TuSimple Holdings Inc (TSP) 2022 Q2 法說會逐字稿

Good day, and welcome to the TuSimple Second Quarter 2022 Earnings Conference Call. (Operator Instructions) Keep in mind that this call is being recorded, and there will be a replay available at ir.tusimple.com following this call.

I would now like to turn the conference over to Ryan Amerman, Head of Investor Relations for TuSimple. Mr. Amerman, please go ahead.

Thank you, Howard. Good afternoon, and welcome to our second quarter 2022 earnings call. With us today are TuSimple's Co-Founder and Chief Executive Officer Xiaodi Hou and interim Chief Financial Officer Eric Tapia. Xiaodi and Eric will review the operating and financial highlights, and then we'll take questions.

As a reminder, TuSimple's shareholder letter and a replay of this call will be available later today on the Investor Relations page of our website. This call is being recorded. If you object in any way, please disconnect now.

Please note that TuSimple's shareholder letter, press releases and this call contain forward-looking statements that are subject to risks and uncertainties. These forward-looking statements are only predictions and may differ materially from actual future events or results due to a variety of factors. Please refer to the risk factors detailed in our SEC filings.

We will also discuss non-GAAP financial measures, which are not prepared in accordance with generally accepted accounting principles. Please refer to the safe harbor disclaimer and non-GAAP financial measures presented in our shareholder letter for more details, including a reconciliation of the non-GAAP measures to the comparable GAAP measures.

I will now turn the call over to Xiaodi to begin.

Thank you, Ryan. Hello, and welcome to our second quarter earnings call. Today, we're excited to update you on the recently announced leadership additions. Continued progress we've made with developing our proprietary IP, newly announced partnerships and efforts we've made on the regulatory front. The next 2 to 3 years are all about bringing efficiency into everything that we do, and we're confident in our ability to deliver advanced autonomous trucking at a commercial scale to do it faster, more efficient and sustainable future on the road.

Before we provide an update on the company, I want to spend more time talking about safety. Yesterday, the Wall Street Journal ran an article that questioned TuSimple's commitment to safety. First, I would say that I am extremely proud of all that we have accomplished and are accomplishing at TuSimple. We are the first company in the world to execute a successful Driver Our test, an 80-mile stretch of highway in Arizona, and we've repeated multiple times since. In the past 7 years, we've had 8.1 million miles of on-the-road testing, and precisely one incident that TuSimple's truck is responsible for. In this incident, an error occurred when a test driver and safety engineer tried to reenter autonomous driving mode before the system computer was primed to do so and the trucks swerved, making contact with the highway barrier. No one was hurt. And the only evidence of the accident were a few scrapes and some minor damages on our truck.

I want to be very clear, while this was caused by human error, as the CEO of TuSimple, I take responsibility for it. At the end of the day, these kind of traffic accidents are what we are in the business of trying to eliminate from our roads to build a safer, more secure and more reliable freight delivery system across America.

When this incident happened, the very first thing we did was ground the entire fleet and began an independent investigation to determine what happened. After doing so and determining the cost, we upgraded all of our systems to make sure this could never happen again. There are some who have questioned how this could have happened in the first place, that's fair, but it is also important keeping this perspective. Collectively, our engineers have spent millions of hours building a system that has caused one incident in all 7 years of operation. No one was hurt. No property damaged. And we reported to the appropriate government agencies as we should have. Considering that NHTSA reported 500,000 large truck accidents every year, I think we have a pretty good average and that our technology is working.

With all of this in mind, we are a company in transition as we scale out from the research and development to commercialization. This also caused some questions the changes I've started making as a CEO and whether we are prioritizing safety. But here is my commitment to all of you, the stakeholders of TuSimple, we will not achieve commercialization until we are sure we can do so safely. Every decision I make starts and ends with evaluating the safety of our company technology. We encourage our incredibly talented team to raise concerns if they have them and come forward with the idea of how to do it better. We have a strong safety record, and we will continue to make it as strong as possible.

Now moving on to the quarter. In June, we announced several leadership changes. First, Pat Dillon, our former CFO, decided to leave the company to pursue other opportunities, has played an instrumental role in setting up TuSimple to succeed, and I wanted to personally thank him for his contributions.

Eric Tapia, who you will hear from later in the call, has taken over the role as interim CFO. Eric has been serving as our Global Controller and Principal Accounting Officer over the past year and came to us from Grainger and prior to that, KPMG. Eric's skill set fits very well with where TuSimple is in our life cycle, and we are fortunate to have him step into the CFO role.

We also announced a few additions to our leadership team. Isabella Zhou was promoted to Chief HR Officer. In this role, Isabella will be responsible for organizational planning and ensuring that we identify and develop and retain top talent.

Ersin Yumer was promoted to Executive Vice President for operations. In this newly created position, Ersin will oversee our AFN, working closely with both our technical and operational teams to ultimately support Driver Out operations and continue our journey towards commercialization. Ersin joined us as Senior Director of Machine Learning and was most recently Vice President of Algorithm. This as well as his years of experience in the AV industry make him a perfect fit for this expanded role.

Lei Wang was promoted to Executive Vice President for Technology. This is also a newly created position. In this position, Lei will take over many of the responsibilities I had as the CTO, thus, freeing me up to spend more time on the overall strategy of the company.

We are fortunate to have such talented individuals within the organization. These organizational changes are designed to help us achieve the vision that we have for TuSimple. At our Investor Day in May, we discussed that we are in the process of expanding our operational design domain from Arizona to Texas and from nighttime to daytime. A critical component of this is building out our physical infrastructure in Texas.

Last quarter, we announced additional terminals in the Texas Triangle. And our team is busy executing on our plans to continue the build-out. What is critical as we migrate to Texas is to ensure we have the ability to create density and scale in our network. This requires a continued investment in our physical infrastructure, trucks and technology and further developing strong partnerships. We intend to further validate the commercialization opportunity by utilizing real-world commercial routes to continuously improve our cost per mile metrics and continue our journey towards commercialization.

Next, I'd like to talk about our continued success in building a valuable patent portfolio. During this past quarter, we expanded our portfolio with 37 new patents. As you have heard me say, and as I will continue to emphasize the next 2 to 3 years is all about bringing efficiency into everything that we do. So let me provide you a few examples of how we're doing that with our patent portfolio.

The first example I'd like to highlight is the patent that covers automatic traversal of routes that lowers fuel consumption to reduce costs. Fuel is a major cost of trucking operation, accounting for roughly 1/4 of the truck operating costs. We've already shown our AV technologies over 10% more fuel efficient than a human driver. Route optimization on top of improved fuel efficiency has a potential to lower fuel cost even more for our customers.

Next, we were issued 2 patents dealing with sensors. The first patent covers technology to calibrate sensors while the truck is moving down the road. The second patent covers sensor cleaning techniques for frozen solids. It is hard to overstate how critical sensors are to any AV platform. Technologies that improve the reliability and reduces maintenance costs is critical in reducing the overall operating cost of TuSimple's ADS platform. We will continue to invest in further expanding our IT portfolio.

In July, we announced our partnership with Hegelmann Group. Hegelmann is the major European transport and logistics provider, and we're excited to be part of their expansion plans in the United States. We believe that this additional partnership is another validation point of the value our purpose-built Level 4 truck will bring into the trucking and logistics market. We look forward to work with Hegelmann and our many other partners in the logistics ecosystem as we develop and commercialize our technology.

Last but not the least, I'd like to spend some time on a regulatory update. We're happy to announce that Tom Jensen has joined the company as the Vice President of Governmental Affairs, to lead our efforts in this very important area. Tom comes to us from UPS, where he spent more than 20 years lobbying at both the federal and state level on various issues impacting the transportation industry. We're happy to have Tom on board. Welcome.

While we are on the regulatory topic, in Q2, we continue to work on building now the state legislative and regulatory environments favorable to commercial AV operations. With Kansas and West Virginia enacting legislation granting Level 4 truck authority, this increases the number of states that permits Driver Out deployments to 28 while 44 states allow Driver In operations.

Additionally, I'm proud to be the signatory on the recent letter sent to Governor Gavin Newsom of California from leaders in the autonomous truck industry. As you know, California [critical home state] currently does not allow for AV truck testing despite a 2012 bill that was signed into law intended to begin the rule-making process. We are pushing for a common sense regulation so that California joins the other 44 states in the country that allows for AV testing.

With that, I will turn the call over to Eric to discuss our financial results and updated guidance.

Thank you, Xiaodi.

Beginning with our reservation program. This quarter, we added 10 new truck reservations ordered by a long-time partner. We're excited to continue working to bring our world-class AV technology to our partners. Our reservation total at quarter end stands of just under 7,500 trucks. We continue to see significant interest from our fleet partners and continue to hold the standards for reservations. We only work with partners capable of implementing our AV technology and those that are ready to make a financial commitment alongside the reservation.

Now shifting gears to our financial results for the second quarter of 2022. We reported $2.6 million of revenue in the quarter, an increase of 70% year-over-year and 13% sequentially. Compared to last quarter, we have kept our revenue fleet relatively flat as new trucks are prioritized and deployed for Driver Out testing operations. The year-over-year growth was driven by revenue models as well as pricing. Sequentially though, revenue was primarily driven by better utilization of our existing fleet while pricing was flat.

We expect to keep our leased and owned revenue fleet flat over the course of '22, and we'll continue to improve utilization as needed. That said, I think it's important to share our view on revenue models for the next couple of years. As we gradually transition to Driver Out commercialization, revenue models will be generated as we support our technology needs. This will essentially validate and advance our technology as we support our customers through either pilot or revenue-generating programs that showcase the benefits of our Level 4 technology.

Moving to expenses. We spent $86 million on total R&D this quarter, including $22 million of stock-based compensation. This compares to $76 million in the same period last year and $78 million in the first quarter of '22 or up 9% sequentially. Excluding the impact of stock-based compensation, R&D increased 4% sequentially as we continue to invest in our technology, R&D assets and core tech talent. We will continue investing in commercialization of our world-leading technology and key talent while managing our R&D dollars as efficiently as possible.

On the SG&A front, we spent $22 million during the period, including $3 million of stock-based compensation. This compares to $44 million in the same period last year and $32 million in the first quarter of '22 or down 32% sequentially. As a reminder, this quarter, we lapped the impact of stock-based compensation related to our IPO last year, which explains some of these deltas. Excluding the impact of stock-based compensation, SG&A was down 19% sequentially. During the first and second quarter of '22, we had some favorable noncash accounting true-ups in SG&A of approximately $2.6 million and $4.4 million, respectively, which we do not expect to repeat in future quarters. Excluding these items, SG&A was down 10% sequentially from improved cost leverage. We continue to be highly disciplined in controlling our SG&A spend and prioritizing the spend that really contributes to our technology and commercialization objectives.

Our loss from operations was $111 million in the second quarter of '22 compared to a loss of $121 million in the same period last year, and $112 million last quarter. Our adjusted EBITDA loss in the second quarter of '22 was $83 million, which compares to $66 million in the same period last year and $80 million last quarter. In the second quarter of '22, we had $3.8 million of CapEx, primarily related to equipment and facility investments. We ended the second quarter of '22 with a cash balance of approximately $1.16 billion, a decline of $81 million versus the previous quarter.

Now turning to guidance for 2022. We are updating our full year guidance. While our revenue guidance remains unchanged, we have updated our guidance for adjusted EBITDA, stock-based compensation, CapEx and ending cash balance. This was done to reflect our increased focus on expense control and disciplined capital deployment. We now expect to finish the year as follows: our adjusted EBITDA loss to be between $360 million and $300 million -- $380 million versus the prior guidance of $400 million to $420 million. Our stock-based compensation to be between $100 million and $120 million versus prior guidance of $155 million to $175 million. And this is a result of hiring slowdowns and stock price impacts.

We're also reducing our CapEx guidance by $10 million. This is primarily driven by a reduction of our investment in administrative facilities. Lastly, we expect to end the year with approximately $950 million of cash on the balance sheet versus the previous guidance of approximately $900 million. While we do not intend to issue our '23 guidance until we report our fourth quarter results, I think it's worth pointing out a few important items as we move through the second half of '22 and into '23.

First, we are in the process of upgrading most of our older trucks to the newest AV hardware technology. This process will continue through '23 and together with the physical infrastructure and our partners, will help us create the network density that Xiaodi was referring to earlier. Second, while we plan to introduce some new trucks into the fleet, our ability to add a significant number of trucks is difficult given the challenges in purchasing new or even slightly used trucks. Lastly, we plan to continue to invest and add terminals to the AFN, primarily and around the Texas Triangle. Our intention is to do this in a capital-light manner, partnering when possible. We look forward to updating you on our efforts in the coming quarters.

I'll now hand it back to Xiaodi for a few last remarks.

Thank you, Eric.

The progress we have made over the last 2 years is significant. We have a built a first-class team from the ground up and created a safety-first culture to develop our industry-leading technology. I am proud of our work at TuSimple and remain completely confident in our progress to deliver advanced autonomous trucking at a commercial scale to build a safer, more efficient and sustainable future on the road.

With that, we're ready to start the Q&A session.

(Operator Instructions) Our first question or comment comes from the line of Ravi Shanker from Morgan Stanley.

Xiaodi, thank you for the details on the accident. Can you share some more details on what changes you have implemented since? Were they on the technology side? Or were they on the operating procedure side? And also, would you like to refute any details in the WSJ story that came out yesterday?

Yes. First of all, we have an overhaul of our human machine interface so that we are guaranteeing that the human error won't cause the machine error that's on the machine side. And also -- we have also tightened up our training procedure to make sure that all of the operations are compliant.

And in terms of the other aspects of the article, I think there's a lot of misunderstanding of it, especially regarding to our compliance when it comes to reporting this incident to the authority. I wanted to make several points to be clear in here.

First of all, we reported the incident to the appropriate authority and has been in the public domain for more than a month on the NHTSA website. And the second is that the truck scraped a median, resulting in no injury and no property damage. There were no requirements to contact the law enforcement. But we did report its incidence to the Arizona Department of Transportation and NHTSA on time.

Got it. And so there has been no follow-up? Like there's been no changes in your ability to run your trucks or do any testing in [the media aftermath], right?

No, no. Neither the FMCSA nor the NHTSA has asked us to do any changes. And as part of our own review process, we actually, as I said, implemented a lot of solutions to prevent this type of thing from happening again. And our new fleet with the upgraded software has since been on the road.

Great. And lastly as a follow-up question, I think this patent that you just received on the frozen sensors is an important one because I believe that was a major, if not the only, obstacle to you actually running your trucks in the snow. Is that the case? And do you feel like you're ready to start commencing some snowbelt tests this winter?

This is a complex issue that we're facing to running in snow and ice. We are actually developing the sensor cleaning technology, and this is going to be part of our solution that we have received the patent from. But as of today, we're still developing this full solution.

Our next question or comment comes from the line of Scott Group from Wolfe Research.

I was wondering, given some of the safety concerns out there. I know you guys haven't given data on disengagements in the past, but I thought maybe appropriate, anything you can share in terms of where that trend is and where it's been.

I think we -- in the past, we don't explicitly give out the disengagement data. I don't know where -- what exactly of the data are you referring to?

No, I'm saying you -- I know you haven't given it in the past, but I was wondering maybe if you wanted to make some comments on -- just given some of the safety concerns that are out there, maybe if you want to just give some high-level comments around where you -- where that metric is today, where it was, what kind of progress you're making maybe...

Okay. First of all, I think the lease engagement is not a very useful metric for marking the progress of technology of autonomous driving. And second, our incident, specifically on April 6 is not a disengagement. It's actually -- the issue happened at the engagement phase of the system where the system is not ready.

Is there a better metric that you think is worth sharing with us?

I would say that overall -- the overall metric would be that is tangible and needed to read will be cost per mile, but that only covers part of the system. As you know, the system is very complicated. I don't think there is good, easy to read, a numerical metric to cover up the -- all of the safety part of it.

Okay. And then just a question on the truck reservations were -- in the quarter were flat. Are you in -- having dialogue conversations with customers? Do you think that the truck reservation count starts rising again in the near term? Or is that less of a near-term focus?

That's a great question. I would say yes, we are still engaging with our partners. Our pipeline for potential orders is pretty healthy as you know. For me, we're definitely focusing on the customer that have the capabilities, right, to introduce our technology into their network in their fleets. So that in and by itself tends to be the filter.

Just as a reminder, some of these orders for purpose-built Level 4 trucks are -- they won't hit the market until several years from now, right? But again, to answer your original question, yes, we're still seeing momentum. We're still engaging partners and there's interest, and the pipeline is healthy.

Our next question or comment comes from the line of Ken Hoexter from Bank of America.

Xiaodi, maybe just take a step back and talk to us bigger picture, what's the status of commercial operations? It seems like you're expanding from Arizona to Texas. What is changing in the base case here? Is there a time frame? You had talked about getting more trucks and operating retrofit and getting that up and running before the fully built, purpose-built by Navistar in which you had pushed out from '24 to maybe '25, '26. Maybe just take a step -- what is the status of rolling out that commercial capability in '23, '24? What are we waiting for now?

Yes. We are -- the team is not waiting for anything. The team is super busy every day. And we have actually several aspects of it.

First, the safety. We're actually having a higher target when it comes to everything about like the safety design and the system engineering side. So we're basically having a more tightened safety standard and the validation process that goes through the -- every corner of the system of the engineering part.

And the second is hardware. We're actually introducing new components. For example, the new type of sensors and the more reliable, next-generation actuation components. So basically, we're rebuilding the next version, which is getting closer to the production truck, but still it's like prototype type of chassis and the hardware team is building on that.

And third is that the software and algorithm team is trying to, on one hand, meet the safety standards set forth by the system engineering team, and on the other hand, we utilize the new sensors and new capability brought the new OEM and Tier 1's -- sorry, no new Tier 1's components to run. And on the other hand, the operational side, we are actually seeing a lot of little issues -- not really issues, like little opportunities here and there for improving the fleet efficiency. And that is exactly what we -- I said earlier, by identifying the scale or identifying the routes and populating more trucks into a smaller network operation, we have a long to-do list to do. So that's the thing that we're doing right now. And now we're targeting at the end of next year, so that you will see some strong results in Texas.

So it sounds like you have to rebuild -- or you are rebuilding your entire truck specs now. Did I hear that right as far as getting new upgraded sensors on each part of the truck? Is that why we should not expect any change in trucks and service? I mean it seems like you're getting some...

No, we're like -- oh, sorry. It's more like an upgrade. Considering all of the Tier 1 hardware, they have new versions and the new kind of LiDAR and the new type of cameras, it's all like an upgrade.

So upgrade of existing versions, but I thought you had to start handing in your final specs to Navistar at this point to be able to get that commercialization? Or is that pushed out as well?

No, no, no. This is a different thing. We actually have a lot of older trucks that we've purchased several years ago, right, and we have to upgrade them so that they can be aligned with the newest version of the hardware and software.

All right. My last one, if I can sneak one more in, is you lowered your spend targets. I just want to understand, we're getting no more additional miles in the interim. What's the process? Again, I'm going to go back to the first question. What's the process between here? Is it just waiting for Navistar to get the commercial built? Or is there a reason why you don't want to put more trucks on the road, you don't need more or you're getting enough info maybe round this out there?

I think as long as we can mobilize all of our truck assets, our older version of the truck, that is purchased like 3 years ago and with pretty old hardware, and we upgrade them. And then we have a decent fleet, all of them are running the cutting-edge software version. We're happy with the testing volume of this fleet.

And just to chime in, and I'm going to refer to some of my remarks. We are -- we're also transitioning our entire fleet. And when I say entire, when you think about our trucks, we have revenue trucks, revenue-generating trucks, hauling freight, but we also have testing trucks that are solely used in our testing facility. So what happens from now through the end of the year and beginning of next year is we're leveraging the entire fleet, right, to certainly -- not necessarily to optimize revenue models, but really to focus on the efficiency, scale and bringing true cost per mile reductions by testing the technology in real-world scenarios.

I think that, for me, is a pretty strong proof point of our industry. If we can really show true economic scale with trucks that we control in trucks that are really hauling freight in the real world, that will be a pretty impactful event. So for me, that's the focus. Even though, yes, we reduced our targets, we are definitely investing in our fleet. We're investing in our technology, certainly in our talent as well. We've been very frugal in other aspects of our business, like administrative facilities. But when it comes down to the core assets, that will, again, show that commercialization opportunity, we're definitely investing the dollars.

I guess one last thing. Somebody just brought up the China sale update. There's no update on the sale process. Is there just because -- there was no mention of it.

There's nothing I can share right now.

Our next question or comment comes from the line of Colin Rusch from Oppenheimer & Co.

So I'm curious about talent acquisition here. You put together a high-quality team that's pretty robust, but in terms of attracting talent and maybe upgrading in certain areas, I'm curious what those prospects look like for you at this point.

First of all, I think we are pretty proud that we do not currently have a hiring freeze. We -- however, we have slowed down hiring from what we have planned at the beginning of this year. And we are very judicious when it comes to backfill of the vacant positions. So overall, we are becoming more careful given the general economic situation right now, but we're not shy in hiring the top talents.

Okay. And then in terms of some of the vendors that you guys are working with, there's certainly some technology development that's still going on in terms of sensors, sensor fusion. Can you just give us an update in terms of how you're seeing that progress relative to time frames and commercialization that you guys are targeting with some of those elements that maybe aren't completely finished at this point?

Yes. I think the most exciting thing that I see from the sensor development is actually not about the sensor capability development, but rather the sensor reliability development. And this is something that we need exactly to conduct more Driver Out brands and in the very end, commercialization of the system. And this is very encouraging. You see many of the components are maturing from A samples to B samples. And the time line of these suppliers, they actually pretty -- get pretty clear and they stick with their time line, which is also encouraging to us.

Our next question or comment comes from the line of Todd Fowler from Key.

So Xiaodi, on the path to commercialization by the end of '23, you kind of have 5 milestones laid out in the newsletter tonight. Can you give us a sense, as we sit here kind of looking at the second half of '22 and then into '23, what's realistic here for the rest of this year and then what we should expect into '23?

Sure. Yes. From all of those metrics, I would say that removing the support vehicles is the biggest -- are the biggest tool. And in order to be able to remove support vehicle, we need to upgrade our hardware so that we can improve the reliability just as I was -- as what I have just said that the sensors are getting more reliable, and we're very happy that this happens as we planned.

And we are currently building this new type of trucks. That is taking -- normally, it's taking a very long time, but we're doing our best to shorten the time. We do not anticipate -- therefore, we do not anticipate to reduce or remove the support vehicle in this year because probably by the end of the year, we're barely finished the buildup of the hardware. And while at the same time, the system engineering team working with the algorithm team are working on strengthening the system from every corner. This is what we're doing right now. And this is also what you will see in the later update, the -- on the plan, the Union Pacific Driver Out commercialization run. That is something that we will be basically having to show as intermediate stop of our strengthened software for this year.

So is it fair to ask about, I mean, what percent of the fleet needs to be upgraded and what percent of the fleet has been upgraded? Because when I square kind of the comments that you made to Ken and Eric, your comments at the end of the prepared remarks, it really sounds like that this is very critical for '23. So is that something that you plan on sharing where you're at with the fleet upgrade? Or how can we kind of track where the progress is with that?

Sure thing. Let me chime in. So just to put the numbers behind that response. So we have 100 trucks globally, right, of which 75 of them are in the U.S. and both 75, they're split between revenue and testing. So a number of trucks that need to be upgraded, it's on the 35 to 38 unit range.

And from an incremental CapEx, incremental OpEx, it's definitely reflected on the updated guidance. These are amounts that -- they're not immaterial, but they were primarily funded through savings in other areas. So we feel very comfortable supporting and completing those upgrades by the end of the year for, again, that testing and revenue fleet.

In '23, we'll continue that to upgrade in the investment on the trucks, again, to ensure that they are running real commercial routes. But we don't expect to start seeing cost per mile improvements likely until the second half of next year, which for me, that's the biggest metric that we'll use to show progress, right? We can show and share updates as we go. But for me, quantitatively, we'll need that true cost per mile improvements to show that the improvements are paying off.

(Operator Instructions) Our next question or comment comes from the line of Brian Ossenbeck from JPMorgan.

Xiaodi, you mentioned you did all the steps as required and even some that may be not required when reporting the accident -- of the incident. Is there -- as a leader of the industry, do you think -- or maybe the industry in general has to be more proactive as there are more miles and more potential incidents that happen rather than relying on the government and posting on that website? Do you think there's room for you to do more and to kind of disclose those more on that upfront? But I guess not waiting for the government to put it on their website and for something like this to happen again. Just want to get your thoughts on that.

Yes. I would say that there's kind of a general dynamics. If there are certain regulation works like certain common practice for every player in the -- in this field to be more open and transparent, definitely, we're all for that. I would say that there is a recent change from the NHTSA website that we were getting everything publicized.

And first of all, I think we were not shy of being transparent, and we actually report to all of that. And you know what, in this call, we also share even more information to that. For example, we told you our total miles, which, by the way, is not included in the NHTSA report. And on the other side, I think we strive to keep our partners and stakeholders informed as possible about our progress. With this, this was an incident with a correctable flaw, not a material change to our business model.

Okay. Understood. And then on the U&P (sic) [UP] process with the Driver Out run, can you just give us an update on that? It sounds like that's going to be something going on later this year? And then just I guess, more broadly speaking, now that the first wave of Ghost Rider is done, can you just comment on Driver Out applications at this point in time and what's planned for the rest of the year?

Sure. Yes. The overall framework of the safety case is there, but we're actually substantiating this with a lot of more stronger requirement that makes the overall behavior for all of the operational aspects even more safer, so that's what keeps us busy.

And in terms of specific -- the Union Pacific Driver Out run, we unfortunately have encountered a complete road closure in front of the distribution center at our destination point. So we are actually actively looking for adequate solutions, but that means that the complete road closure delayed us for a couple of weeks. Not months, but a couple of weeks. Right. And within that, the Union Pacific run, UP run, we actually not only strengthened our system, but also we adopted a new type of trailer with unknown load, and this is actually adding quite a bit of the new elements into our safety case that we have to validate through the unknown ways of an unknown new type of trailer. And we have actually conquered that part of the test. We're still finalizing the tests. You will hear us in the near future about the run.

Okay. And then just maybe to touch on the comments made earlier to the previous questions. It sounds like all in, there's clearly some other challenges that maybe have popped up in the supply chain. It sounds like in terms of the upgrading the trucks, not getting as many as perhaps you might have thought. But is it still safe to assume that you're on track for the wider commercial rollout at the end of at the end of '23? Or maybe some of that buffer has since been impacted by these events? If you could just clarify that, please.

Sure. Yes, of course, there are challenges here and there. But I think when we do the design, we are very cognizant of all of the changes. And the team is doing a terrific job in keeping our time line, still on track. And we're still confident of all of the promise that we have made on the Analyst Day.

Our next question or comment comes from the line of Joseph Spak from RBC Capital Markets.

Just to, I mean, quickly follow-up on that. That time line comment also goes for the production truck time line in 2025? Or is there any change there?

So far, there's no changes on that.

Okay. One, I guess, quick callout, clarification and then, I guess, one bigger picture question. The -- you mentioned the truck reservations, which I think you added 10, but then you also sort of detailed the Hegelmann partnership. And it sounds like there were some initial reservations, but that happened after the quarter closed, so that's not in that number. Is that the right interpretation?

Yes. Hegelmann is in our backlog.

It is. Okay. Because that was -- and you -- because you report at end of quarter and that happened, I thought that came out in July, so I just wanted to understand that. But okay.

And I guess the -- going back to the accident and again, this is understandably a big question. But look, I think even you would acknowledge even when the system is "ready" accidents are going to occur. So how are you going to communicate to your partners and to your stakeholders like what is the definition of safety you're striving for? Like how do we know when it's safe enough for you, safe enough for your partners and safe enough for regulators, all sort of constituents, what should we really be looking for?

I think we will -- like basically, by definition, no one is omnipotent where being able to have like a strong, clear, 100% statement about safety. It's more like we show transparency and our -- we show our methodology, and we have evidence to prove to our customers and our OEMs about our safety principles and context.

And the thing is that for Driver Out, we do have that thing very well prepared. But unfortunately, the incident that's shown up here is more of the operational safety. But nevertheless, I think, as I said in the beginning, I take the responsibility of this incident. I think the company needs to be more -- putting more effort in the operational safety aspects, especially given the fact that we're not going to have all of our truck Driver Outs tomorrow, there will be always the human machine interface out there, and there will be also the operational elements in the system. Therefore, this is the area where definitely we are strengthening right now. We have done a lot of efforts since the incident, and we're going to be more cautious and be more (inaudible) in that aspect.

Can you maybe share -- and I understand different partners might be -- might have different requirements, but can you share anonymously, of course, like what a partner -- like how are they defining safety? Like what's going to give them enough comfort to say we're ready to go?

Because the customers are -- I'm mainly talking about like the hardware-related partners. The partners are -- this is -- Level 4 autonomous driving, it's a new thing for everyone. The partners are open to us to hear about our safety case. And basically, as long as we show them the safety case at a high level -- of course, if it goes to really detailed level, that's a proprietary thing that we will not share. But the high level safety case, once we share with them, they're on board with the principle and the structure of it and with the evidence that can show the overall safety. And that is how we are collaborating with our partners.

But I guess I meant the freight partners. What gets them comfortable -- what gets them over the hump?

I think the period is the same, but the granularity is probably 1 or 2 tiers less because the freight customers are less technical than the hardware partners.

Our next question or comment comes from the line of Rajvindra Gill from Needham & Company.

Apologies if this question was already asked, but just joining from another call. Just following up on that, how do you think that the National Highway Traffic Safety Administration is going to, I guess, adjust their view on what conditional automation is versus, say, Level 4 automation? Are -- do you anticipate that they will provide more stringent requirements between the different levels of autonomy to prevent future accidents like this in the future? I'm just wondering if you've had feedback from the NHTSA as a result of this particular accident and what's been their feedback.

Yes, sure. First of all, I don't think we have any views of how the NHTSA would, in the future, modify their regulations. But what I can tell you is that FMCSA and NHTSA come to TuSimple, they did their thorough analysis of our process, of our -- every step that we do and our -- all our independent investigation results. And there is nothing that's based on anomaly were inappropriate. They gave us -- so far, we implemented the solution on our own, and they do not give us any further recommendations on how to change that or order that. And of course, the investigation is not completed, we remain confident about the investigation.

Right. I think it goes to the heart of the technology and of the business model where the goal essentially is to remove the driver in order to save cost and efficiencies. And there's an impression that Level 4 autonomy is going to happen in freight faster than passenger vehicles. So I'm just curious, if that gets pushed out or if that gets stalled in any way, how would you kind of react as a publicly traded company if that -- if there are new rules or new legislation that would prevent that from happening?

Yes, I would say that, first, we don't really see any signal of that happening or about to happen. And the second is that, of course, we're not going to break the law. And lastly, I think we're going to demonstrate transparency and adherence to the right protocol to demonstrate to the government agencies that -- basically, to give them confidence through our process.

And how does this investigation affect your rollout, your anticipated plans of future Driver Out programs? Again, you might have answered this with previous questions. But my sense is that this year, there's going to be several Driver Out programs. Expanding on the Texas Triangle, for instance, is that being delayed or pushed out because of this crash?

Yes. First of all, I think the promise is about Union Pacific Driver Out run. For Texas, our deadline is set for next year by a lot of other improvement, removing the support vehicles and holding commercial runs, day and night. Yes. So those are for next year. I don't think that we have our deadlines in this year and all of the Driver Out activities. And preparation [of our Driver In tests] are not affected by this incident.

And as mentioned, I think the fact that when this event occurred, we stopped our fleet and we root caused what really occurred and [really -- applied and fixed. And now we're testing our current bots using this basically. We] feel very confident that this is an isolated event, and this has not had any impact to our plans. And again, we're same behavior or reaction that we've had on previous events, that we are taking it seriously, and we'll do that as expected going forward, and we feel confident that this issue that created the accident has been solved.

I'm showing no additional questions in the queue at this time. I'd like to turn the conference back over to management for any closing remarks.

Thank you for joining the call, and we look forward to attending many conferences this quarter and giving you another update next quarter. Thank you.

[Ladies and gentlemen, thank you for attending] today's conference. This concludes the program. You may now disconnect. Everyone, have a wonderful day.