AUR - NASDAQ

Greetings and welcome to the Aurora Innovations Fourth Quarter 2024 Business Review Call. At this time all participants are in a listen-only mode. A question-and-answer session will follow the formal George presentation. [Operator Instructions] As a reminder, this conference is being recorded. It is now my pleasure to introduce, Stacy Feit. Thank you, Stacy. You may begin.

Thanks, Julian. Good afternoon, everyone and welcome to our fourth quarter 2024 business review call. We announced our results earlier this afternoon. Our shareholder letter and a presentation to accompany this call are available on our Investor Relations website @ir.aurora.tech. The shareholder letter was also furnished with our Form 8-K filed today with the SEC. On the call with me today are Chris Urmson, Co-Founder and CEO; and David Maday, CFO. Chris will provide an update on the progress we have made across the key pillars of our business and David will recap our fourth quarter and full year financial results. We will then open the call to Q&A. A recording of this conference call will be available on our Investor Relations website @ir.aurora.tech shortly after this call has ended. I’d like to take this opportunity to remind you that during the call, we will be making forward-looking statements. This includes statements relating to our future financial and operating performance and our financial outlook and guidance, including expected revenue for the 2025 fiscal year, our ability to reduce costs and general expectations beyond that year, the safety benefits of our technology and product, the achievement of certain milestones around and realization of the potential benefits of the development, manufacturing, scaling and commercialization of the Aurora Driver and related services, including relationships and anticipated benefits with partners and customers and on the timeframe we expect or at all the market opportunity, our product's ability to reduce fuel use and emissions, the expected future market size, our expected market share, the efficiency of our validation process, our remote assistance efficiency for driverless operations and profitability of our products and services, regulatory tailwinds and framework in which we operate, expected cash, Runway and overall future prospects. These statements are subject to known and unknown risks and uncertainties that could cause actual results to differ materially from those projected or implied during this call. In particular, those described in our risk factors included in our annual report on Form 10-K for the year ended December 31, 2023 filed with the SEC as amended as well as current uncertainty and unpredictability in our business, the markets and economy. Additional information will also be set forth in our quarterly report on Form 10-Q for the quarter ended December 31, 2024. You should not rely on our forward-looking statements as predictions of future events. All forward-looking statements that we make on this call are based on assumptions and beliefs as of the date hereof and Aurora disclaims any obligation to update any forward-looking statements, except as required by law. Our discussion today may include non-GAAP financial measures. These non-GAAP measures should be considered in addition to and not as a substitute for or in isolation from our GAAP results. Information regarding our non-GAAP financial results, including a reconciliation of our historical GAAP to non-GAAP results may be found in our shareholder letter, which was furnished with our Form 8-K filed today with the SEC and may also be found on our Investor Relations website. With that, I’ll now turn the call over to Chris.

Thank you, Stacy. Reflecting on 2024 and the start of 2025 it's been a monumental time at Aurora as we near the culmination of years of innovation and preparation for our commercial launch planned in April. The word Aurora means the dawn, and we're standing at the dawn of a new era in transportation, one defined by greater safety, mobility and efficiency. Our focused strategy, grounded in safety, positioned for scale and enabled by financial discipline, continues to differentiate Aurora as the leader in autonomous trucking. We've made tremendous technological progress. We're approaching closure of the safety case for the Dallas to Houston launch lane with ARM reaching 99%, and we've also been approaching our targeted 100% API loads commercial launch estimate since mid-October. On the financial front, we've consistently demonstrated strong financial discipline, manage our cash use under budget. We also further strengthened our liquidity position with a successful capital raise last summer and ensuring we have the resources necessary to fund the initial phase of our scaling strategy. And at the start of this year, we announced a 3-way partnership between Aurora, NVIDIA and Continental, solidifying another key enabler to successfully deploy at scale. Our industry is also fortunate to have a supportive regulatory environment. Today, driverless deployment in the U.S. is already allowed by the federal government. And at the state level, under existing laws and regulations, autonomous trucks can be deployed in the vast majority of U.S. states, including our Texas launch market. We're also optimistic that the new Presidential administration's enthusiasm for innovation, safety and a nationwide framework for self-driving vehicles could further support this favorable regulatory environment for driverless deployment in the U.S. With nearly all the pieces in place, Aurora is poised for an extraordinary year ahead. Our Analyst and Investor Day in March 2024 marked a defining moment for the investors who have supported our development journey. We gave them a chance to experience driverless truck rides and a first look at how our driverless trucks navigate advanced road scenarios at our test track. I'll never forget walking up to the track with our investors and analysts seeing the truck speed by entirely driverless. At that moment, Aurora's vision became clear for everyone. As we approach the final steps of development, we're just months away from starting our commercial journey and making our vision a reality. We plan to launch our first driverless trucks hauling customer loads between Dallas and Houston in April. In order to commence driverless operations, we must first close the safety case for the Dallas and Houston launch lane. Our safety case framework is a comprehensive evidence-based approach to confirming that our self-driving vehicles are acceptably safe to operate on public roads. We quantify our progress toward closing our Dallas Houston Launch Lane safety case through the arm, a weighted measure of completeness across all claims of the safety case for our launch lane. We remain the only company in the industry that has provided this level of transparency. As of the end of January, ARM was 99%, up from 97% at the end of October, driven by the closure of a number of software claims. We've made meaningful progress on final behavior refinement and validation and expect to complete the remaining elements of our safety case over the next several weeks. Let's take a look at our rare Surface Street scenario we just recently encountered in Houston. In the video, on Page 5 of our presentation, the Aurora Driver comes upon a large funeral procession, an unpredictable and complex scenario. As the truck approaches the intersection, the Aurora Driver detects a police motorcycle directing traffic at an uncontrolled intersection, which could potentially block the truck's lane. It slows its speed as it prepares to stop and requests input from a remote assistance specialist. Importantly, given our system architecture and strict security protocols, our trucks cannot be operated by remote assistance specialists. All driving tasks can only be executed on board. Within seconds, the remote assistant specialist sees the police officers just blocking cross traffic and not the trucks lane and confirms the Aurora Driver should proceed. With the detection of the active police motorcycle at sufficient distance and rapid remote systems response, the Aurora Driver powered truck was able to continue on its journey without stopping. Adeptly handing scenarios like this helps reinforce our confidence that the Aurora Driver will behave appropriately even in the rarest Surface Street scenarios and the successful interaction with our remote assistance specialist underscores the power of the complete systems performance. Turning to rare construction scenarios. The Aurora Driver is also delivering impressive performance. In the video on Page 6, an Aurora Driver powered truck encounters an upcoming construction zone warning. It lane changes to the left to prepare for the lane closure ahead and then seamlessly navigates a highly complex traffic crossover. As you'll see, this type of configuration redirects traffic to the opposite side of the road, bypassing the work zone with lane separated by barriers and marked with cones. This is a great example of truly skillful performance of known construction zones. Once fully clear of the construction zone, the Aurora Driver returns to its preferred right-hand lane and continues its journey. Strong performance in these types of scenarios support our solid API results. API is another key metric we use to assess the Aurora Drivers performance and commercial readiness. The indicator penalizes the use of on-site support, which will be the most expensive support provided to enable the Aurora Driver. We're focused on driving up the percentage of commercial loads that did not require any form of on-site support, which we refer to as 100% API loads. As a reminder, we do not anticipate that aggregate API will ever reach 100% and even at launch because certain situations will always require on-site support. However, we believe the percentage of 100% and API loads is a strong indicator of our progress and expect this metric to reach approximately 90% by commercial launch. We've been approaching 90% since mid-October, Specific to the fourth quarter, excluding the first two weeks, 88% of loads at 100% API with many weeks exceeding our commercial launch estimate of 90%. This puts us in a strong position for our planned launch in April. During launch, we expect to operate up to 10 trucks commercially starting with one driverless truck and then transitioning the balanced driverless operation. We're deliberately starting with this crawl-walk-run approach as our early efforts will be focused on exercising the full product suite to ensure a seamless launch while demonstrating the value proposition for our customers and continuing to build trust with all of our stakeholders. In the second half of 2025, we'll focus on expanding our product capabilities to include night driving and rainy conditions. Beginning our lane expansion strategy with driverless operations on the Fort Worth to El Paso lane with further extension to Phoenix and increasing capacity of tens of trucks by the end of the year. Growing demand for autonomous trucking underscores the critical role the Aurora Driver will play in addressing industry challenges. As freight volumes continue to increase and shipping distance expands, the rare driver is uniquely positioned to help solve staffing shortages and enable more productive and efficient transport. Aurora Driver powered trucks operating at high levels of autonomy are already achieving best-in-class fuel efficiency, 15% above the industry average. This isn't just a marginal improvement. It's a clear example of how autonomy can deliver tangible value in fuel saving and sustainability. As we work with customers and more deeply integrate the Aurora Driver with their operations, we see the potential to reduce fuel use and emissions by up to 32%, as we discussed in our sustainability white paper published last year. This will help the industry reduce emissions and bring down operating costs. This is just one example of how the Aurora Driver is meeting today's challenges and shaping the future of freight transport. With a mutual focus on sustainable operations guided by safety, Aurora and Volvo Autonomous Solutions, or VAS, continue to make significant progress in our partnership. During the fourth quarter, together with VAS, we launched pilot operations with DHL Supply Chain with the purpose-built Volvo VNL Autonomous powered by the Aurora Driver. We're initially hauling DHL Freight on two lanes, Dallas to Houston and Fort Worth to El Paso. We also continue to autonomously haul freight for our other pilot customers, including FedEx, Werner, Schneider, Hirschbach, Uber Freight and others. Cumulative to date, we have autonomously delivered under the supervision of vehicle operators more than 9,500 loads, driving over 2.6 million commercial miles with nearly 100% on-time performance for our pilot customers. We're also excited to have recently executed an MOU with JB Hunt. As we prepare for commercial launch, we continue to run our Partner Success program, which gives customers the opportunity to more deeply evaluate and assess the Aurora Drive's performance as a final step to move forward with driverless operations. Hirschbach recently evaluated our system, leveraging the expertise of some of their most seasoned professional drivers who collectively represent over 75 years of on-road experience. These drivers have logged millions of miles and have seen everything in the road can throw at them. So understandably, they came into the program with a bit of skepticism. But during their rides, they were blown away by the Aurora Drivers performance and in turn, Hirschbach is ready to go driverless when we are ready. For me, it was awesome to hear how experienced truck drivers get it about what this technology can mean for their industry. You can hear from these drivers firsthand in the video on Page 8 of our presentation. Their shared belief in how autonomous and traditional trucks can work hand-in-hand to improve road safety and transform freight transportation is a true testament to Aurora's mission. Since our founding, our objective has been to deploy self-driving technology at scale our OEM and Tier 1 partnerships with Volvo Trucks, PACCAR and Continental are unmatched in the industry, and we believe position Aurora is the only company capable of deploying autonomous trucking at scale. As I mentioned at the beginning of the call, in January, we further enhanced this ecosystem with a 3-way partnership between Aurora and NVIDIA, my apologies, and Continental solidifying another key enabler to successfully deploy at scale. NVIDIA's DRIVE Thor system-on-a-chip will be integrated into the Aurora Driver hardware kit that Continental plans to mass manufacture starting in 2027. And production samples of DRIVE Thor are coming in the first half of 2025 to start testing. DRIVE Thor will be the core of the primary computer for the Aurora Driver, which we're developing with Continental who will manufacture it. We also continue to make meaningful progress on other aspects of this generation of the hardware kit. During the fourth quarter, we completed the integration of our FirstLight LiDAR chip into a single photonics engine. Notably, the prototype performance is meeting our requirements for our next-generation LiDAR. And in January, Continental and Euro achieved another partnership milestone and are now beginning a sample builds and firmware development. The significant progress we're making towards this generation of hardware is critical as it will unlock true scale on the order of tens of thousands of trucks. While we work toward Continental start of production in 2027, our third-generation commercial kit. We also continue to advance our second-generation commercial hardware kit. We plan to introduce this kit later this year to support our ambitions to scale to hundreds of millions of miles traveled autonomously. This generation brings exciting performance gains. And importantly, we expect it to drive a step function reduction in our hardware costs, which is a critical element on our path to self-funding. During the fourth quarter, we received eight samples from our contract manufacturer, Fabrinet and have integrated this prototype kit into its first vehicle for testing. Aurora has always been a mission-driven company with an immensely capable team, bold enough to dream big and skilled enough to make those dreams a reality. We're on the cusp of our planned commercial launch, a pivotal step toward realizing our mission and our team is more focused and energized than ever with a team like ours impossible becomes achievable. Our extraordinary progress would not be possible without the unwavering commitment from our team, partners and investors. Thank you for your trust. We believe 2025 will be a defining year for Aurora as we begin to reshape the future of freight. With that, I'll now pass it over to Dave who will review our financial results.

Thank you, Chris. Let's discuss our financial results. We have provided a summary on Page 13 of the slide deck for reference. During the fourth quarter of 2024, we continue to demonstrate strong fiscal discipline. Fourth quarter 2024 operating expenses, including stock-based compensation, totaled $199 million. Excluding stock-based compensation, operating expenses totaled $164 million. Within operating expenses, our R&D expenses, excluding stock-based compensation, totaled $142 million. This amount includes $714,000 in pilot revenue, which we record as a contra R&D expense. 2024 pilot revenue of $2.8 million increased 59% year-over-year. SG&A expenses, excluding stock-based compensation, were $22 million. We used approximately $142 million and $611 million, respectively, and operating cash during the fourth quarter and fiscal 2024. Capital expenditures totaled $8 million and $34 million, respectively, during the fourth quarter and fiscal 2024. This cash spend was below our externally communicated target for both the quarter and the year, reflecting our continued commitment to fiscal prudence. We ended the year with a very strong balance sheet, including over $1.2 billion in cash and short-term investments. Given efficiencies we found in the business and cash preservation decisions we have made, we now expect this liquidity to support our planned commercial launch and fund our operations into the second half of 2026. In 2025, we expect quarterly cash use to be within the $175 million to $185 million average range. This accounts for an increase in capital expenditures and continued development of our new hardware programs as we prepare to scale our business. At commercial launch, we will begin recognizing revenue. This will include driverless revenue as well as continued pilot revenue, which up to this point has been recorded as contra R&D expense. With our deliberate approach to launch, we expect our 2025 revenue to be modest in the mid-single-digit millions. For modeling purposes, we expect revenue to build sequentially throughout the year. Revenue recognition associated with our drivers launch will be a meaningful milestone for Aurora, but will have a negligible impact on our overall financials during our launch year. Our focus in 2025 will be on expanding our driverless operations to prove the promise of the Aurora Driver technology. In addition, our team will be focused on key cost reduction levers, including the introduction of our next-generation hardware kit to support achieving our initial scaling cost reduction initiatives. With that, we'll now open the call to Q&A.

[Operator Instructions] And our first question comes from George Gianarikas with Canaccord Genuity. Please proceed with your question.

Hi, good afternoon, everyone and thank you for taking my questions. So I'd like to be optimistic and maybe look past the April commercial launch and just try to glean some incremental insights from you and how comfortable you feel about scaling and this is sort of a multipart question. First, maybe some -- tell us about the confidence you have in lane scaling through -- especially as you mentioned, the operational remote assistance specialists that you have to deploy in one of the examples that you shared? And also from a hardware perspective, particularly as you mentioned NVIDIA, can you -- maybe can you break down in a little bit more detail some of the cost reductions you can achieve from a hardware perspective as you sort of scale the operation.

Yes, happy to, George. Thanks for the question. So on the lane scaling side of this, I'm very optimistic. Right? Part of the reason and are part of the thesis behind going after the trucking application is that the self similarity of the freeway network in the United States is high. Meaning that one bit of freeway looks very similar to another bit of freeway. So if you can drive in one place, you should be able to drive in another. And we're seeing that in practice. Last quarter, I think we shared an update on our ability to rapidly map the route between El Paso and Phoenix and then operated at a comparable level of performance to what we're seeing in other places. So we expect that to roll out relatively quickly. And as we look to '25, I think we'll see some evidence of that through the course of this year. One of the things I would highlight is that -- as I look at this problem, each new lane will become actually easier to roll out. So as we get across the line for the first lane in Dallas Houston, that's where we've built all of the core competence for driving. And then as we move to a new lane, there's a small set of incremental new capabilities that are needed that we layer on top of what we already have. And we've talked in the past about the fact that Fort Worth well Paso, the big difference really is that you have to operate through a custom border patrol station. And we do that today. We just need to validate that work. And that's a modest amount of work compared to what we've accomplished over the last several quarters. So really optimistic and confident in our ability to do that. On the hardware side, we've laid out a strategy where initially, we're going to launch with our first commercial -- first hardware generation that's commercial. We have that in hand, and we can produce that up to small tens of units, right? We don't want to be a manufacturing shop. We've talked in the past about the importance for us of focusing on what we do best in the world. And that's not making widgets. We're really good at designing them, we're really good at integrating them. And so that's where we brought into play our partnership with Fibrinet and you heard that we're already integrating the early artifacts from that partnership onto trucks and beginning to the bring up process to that. And that will support mid-level scaling. So getting it from tens to small thousands. And then we've been planning ahead, as you know, with the partnership with Continental, which truly gets us to automotive scale. -- right? They understand the processes and they understand the manufacturing requirements, and they know how to be part of the just-in-time manufacturing pipelines for our OEM partners. And so with the forethought that we put into building that partnership with Continental and the work we've done in building the partnership with Fabrinet, I have a high degree of confidence in our ability to deliver the hardware necessary to build this business. And with each of those, as you know, as you increase scale in production of a widget, the cost almost certainly comes down because of all of the investment in both the NRE of engineering and the manufacturing costs that you have to spread over the number of units that you produce. And so as we look at this, we have a very high degree of confidence in the cost coming down. We already have detailed numbers on that and seeing really good measure. And we expect to, certainly with the Continental hardware get to that very healthy margin business that we've talked about for a number of years. So overall, when I look at the scalability from a software perspective and the ability to roll across lanes, very high degree of confidence and equally high degree of confidence in our ability to deliver the hardware that will both support that business and do so at a price point that allows us to realize our objectives. I apologize that maybe more than you were looking for?

No, that's great. Just a follow-up on the NVIDIA partnership. At CES, there were a number of companies that talked about their NVIDIA partnership. So maybe if you can go in a little bit of detail as to -- and I think you did this at your Analyst Day, but how proprietary your hardware is relative to others that may be using the same GPU?

Yes. So, the GPU is obviously -- and in this case, we're talking about an SoC, an integrated system on a chip from NVIDIA. That SoC, of course, is an off-the-shelf product. And it's an important part of our product. But the really magic in this application is how that's integrated in the system, how that's designed for manual manufacturability, reliability scale, whether it meets the performance requirements to deliver a safe product. And that's the magic that we bring in Continental bring to complement the NVIDIA magic. So the thing itself, not particularly unique. -- the way that we're using it and the value we're creating around it unique.

George, I just heard from someone via e-mail that there was a bit of a cutoff in the connection on our side and maybe there was a loss in the hearing. Could you hear the answer? I'm not sure if it may have break?

I heard the whole thing.

Okay. Great. Thank you for confirming.

And to whoever it got cut out with it, we apologize for that.

And our next question comes from Andres Sheppard with Cantor Fitzgerald. Please proceed with your question.

Hey everyone, good afternoon. Congratulations on the quarter and all of the recent developments, maybe just to start off, regarding your commercialization strategy, I'm wondering if you can maybe flesh out a bit more your crawl, then walk, then run approach. So obviously, starting with 1 driverless truck and then ramping up the tens of trucks by end of the year. Just wondering what is the best way to think about Q3 and Q4 of this year in terms of those trucks in operation. And also, is there a certain number that you might be targeting for commercial loads and cumulative miles driven by your April date?

Yes. So maybe I'll take the first part and talk a little bit about how we introduce it, and then I'll let Dave talk more about the numbers. So for us, the crawl walk run is really about acknowledging that this is a kind of a change, a step function change in transportation, and we want to make sure that we're able to bring stakeholders along on that journey. As we closed the safety case for Dallas to Houston, we're confident in our ability to operate multiple trucks in that lane. But other stakeholders kind of expect a natural progression. And given that while it will be kind of -- well, not kind of -- while it will be an important moment and kind of a watershed moment in terms of delivering this technology, it won't have a material impact on our financials over the course of the next year. So why not operate in a way that works well with local stakeholders and regulators and just kind of build up in a way that they'd expect and gives further confidence in our approach and collaborative nature. So that's how we think about that. We'll start with 1 and then we'll build to, as you said, up to 10. And then by the end of the year, be at tens of trucks is our expectation.

Yes. And thanks, Andrew, for the question. Let me talk about the last part. It may not be the message you want to hear. But I think for us, what Chris said is important for us, it's about ensuring that we have a great product that delivers on the value that we've promised to our customers and working closely with our partners and our regulators. So for us, we're really focused in on one, demonstrating that we have a commercial driverless product that everybody is going to love. And the second thing is continuing to expand the technicological areas where we can operate in. So going into nights, going into range, going into new lanes, those for us are more important thresholds than the absolute number of miles that we drive because we really want to have a great product first and the miles are going to come naturally. The other thing is that we also want to make sure that we get to our next-generation hardware kit. So we don't want to abnormally drive a ton of volume into the first generation hardware kit as we're waiting for the second. So for us, it's a very deliberate approach. We want to make sure that our customers are completely satisfied with the product we're delivering to our expectations. And for us, technology proof points, we think in '25 is more valuable than a mileage guidance.

Wonderful. That's super helpful. I appreciate all that color. Maybe just as a quick follow-up. So for 2025, you disclosed, you expect your quarterly cash use to be in the range of $175 million to $185 million. I'm wondering if you can maybe give us a bit more granularity as to how to think about CapEx specifically for this year, CapEx was about $8 million in Q4 and $34 million for the year. Just trying to figure out what's the best way to model it for this year?

Yes. So I don't expect a substantial increase in CapEx. There certainly will be some increases in CapEx -- but I think, overall, let me try to describe it in the OpEx. I think we have all the necessary people and operations in place to scale a business we'll have some slight increases on OpEx as we bring in our new programs like our new hardware programs. But I think for both of those, they're going to be pretty modest increases. I would expect to see a modest increase in the OpEx cost and the remainder that fills to that $175 million to $185 million to be in the CapEx. So I think that should be pretty helpful for you for your modeling purposes.

Very helpful, indeed. Appreciate all the color. Congrats again on the quarter. I'll pass it on.

And our next question comes from Scott Group with Wolfe Research. Please proceed with your question.

Hey, thanks. Afternoon, guys. I am wondering, anything specific you guys still need to accomplish in order to be ready for April?

No, there's just bits of work around -- sorry, thanks for the question, Scott. As we look at this, we talked last time about just needing to kind of close out the last set of things. And there's just some work to do, not a whole lot. As we pointed out, we've moved armed to 99%, and that's reflective of the small amount of work that's left to close the safety case. And API is approaching the bar that we have set for 100% trip API is approaching the bar that we set for being able to launch. So we're feeling really quite good about this. We see that in the performance on the road and the quality of the experience. And so we're -- honestly, I'm really pretty jazzed right now. It's happening the way we kind of expected.

I totally appreciate the answer about this year is not about number of miles and trucks and whatever, that makes sense. But as I guess, when do we start to see like a real ramp, meaning ending the year with tens of trucks, does that go to hundreds of trucks next year? Is next year like a step function year -- or is it still more gradual in '26 and then the step function is a little bit later?

Yes. Scott, let me take a shot and then maybe Chris, you can add on to it is going to be a step function here in terms of kind of metrics that you guys are more used to seen. Again, once we have confirmed with and started our rollout and we have demonstrated that we've got a really reliable and great product. And then we're continuing to launch on lanes. As Chris mentioned, we'll be able to launch even more lanes in the subsequent years, we expect to have a step function change in just how many trucks were operating and how many miles we're driving. And the miles driven is going to increase, not just based on the number of trucks that are out there but the utilization of those trucks, right? We expect to, in 2016, really be able to demonstrate a substantial utilization increase versus what people are used to. So we're going to do it both from utilization and truck, but to us is that step function year in terms of really seeing solid financials. And I think we've mentioned before, 1 of our objectives is to demonstrate positive gross profit. in 2016. We're still working towards that. That's not a guidance or affirmation at this point, but that's what we've talked about at our Analyst and Investor Day last time, and that still is an objective of ours. To be able to do that, we have to have a step function in our -- the mileage that we're driving.

And just to add to what Dave said, I really think about '25 as accelerating our advantage. Right, that when we launch commercially, we'll have turned the crank the whole way around for the first time and the momentum starts to build as we close that out, we launched the next lane as we launch new capabilities and features. And it will get the team and the company into the rhythm of deploying the extensions of the product that we need to support our customers and support the scale that we need. It will allow us to definitively answer the questions that George led with today where we have high confidence today. We've got early signal of it, but we'll really be able to put those questions to bed and be able to demonstrate the scalability of this into a market where the technology and the product is desperately needed. So where -- like it's just an exciting and fun time here at the company, and I'm really looking forward to it.

And then maybe just one last one. When do you think you'll start having we'll start seeing deliveries of trucks to third parties, meaning not an Aurora own truck, but a customer owning it?

Yes. I think -- so at the very latest, when we have our hardware kit that's being produced by Continental, that is expected to be a line side install with our OEM partners that will go directly to customers. I think there is a path to do it a little bit before then. But we're going to work -- I think that's going to be more customer-specific and their specific use cases. So, I would say the latest, you'll see a substantial -- essentially all driver as a service, delivering to customers in '27. I think there'll be a migration that happens slowly before that, and that will be customer specific.

And I think this is going to be on us, right, that as we roll out that second generation of commercial hardware at the end of this year, as we are able to demonstrate to customers that this can go to places they need that it does it at a level of performance in practice, not just in theory that they're expecting. It's going to become obvious that it's transformational. And I think for the customers who get that earlier, they'll begin to build an advantage through the experience to gain early on in this next phase of transportation. And I think that's going to perhaps drive earlier acquisition and adoption of the hardware.

Our next question comes from Jeff Osborne with TD Cowen. Please proceed.

Thank you. Good evening. Just had a couple of questions on my side. I was wondering how should investors categorize or characterize success with the April launch and the months thereafter recognizing you're not giving guidance as it relates to miles, but just as you were to look back maybe in the fall of '25 over the prior 4 or 6 months that you had been commercialized? How should we give you a report card, so to speak.

Yes. So I think there's a big binary step, which is there's trucks driving on the road, nobody in them, right, that when I think of -- if I put my investor hat on and shareholder hat on, I think of what are the big risks for our business or what are the big questions about our business. One is, can you actually make the technology work it off? The second is, can you deliver the product at a price point where it's profitable? And then the third is can you scale it to the degree that the business becomes self-sustaining and becomes the exciting opportunity that we all expect it will be? And so by the end of this year, we'll hopefully have laid to rest the first question, right? That's our full expectation. There will be trucks on the road with nobody in them and that, yes, we can make this work. Towards the end of the year, you're going to get a strong signal that we are able to scale this product. We certainly will not be operating at a scale where the business is self-sustaining but you'll be able to see that, okay, this is turning over, they're able to roll out lanes, the expansion works the way we expect. And then we'll start to get the first taste of the hardware that will enable initial profitability is our expectation. So like it's going to be a fun year.

Just two other quick ones. How would you characterize the competitive landscape?

It's ours to win, right? As we look out at the landscape, we look at the partnerships we have, we look at the technology and team we have. We think we're well ahead of the competition and accelerating. And that is just incredibly exciting. When we think about an opportunity where freight in the U.S. is a trillion dollar market and we have the opportunity to win for ourselves, for our customers, a big piece of that. I feel very bullish about our position.

Good to hear. And the last one, there was a lot of questions on the cost reduction potential. You've obviously talked about Fabrinet here in '25 and '26 at the lower volumes and then Continental longer term. Are there any other one or two key variables that you would highlight that are initiatives for '24? Or sorry, for '25 that we should be monitoring as it relates to cost out plans?

So as we've talked in the past about the core kind of economic drivers of cost in our business, we talk about the hardware costs and depreciation of that over the lifetime of the truck. We talk about the cost of on-site support. So anytime somebody has to go touch a truck on the road somewhere that's expensive. And that's why we've been talking about this 100% API trip measure. And then the third is the cost of remote support. How often is a remote agent have to support the truck. And then finally, the cost of insuring these vehicles. So, when I take you to those in turn, we've laid out the path forward even on this call with the hardware and bring that cost down. With remote on-site support as we continue to deliver the product, we expect the reliability of it to continue to increase. And so we're confident in that as we move from or -- just from a hardware perspective, moving from the hardware we've built in-house to the hardware that's built from a contract manufacturer to Continental, we continue to expect increases in the reliability of that hardware. And then on the remote support, we've laid out in our business review -- our Analyst and Investor Day last year what the shape of that curve looks like. And what we said is that in '26, we expect to get to a ratio of better than 10:1 for our -- I think it was '26 we said that for our operators. Into '25, I apologize. Into '25, we expect to get to a point of better than 10:1 for the operator to a truck ratio. And then on insurance, we continue to believe that the insurance market is rational and that as we demonstrate the improved safety and thus reduced loss that the pricing of that will improve. So overall, again, we understand what the key levers are to drive cost out of the -- out of delivering the product, and we're making good progress across all of them. Thanks for the great question, Jeff, by the way.

[Operator Instructions] Our next question comes from Mark Delaney with Goldman Sachs Asset Management.

Given the news in research out on DeepSeek, I'm curious, Chris, if you think there are new training techniques that Aurora could adopt. And more generally, if you think there are ways for Aurora to materially reduce its AI training cost in the future?

Yes. It's been interesting to watch some of the news around rapid distillation of these large models and kind of the implications of that. One of the things that it really points to is something we've been saying for years, which is quality of data matters over quantity of data. So if you take, for example, this is not speaking directly to DeepSeek, but the report of the -- out of Stanford, where they replicated the performance of 01 using a relatively low cost, small model by feeding it just 1,000 examples of quality reasoning from -- I think it was from Gemini in this case and basically got 80% to 90% of the performance, right, much less -- I think the headline of the article was trained 01 for $50 or something. And so this points to this concept that we've talked about for a long time of it's not about just the quantity of data. It's making sure you have the right data, making sure it's high quality and it's really serving the need that you're trying to solve. And that's been our approach from day 1. There's certainly other interesting elements of these large models that we're looking at augmenting what we do already, but pretty bullish about it overall.

Very interesting. Dave, you said your objective is to have a positive gross margin in 2026, but guidance. And you talked about one of the key variables being around utilization and getting more miles in 2026. Maybe you talk about some of the other factors, though, in particular, an update around how pricing is trending in some of the commercial contracts and then how you think input costs and operations are progressing toward that positive gross margin target for next year?

Yes. I think some of the things -- like some of the key drivers in this, if you think about the gross margin and just look at and decompose the cost of goods sold on this, some of the elements that Chris said are really the major drivers of that. And so I won't repeat those per se. But I think the other thing is like how well do we operate if we're operating at terminals, how efficient are we at terminals. How fast can we get a truck onto the road? How many people do we need to support that area. And I think if we look at how Aurora approach is, there's a reason why we've started pilot operations three years ago, right? We're really -- we've really learned a lot. We've hired a lot of experts who understand how to get trucks on the efficiently and effectively. And that's what allows us, and it gives us the confidence to know that we're going to have relatively low cost in terms of the terminal operations. piece of it. And that's really the only other element from the cost side that Chris has already talked about, but we want to keep a special close attention to. When it comes to the revenue side, right, like the higher quality product we have, the more pricing potential we have, right? I can't predict the pricing of the overall market, but I would expect that we are going to deliver a product that is safer is more fuel efficient and is more reliable. And I would expect our pricing power to be pretty high once we've demonstrated that. So we're going to focus on what we can control, and those are having a great product that we can charge at the higher end of the pricing average and controlling our costs. So I think we have a path to get there. There's a lot of hard work to do in the meantime. But certainly, I feel confident that this is the right team to do it.

And I'll just add to what Dave said that we continue to believe that the human labor cost is really the pricing umbrella that we have for operating the trucks, right? This is a super important job. The cost of operating -- or the cost of the driver operating the truck continues to go up as it should, given the risks and challenges these folks face. And that's going to be the next best alternative to the ore driver of these trucks. So we continue to believe that this is too.

That's helpful. If I could ask one last one in. In the shareholder letter, spoke about the FMCSA ruling and how that didn't go your way. You did speak though about some workaround; you mentioned operational one in the near term. I don't know if you could elaborate a bit more what exactly that means and you also refer to a potential longer-term workaround even if the law doesn't change and what that might look like?

Yes. So for those of you who maybe aren't in on this is the FMCSA rejected our petition to allow trucks to have flashing lights on them to indicate when they're stopped on the side of the road. Equivalent to the way that tow trucks do or emergency vehicles do. It's also one of the more dangerous -- but the conventional way to do this is you put triangles out behind the truck. It's one of the most dangerous things that truck drivers do is walk down a freeway, putting triangles down. We believe that the decision to reject this application was wrong, and we're positioning the federal government to reassess this. And we're optimistic that with the new administration that there's a really good opportunity, a good possibility that, that will get changed. That said, that rejection has no impact on our ability to launch commercially we see both operational opportunities to mitigate this and technical approaches that allow us to operate in compliance with the law. And you're not going to get more out of me on that, Sorry, Mark.

I'll pass it on.

Yes. Thanks very much for the question.

Our next question comes from David Vernon with Bernstein. Please proceed.

Hi there. This is Justine Weiss speaking on behalf of David Vernon. So if you look at Slide 13, there are still a few areas where we know driverless operations and autonomous trucking are prohibited. So I'm wondering if you've had conversations with regulators about potential for more sweeping regulation under this new administration that could make it easier for you to enter into these currently restricted markets?

Yes. I think we approach this with two ways to think about it. One is that I think the biggest lever we have is demonstrating the value in other states and helping that grow and impact their economy and that and see the safety benefits of it on the roadway. And then other states, I think, will want that. We also do expect that there's a real chance that the new federal administration will actually push for a nationwide standard for this, and that would help mitigate the somewhat patchwork set of regulations we operate within -- across the United States. And the reason why we have some optimism that for that is that Transportation Secretary, Duffy, in his remarks before Congress actually spent a fair bit of time talking about automated vehicles and talked about the importance of the technology for safety but also for America to be a leader in this space. Ed talked about the value he sees in there being a national framework for this. So we're optimistic that this administration will continue to be a strong supporter of the technology.

That's great to hear. And then I guess another question I'm wondering about is how are things progressing with a shift towards single operators other than two vehicle operators? Like what percentage is single operator? And how should that change by end of 2025?

Yes. Well, I think the missions that are going to matter by the end of 2025 are going to be no vehicle operator. So I think that's going to be really exciting. A lot of our testing will continue to be a blend of a single and dual operator. And it's really about the mission and making sure that we could operate the missioning question in a way that is thoughtful and safe on the roadway. So, we actually don't even internally track the percentage of missions. So unfortunately, I can't share a number on that. But it's really about, again, operational efficiency internally without impacting safety.

Okay. Great. And then just one more if I could throw it in. So you speak about increasing capacity to tens of trucks by the end of the year. So could you maybe help us frame the upper bound of what tens of trucks could actually look like? Like what's the bear and bull case maybe on that?

Yes. I don't know that we have anything additional, Justin to share on that right now. I think, again, I would focus us on the sequential growth in the mid-single-digit millions for revenue. I think that that's kind of a more important thing. I think for us, again, we're really focused in on the execution in the early stage. For us as we continue to launch more lanes, the demand for trucks is going to be really high, and we're going to have to fill that demand with our partners.