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

Thanks for the question. Yes, we are really excited about where we're operating today, right? We have 10 driverless routes approved today. We have 12 routes overall that are active, the last 2 being up and back to from Dallas over to Oklahoma City. These routes are foundational for our 2026 volume -- contracted volume allocation. So we're basically going to take those 200 trucks and operate them on those areas. Obviously, we can't stop with that. We have a build plan and aspirations to grow. We will continue to front-load new routes where we can continue to offer. A lot of that will be based on customer interest, right? Some customers want to go maybe to Memphis or to Atlanta or we added Laredo not too long ago, and that was kind of really strong customer demand. We'll connect some of the triangle, the Texas Triangle. So we're going to go where our customers want us to go. I think Chris mentioned when we're sufficiently generalized and able to go in a lot of areas, it doesn't take long to do that. And so we're really going to follow the customer demand as the key component. I think when you look at the end of Q3, we're going to be operating 20 to 25 trucks. end of the year, 200 trucks, but like we'll have to have more routes, and we'll announce those routes in advance, but you'll have to have more routes to scale the business in '27. I kind of look to where we are in '28. When we start '28, we're going to be operating through a large portion of the Sunbelt and even up a little bit north of that. We expect that, that's going to represent about 60 billion vehicle miles traveled, including the addition of California that recently announced a process to go to driverless, and we have that permitting process. So we're really excited about where we're headed and more to come. We really want to be customer-focused in terms of the next priority.

Great. Our next question, you have more supply capacity than you were contracting for even 2026. I think you have built enough muscle in sales. Thought about what's required transitioning from an R&D lab to a commercial company. Do you have the right people in place, evidence of customer benefits you claim to be transformational?

Well, this is a big question so I'll try to hit most of them. I think the one thing is, I think for 2026 kind of our supply and demand are well matched, right? Like again, we said we're going to have 200 trucks operating. That's what we have. So if we had more capacity, I'm sure we can fill it, right? It's a [ kind ] of you got to balance the supply and demand side pretty well. We are very selective in terms of how we think about partnerships and partnership strategies. It's not just about running a couple of trucks this year. It's about the ability to scale a driver as a service business. And so there might be some customers that aren't as valuable in that area. So we have to kind of segment them differently and continue to demonstrate the proof points. In terms of like the transition of the team, like our R&D team is great. They're going to continue to give us this outstanding leadership position in terms of technology capability, the only company that is operating [ drivers ] and public roads. The sales team, the business development teams, they are growing appropriately with where we're headed. We're adding incremental skill sets. It's not just about the sales. It's about the customer support. It's about the integration with the fleet partners. And so it's a little bit more than just selling the product. This one takes a little bit more integration. In terms of transformational elements, look, we wouldn't be here today if we didn't think that were transformational in terms of safety and what we can provide the industry. So I think that's really -- but if you want a couple of others, we have the ability to double the asset utilization. So let's say, a 200 truck fleet can operate like 400 trucks operating with human drivers in terms of the service. So that's very transformational. And if you think about the cost structure of this industry, it's a tremendous industry. It's been around forever. They're super efficient. This is a technology unlock that can drastically change the total cost of ownership perspective. And that's what transformational technology is supposed to be able to do. And we can do the cost the safety, the efficiency, the asset utilization, there's very few technologies that can hit all of those levers. So we're really excited about where we are and how we're going to go execute.

Well said, Dave. And our next question, what are the limiting factors stopping you from kicking into the next gear and running faster now that you have real autonomous trucks?

Yes, this is exciting. If you've heard us talk, we consistently talk about crawl, walk, run, right? And the first generation was really about crawling, making sure that we could deliver something safely road, really start to pilot value for customers, really make sure we're getting the safety right. We're doing something, as Dave said, it is transformational, the outside world looks at and is surprised that it exists. And so we want to be good stewards of that is [ going ] to add the vanguard of this. We're now definitely into the walking and starting to pick up our pace phase. The fact that we're going to go from a handful of trucks to 200-plus trucks on the road at the end of the year like that is a big deal. It is kind of exponential scaling. And then we expect next year to take another large step in the number of vehicles we have on the road. So we're very excited about this. You're going to see us continue to accelerate. I don't think there is anything to stop us for accelerate. That is -- this is a deliberate pace. We put it -- we're executing at. We're excited about it. We're excited about the ramp that's coming, and we'll look forward to continuing to build the business responsibly and create a lot of value here.

Moving down the list, our next one. How do you cope with difficult to detect conditions such as flooded roads and black ice?

This is a great question. Again, this is where you don't get disc yellow it, right? Like we take the process of putting these vehicles on the road really seriously. We have a whole systems engineering function whose job it is along with our engineering teams to think about, okay, here are the challenges of the vehicle face on the road. Here's how the nominal system responds to it. And then what are the mitigations or the additional safety features we put in place around that. So for our system, when it's driving down the road, it's constantly monitoring conditions. If it gets into a condition where it's starting to feel kind of to humanize it uncomfortable, then it will take an appropriate response. That may be it will slow down to 55 miles an hour from 65 or 75. It may be that if it's appropriate, it will pull off the road to find a safe haven and let weather pass if it's a particularly heavy storm or something like that. Or it may be in really particularly bad conditions, it may be pull the shoulder and stop. you blow out a tire, that's what it's going to do. And so the ability for the system to understand the environment around it, respond appropriately and make sure we've kind of done our work upfront to identify hazards to -- so that we can respond is key to the whole process.

Great. Moving on the list. Will Volvo trucks be solely equipped with your hardware, we see that they are also working with other virtual driver competitors.

Yes. Volvo, again, another great company, proud to be a partner of. They've shared that they expect to be operating drivelessly in Q1 '27 of and then they're going to build to a few hundred driverless trucks, 300-plus driverless trucks next year. We're excited to be the launch partner with them on that. We're looking forward to having our driverless trucks on the road. We don't necessarily expect this to be a winner-take-all market, but we do expect a winner to take most and we feel pretty darn good about our position in that race.

Great. And our next one. Hi, Chris and David, thanks for doing this. Since launching driverless operations, our partners seeing cost savings or other business benefits and was the most encouraging feedback. And looking ahead 12 months, what do you think long-term Aurora investors should be most excited about and why?

Yes. Thanks. Great question. I think the strongest feedback we get is that they want more, right? But you can use a lot of platitudes. You can talk about the elements of it. But at the end of the day, our customers are coming back and saying, "Yes, we'd like to grow the fleet that we're operating with you and ultimately and operate these assets. So we think that's really excited. The pipeline is strong and the customer interest, we actually have more demand than we can supply this year, which is exciting as a company, an exciting place to be. And what we see is this kind of multiplicative factor, right, that as we are able to bring exciting customers, some of the luminaries in the field into the fold to be customers of ours that gets more interest. There's a lot of, hey, that company that I respect thinks this is a useful product. I should check that out. We end up having those conversations and the flywheel kind of accelerates. So really just excited to see the investment that the team has made here in time and the investment that our investment partners have made with us start to translate into real value to customers and ultimately value to the company.

Great. Switching gears a bit here. more of a technical question. What gives you confidence in the simulator accuracy so that Aurora can handle correctly and timely in edge cases, such as a pedestrian getting thrown on to by another vehicle or a multi-vehicle pile up on IC Freeway or police highway chase in wrong lanes, the crucial factor when scaling operation?

Yes. This is a great question, really insightful. A simulator by itself is just a video game, right, that it's pretty pictures and some physics to actually get to the point where you can trust that simulator and ultimately use as part of the evidence we used to close our safety case. We actually have to ground it, which means going out and gathering an equivalent scene in the real world and one in the simulation and confirming that at each level of fidelity that matters that we're matching. We have a process internally, we call it -- I think it's the STAR process where we'll go through and ensure that our simulator is actually fit for purpose for whatever the test we're going to run is. And this is not just kind of an academic exercise. We've actually, at times, found like, hey, there's a gap between what we expected, what we saw in simulation and what we see at the track. And so then that turns -- creates an engineering iteration loop for us to go and figure out, okay, why is that? Is it something in the simulator, has some of the way the system is plug together. We can then iterate and then ultimately, we can close that out and say, okay, now we're convinced that this works. Example, this was for us with a very this is several years ago now. But with a very high-speed motorbike crossing the road in front of the Aurora driver and blowing through a traffic light, there's we observed that the perception system didn't track that cyclist in simulation the same way that it did in the real world. And so we were able to take that, go figure out that actually guess the thing that we found in simulation was actually something we need to work on in the perceptive system, but there was also an issue with the way the simulator is working. And so by closing that loop, we get to a point where we have conviction that, yes, simulator serves these purposes. And then for other purposes, we have to go and do track testing or we'll go and use real-world data from other sources to get conviction in the performance of the system. We have this massive test volume that comes out of this. And ultimately, the processes that we use, we have review to go through to make sure we have conviction. These things actually valuable work.

Great. another bit of a technical question here. How does Aurora handle sudden extreme weather events that exceed the Aurora drivers operational design domain. Specifically what protocols are in place to balance safety with delivery reliability being on time, when trucks must execute minimal risk maneuvers or pull over during unexpected conditions.

Yes. Another great insightful question here. So first, if you ask any of our customers, it's not a bound safety with on-time delivery, it's safety first, right? And we take that philosophy here as well. And so that is the priority, and we put that ahead of time. This idea of being out of ODD, it turns out once the truck is on the road, you have to deal with what's there. Now it may be that you deal with it by minimizing the time that you're in that condition and get off the road as we talked about earlier, much like a regular truck driver does, right? If you are in a monsoon rain in Arizona as a truck driver, there's a point where it's so bad, you pull the side of the road or you find a way on an unpatch and get off the road because it is just too difficult driver. We handle that the same way with the Aurora Driver. We also do proactive avoidance. So part of our -- the Aurora services layer is this command center where we are monitoring real-time conditions. And if we notice this, particularly if there's a hurricane or there's particularly bad pattern of weather, we're able to actually proactively avoid encountering that condition, has the vehicle weighted out and then continue to -- on its way. One of the advantages that maybe is not obvious here is that because the Aurora driver doesn't have an hours of service limitation, it's actually okay for it to weight it out. And it doesn't have to need -- it doesn't have to kind of double down and white knuckle it through the weather because otherwise, it's not going to get where it needs to go before it gets to a hotel or is able to sleep in a truck stop. They can just patiently wait, let the weather pass and then get on its way. And again, given that for our customers, it is safety first and then everything else after that, this is something that we feel very confident in.

Great. Our next one is back to more of a business focus. As Aurora expands from its initial launch lanes, what do you believe will be the biggest bottleneck to scaling from hundreds of autonomous trucks to thousands? Is it hardware production, carrier adoption, OEM manufacturing capacity, regulatory approvals or something else?

Yes. It's also a really good question. I think if you look at the demand side and the regulatory side, we really don't see those [ next ] today. I think we've talked a lot about the demand and the value that we can create and the interest that continues to come. So we're pretty excited about that. If you -- and again, looking at some of the value proposition, things we have, like the 2x utilization and the lower cost, I think -- the regulatory landscape is very favorable right now. If you look at the existing laws and regulations around the books, we can apply or operate autonomously in the vast majority of the U.S. today. And the momentum at least at this point, has been pretty favorable for autonomous technology adoption. So we're excited about that. Certainly, in the past and even today, we have to manage the capacity side of the situation, right? We're building our second-generation hardware set that we've just launched. We had to utilize a partner, Fabrinet. They're a leading contract manufacturer and so they're going to build our second-generation kit. And then I think you all know that we partnered with AUMOVIO, they're a leading Tier 1 partner to build a design engineer, manufacture even finance and service our third-generation hardware kit. So we're super excited about that. We have mutually aligned incentives with them. So -- but there's always an element where you have to balance that capacity and make sure and address the key issues. What's great about like our third gen with AUMOVIO is they're way better equipped to handle supply disruptions than we are today, right? Like they're just a more capable company. They've got a couple of thousand purchasing and supply chain folks. And like that's more than our company today. So like we have a really solid partner. I think the other thing we have to match is we have to match the demand coming in from the kids with the OEM supply of trucks, right? And in the past, this has been our primary constraint. But we've been working with our partners to try to eliminate and make that not a constraint anymore, but just continue to make that an advantage of the overall business. Again, Volvo is going to be launching -- next year, we have a really strong partnership with PACCAR. We have really good supply of trucks from international. So we're really managing those, but really, it's really the capacity that we have to manage the most from a constraint perspective.

Great. Thanks, Dave. With the 200 truck capacity for 2026 fully allocated, what level of commercial demand or preorders are you seeing from major enterprise carriers for the 2027 to 2028 DAS rollout. DAS being driver of service. Our Tier 1 carrier is ready to commit to multi-hundred unit fleets, once driverless operations are proven at scale.

Yes. Well, for sure. First is a pretty good example of this, right? They've outlined plans to purchase 500 Aurora Driver trucks starting in '27 and '28 in the DAS business model. So that's one example. We are working -- as I mentioned before, every partner that we select is with an eye towards the Driver as a Service business. So it's not just a pilot or getting a couple of trucks, it's always with the long-term plan in mind. We're actively negotiating for additional Dver-as-aService commitments today, and we'll be excited to share those in the future. These are large enterprise customers. So we're really excited about where we're headed.

Excellent. Okay. This is a bit of a logistical one. I'm looking for more logistical information. Does the truck back into DAS without human intervention. Who fuels up the vehicles between stops, who performs DOT inspections, has AI and onboard centers replaced manual inspections do these vehicles bypass way stations?

Yes. Another great question. So today, we're really focused in doing the core. This is where the biggest value is for our customers. So leaving a customer gate, getting down the road and getting to the customer destination and delivering the goods there. That's really where the biggest value unlock is the customer. That's the thing that we want to do. We're really focused on drop and hook type operations. So we're not backing into docks today. Ultimately, we'll do that, right? Driving around the yard at low speed is something we can take on, and we don't see as particularly complicated or difficult to add. It's just not the core focus of the value to our customers. When you're in the yard, there's other things that people have to do. Today, of course, like the most obvious one is just connecting the air hoses, right? That's a human thing that we're not putting robot arms on the truck to do or anything like that. So there's definitely these human loop touch points. If I just go through the elements here, so who fuels up the truck? We're actively running tests to try and find ways to work with existing infrastructure. So partners that have truck stops along the way where the Aurora Driver can drive up, notify ahead of time so that the attend notes coming, pull in, have it get fueled, get sent on its way to be able to leverage the infrastructure that exists and to kind of build that and continue to help and accelerate those businesses ultimately over time. Inspection, that's something that, of course, we need a person to do. There's things where you have to go and look and touch parts of the vehicles. So that will be part of one of the roles that will happen at the customer sites is our expectation. And then, yes, we're able to bypass certain waste stations just like people driving trucks are today. But we're also working with law enforcement to be able to transit through waste stations when required to do so by the authorities. So we'll continue to kind of find ways to meet the world where it is today and integrate the Aurora Driver in a way that is seamless and useful for our customers.

Great. And I think we addressed this one a bit before, but if there's anything other detail you want to provide. What is Aurora doing to address potentially driving in heavy snow? It seems now is going to be a major bottleneck for North expansion. Trucks have to run year-round.

Yes, absolutely. And today, of course, we're operating in the Southern state. And part of the reason to do that is the weather is better. And it's really -- and there's this mythology that driving in heavy snow is going to be this whole other thing. And it isn't. The reason why we operate in good conditions today is because -- that way, we don't have to whole boil the whole technical ocean to solve all of the problems. We can actually create value for customers in the near term. When we think about operating through snow, it's comparable to operating through dust or operating through heavy rain, right? It just happens to be white and more flaky and this is where we think things like our multimodal sensor technology, the combination of LiDAR, radar and camera gives us a huge advantage. This is where things like our proprietary FMCW, first light technology allows us to be able to penetrate through and kind of differentiate falling snow from physical things in the world. We can do clever things. We're looking at the traction response of the vehicle. So like there's work we need to do here. But it's -- like I said, we expect to be operating in light snow later this year. And then we'll find the right time to introduce operating in heavy snow in the coming period years.

Great. And we have time for one more. So our final question for today, and it's been partially addressed but partially not. How and where are Aurora trucks refueled during a highway run? Who fixes a flat tire or motor, et cetera, issue on highway?

Yes, another great question. So first, it's important to realize those [ ops ] who aren't around trucks a lot. They have gigantic fuel tanks. 100-plus mile range. And so they can go a long way. So this is not like my normal car road trip kind of experience. And so first, there's that. And some of our customers will fuel on-site because they see benefits to that. But as I mentioned, we're working with partners to figure out how we can actually have the truck pull into truck stops and use kind of used fueling infrastructure that already exists. In terms of fixing a flat tire or an engine issue, this is going to look just the same as it does today, right? That if a truck breaks down on the road today, they're going to call their service support that someone like [ rider ], who we're working with today. We'll come out and bring a technician who can potentially change the tire more likely than not total vehicle to a place to get a repair. So we -- again, one of our approaches to build in the business is to really work in partnership with other great companies and then to work with the world the way it is today and really kind of fit into the same structure, and that has a lot of benefits. It reduces the complexity of what we have to do and it also allows that seamless integration that our customers and partners are looking for to allow us to build the business quickly and get out there and serve as many customers as possible.

Great. Thanks, everyone, for joining us today and for the incredibly thoughtful questions. We hope you found this town hall valuable. We'd also encourage you to tune into our Analyst and Investor Day next month on September 23. We'll be announcing more details on that very soon. We hope you have a great rest of your day.