it a mix of performance and cost? Or what have you heard from them?

Yes. I got a call from somebody who's messing around me and say, well, your letter -- your company name starts with, hey, that's why you're at the top of the list. No. Well, if you sort based on detection range and it's called points per second. It's a measure of resolution. We're at the top there very, very clearly. There are different optimizations for field of view and whatnot. I think somebody earlier mentioned 360 products were 120-degree field of view that more affects the use case versus the performance. Detection range and resolution, we're at the top of the list, hands down.

And again, my congratulations on your recent win with the transportation OEM. I don't think you're at a stage where you can disclose the OEM just yet. But wondering if there's more color you can share on the contract terms in terms of the time line, competitors involved in the process and the extent to which AI software is expected to be used in the perception and driving decisions.

Yes. So this was a case where extreme range was absolutely required, okay? I'm talking about a kilometer range. And so nobody else could do it. That was the thing, right? There are a few claims out there, people said they could, but you actually have to see and identify objects at that range. That made the difference for us hands down. We're already starting deployment. We've already started it. We have people on the ground. Since I will say a little bit more about it. It's more of an aftermarket type application. The volumes are very nice. But it is an aftermarket application, and we are starting deployment immediately. It will continue to generate revenue this year. We expect it to start ramping up more aggressively revenue-wise next year and into '27 as well. And it's just very cool because it sort of bypasses that whole OEM integration process, but it is very definitely mobility and commercial vehicle aligned.

Understood. I think that's great news. Another hot news item was the advanced negotiations with the top 5 automaker. Again, unless you're willing to like disclose the OEM, is there like a way to talk about the type of programs or vehicles that these negotiations are related to? Is it primarily L3 applications? Or is it more towards the L4 and L5 end of the spectrum? And just like anything on what you're hearing in terms of LiDARs per vehicle as well as whether you're replacing an existing competitor that might be on the platform?

Right. So I can say it is a Level 3 passenger vehicle-related program. And I also want to point back to the partnership we had with NVIDIA and the validation and the credibility that, that brought to the table in the discussion. So it was definitely a catalyst in terms of getting deeper into that engagement. And it was one of the things that I had in mind when I said we have on-ramp opportunities in the next 12 months. That's probably all I can say about it.

Understood. Great. Could we just touch on pricing for a moment? Like I know you alluded to it during your presentation, but curious like where are ASPs for Apollo and OPTIS today relative to competition in the comparable segments? And how should we think about ASPs and gross margins at scale in the medium term?

Right. So let me answer it this way. I think it is fairly well known, and so maybe I'm not putting anything new out there. But in order to even be taken seriously or reach any reasonable step forward with an OEM in the passenger vehicle space. For long-range LiDAR, this means driving up to 85 miles per hour in L3, you need to be well below $1,000, well below. And we're there. We're very competitive in that space. In fact, we've already budgetary quoted in that regard and got the thumbs up. And I think is -- look, the automotive margins are going to be tougher than the nonautomotive. Really the trick for us, automotive drives volume. And then in the nonautomotive space, that scale. And it's not just scale of the quantities being driven by the OEM, but our partnership with Lion, they have a ton of supply chain leverage in components that we care about, enable us to compete well in that automotive ADAS space. But the margins are very, very nice when you start selling these in adjacent markets that are used to paying $10,000, $12,000, $15,000 for a LiDAR. We can sell much cheaper than that given the number I just gave earlier. And that's a really good number margin-wise at that point. And it takes a bit longer to ramp up volumes there or maybe not depending on the speed of the OEM space, but the margins are very, very nice outside of automotive.

Got it. Maybe just following up on that. Like to your point, you mentioned the margins were really nice on the nonautomotive side of the business. But we've seen a lot of your peers also start to focus on the smart applications and smart infrastructure market. How have negotiations evolved over these past couple of years? Like I'm assuming it was relatively easier to get in -- get an inroad with a customer 2 years back when you pivoted as opposed to today, are you seeing more competition from some of those peers?

So the nonautomotive space specifically, is the question. This is easy. the market is so large, that's not our biggest problem. I will say because of some of the trade policies that have been enacted, that has put pressure on the competition coming from China. I'll just come out and say that. But it's certainly not the largest thing that's driving our success here. Very, very often, I would say most often, we're running into a customer that is looking to replace legacy technology. And we're not bumping heads with competition as much as you might think. So it's quite nice from that regard.

Awesome. And then like to your point, you mentioned your partnership with LITEON as a key competitive advantage, but scaling also brings risks, one of them being supply chain, the other being certification time lines and just competitive moves. Which risks concern you the most as you go on this next path for the company with 2 development wins and transitioning into series production over time?

Right. Look, we just -- I will never be comfortable. I'm most uncomfortable when I get comfortable. So we're certainly not relaxing, but the cash runway that we have right now is helping us tremendously to focus on building business and driving revenue. This is like I wake up in the middle of the night every night thinking about this, and the company certainly knows since I'm the CEO that this is our most important mission right now is to drive conversions and whatnot. I'm feeling quite good on supply chain resiliency. In fact, we just went through the first stages of a supply chain audit at a large global OEM. And they're asking questions like, okay, we get it, you can assemble in Mexico, for example, great. Now let's talk about raw materials. okay? And then I have the President of LITEON Automotive sitting next to me like, "Oh, yes, we have 2 options. We have raw materials. We already thought about that. We have a Southeast Asia raw materials option. We have a China raw material option. We got that. It makes sure. So look, our partner makes -- helps me sleep at night on supply chain resiliency. They're really froze at this. Again, I said 40-plus million components per year. They're already shipping into automotive. They know how to do that, and we have a really great relationship, so we can rely on them to take care. By the way, they have expertise in things like functional safety, AutoSAAR, ASIL, all the things that are required to integrate fully into the automotive space, and they're certainly helping us there. We're all about converting customers right now. This is our maniacal focus. And I just -- I'm treating it as make or break for us.

That's great. It sounds like a great partnership. Just wanted to check if there was a question in the audience. Great. Maybe just one last question as we wrap up. You've called Q2 as a critical inflection point. Looking ahead over the next year, what milestones should investors be focused upon? And what would you view as the optimal outcome in terms of customer wins or revenue pipeline, when we hopefully sit here again next year at the same time?

Yes. Look, I want to be talking about how many thousand units we shipped by the time we sit down towards the third quarter of '26. Hopefully, that's consistent with what I've been saying about ramping revenue, but that's all hands on deck on that topic.

No, that was all great. Thanks for a great presentation and thank you for doing this. And thanks, everyone, for joining.

Thank you.