Aurora Innovation, Inc. (AUR) Earnings Call Transcript & Summary

Great. Thank you, everyone, for joining as we continue the Barclays Global Automotive and Mobility Tech Conference. I'm Dan Levy, I lead U.S. autos research coverage at Barclays. And very pleased to have with us Aurora Innovation. One of the -- not many of them, but a few pure plays out there on the trend of true autonomous driving that have efforts both on the truck side and the passenger car side. We have with us today, David Maday, the company's CFO, on the newer side, yes. And David is going to lead with a few slides, and then we'll go into fireside chat. Anyone that has questions, please feel free to send them over to me, [email protected], and I'm glad to ask your questions anonymously. So with that, David, thank you so much for joining.

Thanks, Dan. Appreciate being here. Let me just pop up on the slides and share a few slides and then give everybody a quick orientation on Aurora. So I always like to start off with a little bit of kind of blitzy stuff, so let me start off with a little bit of a video. [Presentation]

All right. So I hope it gets everybody awake and started. Before I jump into it, just so that my label team doesn't kill me, here's our cautionary statement relative to forward-looking statements. Let me talk just briefly for those of you who might need a little bit more information on Aurora. Look, at Aurora, our mission is to deliver the benefits of self-driving technology safely, quickly and broadly. And we're a mission-driven company and everything that we do is based on delivering to that mission. We think Aurora is really outstanding and really set ourselves apart from some of the from some of the other folks and competitors. And really, I think it's based on a really strong leadership team with unparalleled industry perspective and tremendous deep technical expertise. We have a differentiated technical solution in stack for both software and hardware that was designed from inception to work across multiple verticals. We've got an amazing partnership ecosystem, which is set up to commercialize at scale. We have a really defined path to commercial launch, which is set up to be able to rapidly scale and really focus on the long-term opportunities in trucking and then pivot over into riding lead, and then we have a Driver-as-a-Service business model that underpins attractive unit economics. It's a little bit about why Aurora. When we talk about our technology stack and the kind of the main product, it's what we refer to as the Aurora Driver. And it's really built on a common core approach from a technology solution standpoint that will power multiple verticals, including Class 8 heavy-duty trucks like the Peterbilt 579, even driving other vehicles like Toyota Sienna, which we have on the road today as well. We plan to launch in the trucking market, it's an enormous opportunity, $700 billion, maybe even $800 billion in that trucking market with a huge global opportunity. And that's going to be our focus. And then we can then pivot and take on other opportunities once we've scaled the trucking solution in both drive daily and local goods delivery. I mentioned a little bit about our partnership ecosystem. And really, this is the foundation like Aurora focuses on what we are really good at, which is we have a great, amazing, talented group of people that will focus on the technical solutions for self-driving. But to succeed in a business and commercialize it, you really need a strong partnership ecosystem. Whether it's our OEM partners like PACCAR and Volvo on the trucking side that represent almost 50% of the market; our fleet service partners like Ryder. We've got customers that are helping us get ready to launch, including FedEx, Werner, Schneider, Hirschbach and Uber Freight and some others. Or even our Hardware-as-a-Service partner, this is our partnership with a Tier 1, a capable, really industry-leading Tier 1 partner that's able to set up and build a future Aurora hardware kit that we could scale in the tens of thousands. So really, partnership Ecosystem is really a centerpiece to how we think about commercialization. Anytime we want to deliver a product, we need to create value and part of this is by addressing industry pain points in addition to like safety itself. There is high driver shortage -- or there's driver shortage and high turnover, and we can really provide a stable, predictable, scalable supply. There's hours of service limitations which will unlock for higher utilization and faster freight. There's higher fuel costs. And then there's higher insurance costs, which we can do and effectively address with safer operations as well as better data. When we think about our business model, our long-term business model is a Driver-as-a-Service business model. And that's really structured to be highly capital efficient and generate really high SaaS-like margins. That means that we will operate the driver and the carriers like Werner or a Schneider will continue to operate the carrier business like they've done for decades and are highly efficient at. We consistently publish and track our performance relative to our Aurora Horizon roadmap to launch. Our next major technical milestone is the Aurora Driver Ready milestone. That's when we have closed our safety case, which for us, for our launch lane of Dallas to Houston means that we feel it will be acceptably safe to operate on public roads without a human driver. In addition, we have other work streams related to working with our OEM partners on an autonomy-enabled truck, which will support our commercial launch by the end of '24 as well as continuing activities to be able to operate an effective service at scale, whether it's with customers or its lane expansion or operational elements. When we think about launching, we're going to launch in Texas. But what you can see on this map here is that everything in blue allows for autonomous technology and trucking to be deployed across the U.S. So basically, aside from California and a couple of other states, so all throughout the Sun Belt area, which will be our primary launch market for the next several years. AV deployment is permitted today either explicitly where it explicitly allows it to be or implicitly where there's no rules that prohibit it. We have a strong government relations team, but we're really foundationally built on the approach to work collaboratively with regulators as well as our partners and our customers to share information to build trust to ensure that we have the ability to effectively launch autonomous trucking service that will scale safely. And then most recently, in July, I think many of you know, we had a successful, an overwhelmingly successful $850 million capital raise, public and concurrent public and private offering. And really, we see that as continuing to get strong support from the investor community, it validates our plan. It provides us sufficient cash, $1.5 billion as of Q3 to fund ourselves past commercial launch and well into the second half of 2025. With that, maybe, Dan, I will turn it back over to you, and we can do some Q&A.

Great. Thank you, David. Very helpful. Okay. So let's just start with more of, I would say, an industry question. And I'm guessing you've gotten this question a bit over the past couple of weeks. There's a notable player in the AV space that's going through some hiccups. We talk about in AV these like periodic trough disillusionments where our expectations get high and then it doesn't meet our expectations. And the indication -- the missteps or issues of one player get applied broadly. So what should we make right now of the state of AV, the confidence in AV given this latest set of hiccups from another player?

Yes. I think it's a great question. And yes, I've gotten it quite a few times. I think the sentiment pendulum for AV is probably swung a bit too far in the negative. There is certainly justification for that. There's been players that have exited the market, whether it be Argo, whether it be challenges that TuSimple has seen or Embark. Like there's a lot of challenges. And sometimes it's easy to focus in on the negative. And even I know Cruise has gone through some challenges as well. But I think sometimes we forget a little bit about the amazing progress that's been made as well as the tremendous opportunity that's out there. And so certainly, it's swung really hard, so to speak, in terms of the negative sentiment. But I think Aurora maybe stands out a little bit differently as continuing to make a lot of progress. And I think some of the challenges that others have had, while we can't control these factors, it has helped to validate our strategy a little bit and how we think about approaching it. And really, the AV industry has to build trust, whether it's trust with the public, whether it's trust with regulators, whether it's trust with your partners or whether it's trust with your customers. And our approach from day 1 has really always been about the safety-first approach to development and deployment, right? That's why if you even look at our mission. We like it's consciously, it safely is the very first thing. We really pride ourselves on remaining committed to this and really transparent with the industry. If you think about Aurora and how we operate, we're one of the first company that publishes the safety case framework to tell everybody how we think about safety and how we're focused on development. We're a company that puts out an external roadmap for everybody to evaluate our performance on. We talk about the measures that are important to us to succeed and measure ourselves against those. We consistently work with all regulators in building trust. We've got a great government relations team. But really, foundationally, it's all about our approach from Chris Urmson, all the way down to the software engineer that's coding the system with the same mission, the same focus in terms of deployment. And so I think in the third quarter -- I think it was somewhat validated, in the third quarter, we were invited to present at the future -- or at the House Transportation and Infrastructure Committee on the future of autonomous commercial motor vehicles, like we were the only AV company to present. Chris got an opportunity to talk about it. And I think that's based on building a lot of trust and a lot of transparency. I think we try to be humble, we try to be modest about our expectations and the fact that it's going to take time to grow and it's not going to be an overnight thing. And I think we hear this a lot from our customers, our partners and even the regulators that we're doing it the right way, but your point is really valid. Like there's a lot out there that we can't control, and we really need to focus on being transparent and staying close to our mission.

We've frequently thought of the true gating factor of the pace of AV really as the technology and the ability to pull the driver. But how much do you think right now this is, when you talk about trust and the regulators, how much is it getting the regulators confident? And we know especially in California, which has been a place where you've seen a lot of AV development, you can't do trucks there. So how much is -- are the regulators the gating factor right now?

Yes. So I think -- so trust is always going to be critical. I still think the technological challenge of being able to safely operate a 80,000 truck going 65, 70 miles an hour on a highway is still enormously challenging and an enormous feat. And I think you have to give credit to some of the other players that are operating safely without vehicles -- or without drivers. There's tremendous accomplishment, but you can't do any of these in isolation. You can't go out and build trust and have a shitty product. You can't go out and have a great product and not be building trust and transparent and working with people. So I think for us, it's really focusing on the combination of all of them. You can do them all. In Texas, we have great relationships. We have great relationships with the folks in California. Yes, AV trucking deployment is not permitted today, but we're working with them on a construct that they feel comfortable with, that will allow us to start deploying trucks, hopefully in the future. But we also know that we can achieve our mission, and we can be a very successful, viable business and scale our business within the existing states that allow AV trucking today. And I think it's important to note that the U.S. environment, it's a self-regulation and self-certification environment. So -- in most cases, there isn't something that you have to go and pass and say you're safe, that's why building trust with the regulators and sharing your approach and constantly communicating with them is so important. Because that's part of self-certification is having the trust within the public and the regulators that you are doing the right thing.

One more just on the backdrop. But I ask this because these are like really the big questions right now on people's minds. When I think about what we've seen in AV trucking, Waymo actually, which is, I think many would deem to be the dominant player in AV is actually out of trucking now. TuSimple was another prominent player and they've dramatically scaled back, Embark has gone. So why is it that we've seen is a massive pullback from everyone else in AV trucking? And you're actually like the last man standing or the only man standing really. I mean there's others as well. But...

Yes. I think -- so like I have a lot of respect for everybody that is in the industry and does it. This is a really -- technically, this is an amazing challenge. And it's not something that can be done lightly. I think what I would say is if you look at what works for us, right, which is the combination of the leadership team and the technological approach that started with being focused in on both trucking as well as other verticals. And the understanding that we would need a long-range LiDAR sensing system more capable than what exists in the market today. Or if you look at the way we approach partnerships, or if you look at our business model. If you start looking at all those things together in combination, I think that's what sets us apart from others. And when you look at each of the other players, they probably all have a very specific reason why they may or may not have -- why it may or may not have worked out for them. But if you do an objective criteria like all the important things and check all of those, I think for us, that's what's worked really well. And again, each of them are probably a little bit different in terms of why they didn't happen. But we know we need a lot of people. This is not something that can be done easily. So you need a strong engineering talent. You need a very capable long-range LiDAR system, which we think our proprietary FirstLight LiDAR is really steps ahead of what anybody else could have out in the marketplace. You need strong institutional investors, you need strategic partnerships. You need all of these things to succeed. And I think if you look at some of the other players, maybe they were lacking 1 or 2, and it's really hard. And I don't know that you can lack in any of them and be successful.

Right. okay. Let's talk about your path to commercialization. We'll just quickly start with the safety case and this path to autonomy readiness measure, Aurora Driver. So these are all paths, a key set of points that you need to achieve to get to a point where you can begin commercialization. So let's just start. What are the remaining gating factors to get to Aurora Driver Ready? I mean you've given these updates on percentages, you're now 84% ready. What's the path to the improvement? And then how do you get to -- there was this question of it's the edge cases. And this is actually what we saw in this notable AV case 6, 7 weeks ago. There's infinite edge cases. So how do you -- how are you working through the path to Aurora Driver Ready? And how do you address the edge case?

Yes. So that's good questions. Maybe a couple of questions inside there. So let me try to -- again, let me reiterate, like in the U.S., it's a self-certification is the automotive regulation. So you have to have a way to internally decide how you approach safety and how you determine that you're going to be safe enough. And so for us, that's really built on the safety case framework. It's best practices in a lot of safety-critical industries like aerospace or medical devices. So I think that, as an example, it really shows to it. I think closing a safety case for us, like it means that on our launch lane, we are confident that the vehicle, the truck would operate in an acceptably safe manner without a vehicle operator. And so our first big hurdle is completing the Aurora Driver Ready. And after we do that, then we'll continue to improve the capabilities and the operating conditions that it can do on and scale the business. Your point on like the progress and where we're at, so right, we're at 84%. 84% is a representation of kind of like people hours, right? How much amount of work is still required to close the safety case? And I think sometimes -- and it's kind of a weighted measure. I think sometimes people get a little bit confused in terms of safety case and it's like, you're doing all the other stuff, but you're not doing the hard stuff or these "edge cases" or as you would call it, the march of the 9s, right? And what I would say is that we've really architected our system from the beginning to deal with the long tail. We started out with the assumption that there is not a vehicle operator in the truck, right? And so therefore, we can't rely upon a vehicle operator for any instances and then kind of like continuously improve and hope it gets better. We always start off to address the long tail. And what we're doing is we're systematically evaluating our performance of our system, whether it's like with on-road miles, but also in simulation against all the ways that vehicles crash using NHTSA's collision taxonomy. And that's what gives us the ability to have the high level of confidence that once we've closed the safety case, we will have the confidence that we can be operated in an acceptably safe manner. We're always going to continuously improve the system, and that's going to be like, can we handle different operating design domains? Can we -- and those might be what if we're rerouted off the highway for 2 miles? Are we confident we can do that? What happens is like all of a sudden, there's wind gust of 50 miles an hour? I think generally speaking, even human drivers don't drive with wind gusts at that level. But like those kinds of things are going to be capabilities we always enhance at. But really, the NHTSA collision taxonomy is a really important way to think about how we measure and give us the confidence that we are going to be able to get to that endpoint of the closed safety case.

What other technical gating factors are there? Where do you stand on hardware, software, what are the other key technical elements that need to be achieved?

Well, I think there's always going to be technical -- let me highlight -- let me go through like a few of them. So like from a hardware perspective, we are just bringing up on our trucks now our hardware that we are confident will be the hardware that we use for our commercial launch. And so like from a hardware perspective, we're actually quite a bit ahead. And in fact -- and hardware has a long development cycle, so you've got to be a little bit ahead. And so not only are we confident in the hardware solution that we will have on the trucks for commercial launch, but we're also working on future generations. One of those future generations is supporting the Continental deal, which is going to be our third generation of hardware that Continental is going to help co-develop design, manufacture and actually scale up with us. And that's one of our key enablers to scale up in 2027. And that's really a first-of-its-kind Hardware-as-a-Service relationship. And so we have hardware that's already thinking about the second and third generations of hardware. And so we continue -- like there's never one point where we just stop. We're going to continue to evolve it. That's also going to be the case with software, right? If you think about software, once we get to the Aurora Driver Ready milestone, which is a critical milestone, yes, we will be confident that we can put a vehicle on the roads in our launch site that will be exceptionally safe, but we're going to want to continue to improve the performance. When we talk about performance, we talk about how often -- there's 2 measures we use. One of them is like the measure for the safety case and that's our autonomy readiness measure, or ARM. The second one is our autonomy performance indicator. And that's API. And in that particular case, we're measuring the amount of commercially representative miles that we're driving, where we would need -- we wouldn't need on-site human performance -- support, sorry, on-site human support. That right now is at 98%. It will never be 100% because there's always going to be an instance, an engine failure, blown tire, sensor system gets hit by a stone or something like that and goes down. But the ability to drive autonomously and not require on-site support is always going to be something we're consistently improving upon. How often do we need remote specialists? Are we reducing the number of those? Are we driving in better conditions? When we go to our next launch lane, which is -- which will be Fort Worth, El Paso, what are the unique capabilities that we need to unlock from a technical perspective in that area? So what I would say is that -- from a technical perspective, there's never a point where it's like, okay, it's done and it's ready to go. But if you said like what are some of the ones that we most control, the Aurora Driver Ready milestone is the most significant one. And then we'll continue to work on improving our performance to reduce the amount of on-site support that is required, and that's how we measure our API.

Great. Commercialization, so you've talked about commercial launch by the end of 2024. What can you tell us about the launch? I think it's Dallas to Houston, is that correct?

Yes.

Customers. What is it actually -- are you actually hauling freight? What does it entail? Is it terminal to terminal, et cetera?

Yes. So commercial launch at the end of '24 will look like this. It will look like we're operating on our launch lane, which is Dallas to Houston. We're operating between our South Dallas terminal and our Houston terminal. We will be carrying loads for customers. And these are commercial loads delivering goods between point A and point B, they will drop off at our place, we'll be kind of like a middle-mile solution for them. They'll drop off at a terminal, we will take it to the Houston terminal, they will pick it up and carry it the rest of the way. We're actually doing all of that today. We do it today autonomously. The only difference is we do it with vehicle operators in the trucks today. So we're actually practicing as we intend to launch today. We deliver 75 -- more than 75 loads per week every week today, either on our Dallas to Houston or some of that volume is also in our Fort Worth to El Paso because we're planning to do that as well. So we're going to continue to do exactly what we're doing today. We're just going to do it without a human operator. So -- it will be up to 20 vehicles operating autonomously on a regular basis. And we will then -- after we launch, we will continue to expand both the number of loads that we do between the Dallas to Houston lane, but then we will focus on expanding to the Fort Worth to El Paso lane as well. And so that's kind of what commercial launch looks like. We get paid per -- we'll get paid per mile for every load that we deliver. We get paid today per mile for every load delivered, we just -- it's a contra R&D expense today. But once we go to commercial launch, we'll start recognizing revenue associated with it.

Okay. Let's -- I actually I have other questions on this, but I want to move on because I want to hit on some of the other topics. This is all very interesting. Maybe, could you just talk about the path of scaling, okay? So you're starting Dallas to Houston, you're doing next Fort Worth to El Paso, right? So how difficult is it to scale to other routes, to expand more broadly in the Sunbelt? Is this just a matter of more mapping, more terminals? What are the things you need to do to scale this?

It's a great question. So -- if you think about Fort Worth to El Paso and then what comes from there. So let's talk about the technical element first. So the technical element between Fort Worth to El Paso, what are differences that we might see that we haven't done today because we are going to have to basically close the safety case for that second lane, Fort Worth to El Paso, but we only have to deal with the incremental differences between what we already know we can do and what we would have to be able to do in Fort Worth to El Paso. So today, if I look at Fort Worth to El Paso, there's 2 technical differences that we would have to validate and unlock. And those are the ability to decelerate on a large hill using engine braking because otherwise, you'll burn up your brakes. In the Dallas to Houston, there's no situation where there's a big hill. The second one is navigating through inland border control. Both of those in Fort Worth to El Paso today, we operate autonomously with vehicle operators, but we will need to close the safety case for those who do the final validation and verification, so that it's confident that the system will work as intended. Every -- because the U.S. highway system is very self-similar, there isn't a tremendous amount of new technical capabilities that you need to be able to unlock each and every time. It's mostly kind of like are there differences in terms of weather as an example, like is there more precipitation? Is there more wind? Things like that. And there might be 1 or 2 technical things that we have to do that are related to the road system, but it won't be a lot. The other thing is, and it's a little bit of a misnomer, mapping these new lanes isn't all that challenging, right? Like you take our trucks out driving a couple of times, we've built a map. We used our own internal Aurora Atlas mapping solution, which allows us not only to keep a more predictable and reliable map, but we're also able to update it more frequently. If we have a truck on the road and there's an issue where like there's a rerouting because of a lane closure or something like that, we're able to update and notify and update our map almost in real time so that the rest of the trucks know that like, hey, the right lane is closed, we're going to just be operating the left lane today. So mapping itself, not really all that complicated. Yes, relatively easy to do after a couple of trips. And then what technical features would you have to unlock? And then are there any environmental issues that might be different, like, again, the heavier precipitation? So we'll start out, again, Dallas to Houston and Fort Worth to El Paso, then we will unlock our first lane that exceeds an hours of service for a human driver. That could be -- we haven't announced what lane that would be, but like hypothetically, that could be like going from Fort Worth to Phoenix, right, or going from Dallas to Atlanta. Those are 2 potential options where we do that. We expect to be operating in the Sun Belt all from like if California is closed, we don't need to be --California doesn't need to be open for us to be successful. It's a big market. We would like it to be, but it doesn't need to be. But basically, you can go from California all the way to the East Coast on the Sunbelt. And that is an enormous opportunity for vehicles mile-traveled on a regular basis. The other thing is that I forgot to mention is we are not going to be building out this big long terminal network where everything has to operate between terminals. We are going to -- we start office terminals. But quickly, we're going to start going to customer endpoints. Those endpoints might be terminals that they have or it might be distribution centers or things like that, but most all of the locations where our customers already have sites, it's in the same level of complexity and difficulty that we already navigate today at -- where our terminals are because our terminals are off the highway. You still have to be able to navigate surface streets, go through stop lights and things like that. We don't expect that to be challenging, but we will start going in 2025, we'll start focusing on less terminals and more endpoints -- customer-specific endpoints.

So let me try to squeeze. I know we're over time. I want to try to squeeze 2 in. And hopefully, it's going to be lighting round-ish. No, no, good, this is excellent. Your efforts on Aurora Horizon, on the light vehicle side, just where does that stand? And then maybe you could just give a comment on the pace of cash burn. How you're thinking about capital resources? I know you just did a capital raise earlier this year.

Yes. So Aurora Horizon is going to be our trucking subscription service. So that will be our heavy-duty trucking subscription service. So that's our Class A deliveries with the Schneiders, or Werners, or FedExes, et cetera, of the world. We continue to demonstrate the transferability of our technology to operate in light vehicles like with the Toyota Sienna. But our primary focus to start is going to be on launching and scaling a trucking product. We will continue to show some proof points to the outside world that the transferability is real. We do have Siennas operating on the same hardware and software on highways today and in the airport area in Texas today. So we continue to be very optimistic about the long-term potential to expand beyond trucking. In terms of cash burn, we had guided like $175 million to $185 million through a commercial launch. We have about -- we had $1.5 billion in liquidity at the end of Q3. We expect our liquidity to be more than sufficient to not only get us well past commercial launch, but really to get us into the second half of 2025.

Okay. With that, we're over time, but this was very helpful, very insightful. David, thank you so much for your time. Look forward to seeing how the narrative unfolds.

Thanks, Dan. Thanks for having us. Appreciate you.

Great.