TuSimple Holdings Inc (TSP) 2022 Q1 法說會逐字稿

Good day, and welcome to the TuSimple First 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, Ashley. Good afternoon, and welcome to our first quarter 2022 earnings call. With us today are TuSimple's Co-Founder and Chief Executive Officer, Xiaodi Hou; and Chief Financial Officer, Pat Dillon. Xiaodi and Pat will review the operating and financial highlights, and then we will take questions.

As a reminder, TuSimple's shareholders' 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 events or results due to a variety of factors. Please refer to our earnings release and in the sections of our SEC filings titled Risk Factors.

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. Xiaodi?

Thank you, Ryan. Hello, and welcome to our first quarter earnings call. Today, we'll update you on exciting partnership agreements that we believe will drive value in our AFN and share some of our recent technology and commercialization achievements with you.

And we're very excited to host many of you at our Investor Day next week in Tucson. It will be a great opportunity for the investment community to experience our industry-leading AV technology firsthand here, from many of the talented individuals within our organization, and for us to share more details on our long-term plans. We are very much looking forward to hosting you next week.

Since TuSimple founding in 2015, we have had one goal of building the first to market scalable Level 4 autonomous semi-truck solution that will transform the freight industry. Our mission is to change the game. We believe our technology will improve safety, increase efficiency and reduce our carbon footprint.

While there is much more work to be done, our autonomous driving system is now feature complete, and I'm excited to be leading TuSimple towards commercialization. Our commercialization team members are best-in-class, and we have established some of the strongest partnerships in the industry. I'm extremely proud to play an instrumental role in bringing all aspects of the company together to achieve our goal.

Last week, we announced some Board changes. First, we're very pleased to add Reed Werner to the Board. Reed has deep experiences in finance as well as in government and defense, and we're grateful to have his counsel in our Board of Directors. We also announced that Brad Buss has been appointed lead independent director. Brad has been critical in all aspects of our business, particularly as we have transitioned to life as a public company. I greatly appreciate his many contributions so far, value his counsel and looking forward to working with him in a new capacity as he takes on this additional responsibility.

We also announced that my Co-Founder, Mo Chen, will not be standing for reelection. Mo and I have a long-standing and deep partnership that has helped build this entire company. Mo has always focused on doing what will be best for TuSimple. As we have talked about previously, we are targeting more independent directors and a more U.S.-centric Board. We are focused on enhancing governance, and our home market in the U.S. is expected to be the first to commercialize.

So this is a natural and strategic evolution of our Board composition. Mo is fully aligned with this strategy, and he made the decision not to seek reelection to support the company. We thank Mo for the many contributions he has made as a Board member, helping to guide the company from its infancy to what it is today. Mo will continue to be a significant shareholder of the company and will continue to give us his full support while his director term finishes in June. We will continuously evaluate additional independent director candidates to augment our Board in the future and to continue to bolster our governance.

During this quarter, we made significant progress towards commercialization. We now have over 10 Driver Out missions under our belt, and we made meaningful enhancements to our minimal risk conditions, or MRC capabilities, including our Autonomous Driving System ability to pull the vehicle over to the side of the road in certain scenarios. This is a critical capacity in order to be able to remove our chase van support vehicle and improve efficiency in our Driver Out operations.

To test this capacity, we first ran trial in simulation and then on closed tracks. Finally, we proceeded with the Driver Out test drives, where we deliberately triggered MRC command while the vehicle was in operation on Interstate 10. As designated, our semi-truck identified a safe place to pull over, lowered its vehicle speed in lane and then put it off to the side of the roadway.

The MRC capability to pull over to the side of the road may not seem like progress. After all, our AFN is about completing mission successfully. But a robust MRC capability is needed to safely take the Driver Out and to optimize operations. As the world's first and only Driver Out trucking company, we will continue to refine our technology towards commercialization and provide updates on critical capacity advancements such as MRC.

This quarter, we also launched collaborative mapping. This enhanced the fleet mapping capability, allows our semi-truck to share information with each other using sophisticated vehicle-to-cloud-to-vehicle communication. Trucks traveling the same route share dynamic events such as identifying emergency lane vehicles well before the camera or LiDAR can sense the event. This advanced notice is incorporated into the vehicle planning system, giving 2 simple semi-trucks considerable time to make a safe lane change or other driving maneuvers. As our operations scales, the collaborative mapping is expected to help us continue down the efficiency path, and more importantly, magnify and enhance the safety benefits of our AV technology.

TuSimple has one of the broadest patent portfolio in the industry, and the value of our proprietary technology continues to grow. We were issued 21 new patents in this quarter, bringing our total patent count to 408, highlighting our industry leadership in intellectual property globally. Patents issued in the quarter will improve safety and reduce operational costs.

Now, I'd like to talk about Navistar joint development partnership to build the world's first Level 4 production semi-truck. Over the course of the last 2 years, we and Navistar have made significant progress on advancing our technology and refining our production plan. This includes a significant amount of progress over the last 6 months on our bill of materials with Navistar, including deep discussions and negotiations with the supply base. This work is critical to advance both our production trucks but also to utilize our hardware knowledge for advancement of the retrofitted trucks over the near term as we expand our Driver Out operations.

We are also pleased to announce that the production of semi-truck will be manufactured in Navistar's world-class Escobedo, Mexico manufacturing facility. This is an important milestone as we jointly prepare production and is a signal of the conviction that we and our partners have in this program.

As we and others in the industry have highlighted, the last 2 years has seen unprecedented global supply-based challenges. This has delayed us from the initial production truck time line we set in 2020. We and Navistar have refined our time line to have production in 10 prototypes in 2024, with fully-integrated vehicles targeted for 2025. Even with this delay, we believe the timeline that we have developed will lead our purpose-built commercial truck to be the first to market.

In parallel to the production truck program, we will also continue to commercialize our technology with successive generations of more advanced retrofitted trucks, driving efficiency and ramping the scale of our operations into the hundreds of trucks. As our production truck starts coming off the line, we will be able to place thousands and tens of thousands of trucks onto our AFN, fully scaling adoption of our groundbreaking technology.

Year-to-date, we've made some important announcements that will enable our Autonomous Freight Network to bring efficient AV capacity to the trucking market. First, we expanded our AFN ecosystem with Ryder. We added 2 internal -- 2 terminals in Houston and Laredo, Texas. This terminal will provide non-stop access points for our customers, providing bundled maintenance service to support commercialization. By partnering with Ryder, we're able to quickly, and in a capital-efficient manner, add to our AFN terminal infrastructure in Texas. This is an important step to prepare for the expansion of our Driver Out operations over the next 12 months.

Second, we announced that an integration with Werner Enterprise (sic) [Werner Enterprises] roadside assistance to support our Driver Out operations on our AFN. As anyone in the truck industry understands, breakdowns happen, whether Driver Out or not. What's critical, however, is expedited service when assistance is required. Our announcement with Werner gives our AFN access to an extensive network of dealers and repair facilities, offering 24/7 support. This means carriers on our AFN will have their choice of service and support providers, including Werner Enterprise (sic) Werner Enterprises authorized service centers.

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

Thank you, Xiaodi. Beginning with our reservation program. We added 500 new truck reservations this quarter, including one larger reservation of 350 trucks from Loadsmith, a leading third-party logistics and capacity as a service platform. We're excited to be working with an innovative company like Loadsmith and see this as a validation of the attractiveness of our AV technology and the value that we are bringing to the industry. This brings a reservation total at the quarter end to just under 7,500 trucks.

The interest in our technology from our fleet partners has never been higher. As we build on the success of our initial Driver Out missions towards operationalizing Driver Out for customer freight runs, we continue to hold high 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 first quarter. We reported $2 million of revenue in the quarter, an increase of 140% year-over-year and a slight increase quarter-over-quarter. We have kept our owned and leased revenue fleet relatively flat quarter-over-quarter as new trucks are prioritized for our testing operations. The flat truck count and relatively flat revenue mile growth was offset by higher rate per mile driving our revenue growth.

While 2022 does present some headwinds such as adding drivers, trucks and trailers, we expect to increase our revenue fleet size modestly over the course of the year and to continue to improve utilization. We closely watch the pricing environment, and we price competitively with our freight customers.

Moving to expenses. We spent $78 million on total R&D in the quarter, including $17 million of SBC. This compares to $41 million in the same period last year and $82 million in the fourth quarter of '21. The R&D expense declined by 5% quarter-over-quarter, primarily due to a one-time true-up of 2021 bonus accruals. We continue to fastidiously manage our spending to invest smartly in the commercialization of our world-leading technology. We spent $32 million on SG&A during the period, including $10 million of SBC. This compares to $15 million in the same period last year and essentially flat versus fourth quarter '21, also benefiting from a one-time true-up of 2021 bonus accruals.

Our loss from operations was $112 million in the first quarter of 2022 compared to a loss from operations of $59 million in the same period last year and $116 million last quarter. Our adjusted EBITDA in the first quarter was negative $80 million, which compares to negative $50 million in the same period last year and negative $81 million last quarter. We invested $1 million in the purchase of property and equipment during the quarter, primarily related to equipment purchases and facility investment. We also made acquisitions of property and equipment of $1 million that is included in liabilities.

We ended the quarter with a cash balance of approximately $1.25 billion, a decline of $100 million versus the end of 2021. The cash burn in the quarter included the payout of annual performance bonuses and other similar items, which are not expected to repeat in the next 3 quarters.

A quick update on our 2022 guidance. There is no change to our full year 2022 guidance. However, given our announcement regarding the exploration of potential transactions for our Asia Pacific-focused businesses, we do want to provide some context. We expect approximately 20% of our 2022 adjusted EBITDA loss to be related to our Asia Pacific-focused businesses. We are not expecting any material revenue contribution from these businesses during 2022.

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

Thank you, Pat. The technology we've developed is best-in-class. Our path to commercialization is becoming clearer, and I am honored to be leading this amazing company through our next phase of development. We're excited for 2022, which we expect to be a year in which we expand our Driver Out operations, continue to make progress on efficiency and work towards full production of our purpose-built truck. We will have more details to share with you next week in Tucson. I look forward to meeting many of you and highlighting the investments we have made.

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

(Operator Instructions) Your first question comes from the line of Ravi Shanker with Morgan Stanley.

Xiaodi, thanks for the update on the Navistar relationship and the new time line. Just so that we understand, can you give us the kind of step-by-step on what you expect in 2022, what you expect now by the end of 2023? I think you had said some level of commercial implementation by end of '23 before. What happens in '24 and what happens in '25?

Yes. Thank you, Ravi. The Navistar thing is that we have agreed on a lot of things on the building material -- bill of materials. But the things that we cannot control in the world is that the supply shortage on that, and we are finalizing it. But even with this delay, we still believe that we are the only company that go into item-by-item about the items need to be -- needed for building the autonomous truck. And this is our plan for now, that we're going to have the prototype ready by 2024, and the real truck coming out by 2025.

And maybe -- Ravi, it's Pat, just to give a little bit on your question about what is happening in 2023. The important thing to remember is that these programs are running in parallel. So what we're doing with retrofitted trucks, which we expect to continue to improve in terms of the quality of the hardware, the reliability and the fidelity of the hardware, will continue to improve. The commercialization is not tied to the production trucks. So the commercialization, which we'll talk more about in next week's Investor Day, will really be using retrofitted trucks initially starting at the end of 2023 and into 2024.

In parallel to that, we will be making regular and steady progress into the production truck, and those are the specific milestones that Xiaodi laid out in his prepared remarks around production in 10 prototypes in 2024, and we're targeting the fully integrated vehicles on the road in 2025.

Got it. That's an important clarification. And maybe as one follow-up, just on the first quarter itself. It looks like your miles driven, both total and MAP, were flat sequentially, which I think is the first time that's happened. So can you just elaborate on why that was the case? Kind of did you guys run into any COVID operational issues or something?

Sure. Thank you, Ravi. It's a good question. So I think for mapping, we have different phases of development. The first phase is to demonstrate to the world that we can map cheaply and quickly, and that goal has been achieved. And given the finite number of trucks that were operating and the hauling freight on the entire U.S. network, we have to come to the situation where we do not need to map the route where we don't operate.

And in the new phase, what we're focusing on, as you also -- as I also mentioned previously about the collaborative mapping, we focus more on the latency of the mapping, the accuracy of the mapping and the additional functionality of the mapping. But we do not need to expand the miles of mapping because that's only one of many success criteria for mapping.

Your next question comes from Chris Wetherbee of Citi.

This is [Eli] on for Chris. If we could just start with the Union Pacific partnership, you mentioned that was going to be rolling out in the spring here. Just curious if that's happening now and if there are interesting learnings that are coming out as a result from some of the final mile moves there.

So for the Union Pacific collaboration, initially, we were aiming at hauling a trailer, like a typical trailer for Union Pacific in a Driver Out fashion. However, as we go deeper into the collaboration with Union Pacific, we realized that their model is rail -- really, the rail intermodal container transportation. Therefore, it will be a much more significant achievement if we can haul the rail intermodal container instead of a trailer, which is more like a smaller aspect of their business. Therefore, we have to modify our ODD and expand our operations to be able to support the rail intermodal container.

And with that, we set the new time line for the third quarter of this year to commence the Driver Out ride for the Union Pacific for the rail intermodal container, which is an expansion of original plan. And with the new plan will be -- sorry, will be much tighter integrated with the daily operation model of Union Pacific.

Okay. Got it. And then in terms of Texas, so presumably, you want to continue to do Driver Out routes there in 2023 and those will be prototype or upfits, I guess, that will be doing those routes. So you imagine that you will add some more trucks out there for that. And then I guess more on 2022, so there's going to be significant reductions in cost per mile. Are those going to be on the prototype trucks? And what's happening to help reduce the cost per mile there?

There are several aspects. Number one is that we are going to operate in Texas by migrating our operational design domain from night to day and migrating from one lane in -- one Driver Out lane in Texas -- sorry, one Driver Out lane in Arizona to more Driver Out lanes in Texas. You see, those are already 2 operational design domain expansion. And in addition to that, we are going to change some of the hardware to a more product-matured version of the hardware. By doing that, we're going to reduce the operational cost.

And furthermore, we have multiple roles in our initial Driver Out. We have leading vehicle. We have chase van. And we do have a plan to remove all of this in the following, I would say, 2 years.

It's Pat. So maybe I'll just add a couple of things. One, we do expect to give a lot more detail on the Driver Out expansion at our Investor Day next week. And two, just to make sure that the nomenclature doesn't get confused. The trucks that we'll be operating this year and next year will be upfits or retrofits of pretty close to stock at Navistar vehicles. The production in temp prototypes that we mentioned for the purpose-built truck, we've targeted those with Navistar to start rolling off the line in 2024. We haven't specified yet, but they will be used likely for dual purpose, including revenue runs as well as doing all of the typical testing that you would expect in a pre-production phase. So hopefully, that helps clear up sort of the cadence of what will happen this year, next year and then into '24 and '25.

Your next question comes from Dan Levy of Credit Suisse.

I want to go back to the questions asked earlier on the metrics. And I think we're all figuring out how to read into the metrics on a quarterly basis. But look at your cumulative road miles, and the quarter-over-quarter growth there decelerated. It was 14% in the first quarter, down from 17% in the fourth quarter.

I recognize that there's going to be some lumpiness and we shouldn't read too much in that, but maybe you can give us a sense of how we should look into the quarterly progression of these road miles, especially as you're focusing more on commercialization and I think seemingly narrowing the number of routes where you're doing serious work as opposed to a wider set of routes.

I think the actual miles are only part of the story. For example, we do have a lot of simulated miles as well, and all kinds of simulation does not really reflect in the miles. For example, we have hardware in the loop simulation. All this kind of validations are progressing. And I think miles only tells you part of the story, so my recommendation is don't really look into -- too much into that.

And I will let Pat answer the rest of it.

Yes. I mean, Dan, just one point on the math around it, too, is that we're doing this with a relatively flat truck count. The new trucks that we are adding into this are largely going into testing operations. Testing operations are typically running a very specific test to test very specific functionalities. So the overall -- the progression on the additions to the cumulative road miles have been fairly steady over the last couple of quarters.

And to the point Xiaodi made, it's not about trying to grow the cumulative road miles just for the sake of that metric. It's really an output from a confluence of a whole number of different objectives that we're trying to accomplish in developing this technology.

Got it. And then just the second question. I want to follow up on the press release you put out about 6 weeks ago on the exploratory process for your Asia Pacific business. Maybe you can just give us an update on where that stands. And just broadly, maybe you can refresh us on how you're allocating resources between U.S. and China and whether you think it's still possible to concurrently develop autonomous trucking in both regions or whether it makes more sense to purely focus on the U.S. And I know you have operations in Europe as well. But just how you think about the regional focus these days.

The regional focus, definitely we are a U.S. first, U.S.-centric company. That is the first thing because I think in the U.S., we have the best policy to commercialize autonomy, and this is whole company's purpose.

And in terms of the possible fundraising, I will let Pat answer that.

Yes. I think just as a refresher, Dan, and from the -- when we put the press release out. First of all, I'd emphasize that it's a preliminary process. It's something at an early stage, so we don't have a lot of detail to share yet. But the rationale behind exploring a potential subsidiary level transaction for our Asia Pacific-focused businesses is really just seeing a financial arbitrage opportunity that could potentially be beneficial to us. In discussions with market participants, we think it is likely that there is relatively minimal value ascribed to our Asia Pacific-focused businesses and our stock price today.

We do see a great deal of intrinsic value in our Asia Pacific-focused businesses. And we've also seen that there are other pools of capital or market participants that would ascribe value to our Asia Pacific-focused businesses. So we're not making any commitment to taking any particular action or any particular type of transaction, but we do think it's worthwhile for us to explore that. And we think it is potentially a path for us to unlock financial value for the company.

I'd say just one other thing that is important to note is, from a CFIUS perspective, we did put out our press release with all of the material terms of our National Security Agreement with CFIUS. I want to be clear that there is no action required around our Asia Pacific-focused businesses from a CFIUS perspective. So this is purely looking to unlock financial value that we don't think is reflected in our share price today. And again, just -- it's a preliminary process. As we have more as we move through this, we'll certainly make appropriate updates to the market as we go through that process.

And just to be clear, none of this is being spurred by additional capital needs in the U.S. This is purely because you see a financial opportunity, not because you're in a need for additional capital.

Well, I mean, I do think, Dan, it's probably premature to know exactly what a transaction would bring to us and exactly what it would look like. Certainly, there could be opportunities where we could fund some of the investment needs in the Asia Pacific-focused businesses with a separate pool of capital, and that would, of course, have an impact on our -- the cash burn on our resources at the parent company level. But I think it's premature to speculate around that. It is one potential benefit that could come from this, and certainly, a lever that could provide additional runway on our balance sheet.

Your next question comes from Brian Ossenbeck with JPMorgan.

So I just wanted to get maybe a little bit better understanding of the partnership with Loadsmith, what they plan to do with the 350 trucks you have there. And then just maybe a bigger picture in terms of your own PMS when you have your own capacity, how you're going to connect into others. I think it would be helpful if you can just kind of outline where you are on that process and whether or not this partnership with Loadsmith is part of that or something separate.

I'll let Pat answer that question.

Brian, so I would say from Loadsmith, obviously, a company that's been pretty innovative for those that know it in the space, I think with a lot of different partners here, they are all seeing autonomous trucking as a new form of capacity that's coming into the marketplace and a way not to replace the existing capacity that's out there, but, really, a complementary technology. And I think the partnership with Loadsmith is certainly one that is adhering to that general philosophy.

We do have, as you noted, an oversight system. The purpose of that oversight system is multiple. It is to track where our vehicles are and to provide support to those vehicles while they're in operation, but also to communicate with the various shippers, carriers and others that are using the Autonomous Freight Network. And we know that that's a complex process. So being able to tie our oversight system into the systems of someone like Loadsmith certainly makes it easier for us to plug the capacity that Loadsmith has reserved into their operations as well. And we expect to share more details about this over the coming quarters as we get further along in the partnership as well.

Okay. That makes sense. Just maybe touching on the supply chain and the maturity of it. Are there any areas you feel like you still need to invest in, especially now you've gone through more of the redundancy testing with the Driver Out program? You can bring us up to speed on where you kind of are right now and what maybe some of the challenges that you're still facing above and beyond just the obvious one with the OEM trucks.

Yes, sure. I think in one word, the devil is in the details. So you have to crack the shell and just lay down thousands of items in the bill of materials, try to identify one by one and contact the suppliers to understand what is ready, what is not ready.

And to give you an example, in the past, we found the biggest gap here was the domain controller, the computing unit in the system. And none of the domain controller in the system would actually be sufficient for our purpose, so we have to build it ourselves. So that's why earlier this year, we announced our partnership with NVIDIA and we're building this domain controller by TuSimple. Of course, not manufactured by TuSimple, but designed by TuSimple. And this is our attitude in terms of bridging every single gap that we see and try to deliver as soon as possible.

And apart from that, there are some other things like redundant steering, redundant braking. And as we all know, the LiDAR industry, whole thing is new. So we're all racing towards -- basic -- everyone is having their full speed towards an earlier production time line, but things could happen in this.

Your next question comes from David Vernon with Bernstein.

So a question for you on resourcing. Its 4Q to 1Q head count, R&D head count, relatively flat. Is that going to change as we move through the year, looking at expanding into the operations into Texas and things like that? And is there -- is it right to read into that, that more resources are not what you need to kind of accelerate the path towards commercialization?

I think one thing is that you can have 10 women to born a child in 1 month. So the acceleration of talent is certainly going -- not going to be exponentially growing in the future, that's what I see. And also at the same time, we are very capital efficient so that we do not want to waste time on anything.

But we do want to tell you that there is going to be a grow -- a steady growth of our talent and expenditure in terms of research and development because we always feel that accelerating the development is our #1 goal. And we'll take it what it takes.

Okay. So the plan will be then to grow those numbers as we go through the end of '23?

Yes. David, it's Pat. So just a little bit on the figures. Yes, we do expect to see modest head count growth. We obviously saw a pretty significant head count growth as you looked at where we were in 2020, and the heads that we added in 2021. I would say it's -- and you've seen this in the first quarter, some significant deceleration in both the head count growth, and that translates also into the financial metrics around our spending and investment.

So I would say it won't be flat, but I think you'll see some deceleration. And we're going to be very smart from an ROI perspective on how we invest in our people, our non-compensation spending. And essentially, we'll be balancing and really closely monitoring our head count growth, our spending. But obviously, we won't sacrifice our ability to hit the milestones, which we think are pretty aggressive for the next 18 to 24 months. We're going to make sure to invest at a level that will allow us to hit those milestones.

All right. And then maybe just as a quick follow-up. Should you -- are you guys planning to do anything around stock-based comp, just given kind of where the stock is? Are you worried a little bit about retention? Are you going to be thinking about maybe needing to have to chew up some of the key employee stuff? I'm just trying to get as for whether there's some pressure on head count right now being created because of the performance of the stock.

Yes. I think the stock-based incentive is pretty common in all companies like us. A similar principle also applies to us. We wanted to keep the top talent. And of course, with a relatively lower stock price, giving a meaningful amount of stock is the best way to keep them.

Yes. I think we'll keep -- we're not making any changes to the stock-based compensation charge. I think that will be inclusive of regular programs. But we're also making sure that we're investing strategically around particular retention needs and a competitive market dynamic. So no change to the specific stock-based compensation guidance figures. But it's certainly something that we and our Board of Directors are focused on, on retaining our top talent.

Regardless of the stock price environment, this has been always something that we'd had to deal with, whether it's competing with much larger companies. And certainly, the stock-based compensation and the cash compensation is one element of it, but the mission of the company and the ability to work on groundbreaking technologies has always been a really attractive recruiting and retention tool for us, and it continues to be in 2022.

Your next question comes from Alex Potter with Piper Sandler.

Great. So I was wondering, I don't know the extent to which it's possible to quantify this because I know you have a lot of costs associated with chase vans and things like this that make it so that maybe these cost per mile figures that you're achieving right now maybe aren't perfectly apples-to-apples versus what they will be several years from now when you're in sort of pure autonomous mode. But if you correct for all of that, are there any numbers that you can give us in terms of cost per mile in your real-world runs that give you confidence that you're on the right sort of downward trajectory for that metric?

Yes. We will give you a very detailed analysis next week in Tucson for our Analyst Day. And I think maybe in here, I can give you some very top level overall feeling of it.

I think there is a misconception that a lot of people are feeling that autonomy, by removing the driver, you get all of the profits from the driver cost. That is untrue. And the maintenance of the autonomy system is significant, and a lot people have overlooked that part of the cost. But that part of the cost is really the inevitable way for us to go until we reach commercialization, and we're all-in for that goal.

Okay. Great. We'll look forward to getting some of those numbers next week. I guess the second question that I wanted to ask about was related to the NVIDIA partnership that you were just mentioning. Just any -- I know it was about, I don't know, several months ago when you disclosed the relationship to begin with. Any update you can give there regarding milestones or how the relationship has been progressing and the pace of the development so far? Is everything more or less tracking in line with what you had originally envisioned?

Yes. I just wanted to remind you that we're still in the very early stage of a multiyear development. And so far, so good. Everything is on track. And we're very happy that we have this collaboration in our pocket because we have actually developed a lot of things that has to be done by TuSimple, and you can't find anything -- or has to be fulfilled, but you can't find any substitute in the market so that the decision of going this path is -- looks more and more to be the wise decision.

Your next question comes from George Gianarikas with Baird.

First, I'd like to ask about collaborative mapping and any additional detail you could share on how that's a differentiated offering relative to some of your competitors.

Yes, of course. I think we're -- as we are entering from the safety phase, the first phase is safety, and we're entering from the safety phase to the efficiency phase, we realized that the autonomy technology should not be developed as a single vehicle, but should be developed under the grand formula of operating the entire fleet. And this aspect of collaborative mapping is really like sharing one vehicle's perception system with the other one and better facilitating the capability of the motion planning by, I think, additional information. And we believe this is a very advanced technology.

Got it. And just to clarify an earlier point and questions around the Navistar relationship and supply chain. And it sounds like you're saying that the industry's time line to commercial deployment has shifted a little bit to the right based on issues with redundant steering or braking or some -- or a multitude of technical issues that really haven't been solved yet. Is that a fair assessment?

Yes. And I think Navistar and ourselves, Navistar is a great partner to ourselves. And we, on both sides, are having very deep collaboration on this. Really, we are bridging many, many gaps that were previously existed, looking like unsurmountable gaps, but we actually reached most of them. But there are still certain things in the world that we cannot control. We cannot build the LiDAR by ourselves in-house or the redundant steering by ourselves. So that's why a lot of these type things are -- yes.

But the other thing is that I do not really see any technical breakthrough in order -- in this kind of delayed waiting time because all of these are pretty mature technology. They just need to go through different engineering development gates to be able to launch to the automotive grid to be integrated into the system. And it's really about time, not about anything that we're waiting for another physical level innovation.

Your next question comes from Adam Roszkowski with Bank of America.

Adam Roszkowski on for Ken Hoexter. Following up on the Werner partnership, is there a time line here? Any major milestones you are targeting? And have you been able to quantify any dollar impacts to safety, fuel efficiency that can be gained from successful implementation?

Yes. We actually have quite a bit of analysis, quite an in-depth analysis into that. We will see you next week in Tucson to give you a full disclosure of that.

All right. Can't wait. And then on the reservation pipeline, so $500 million added this quarter. Can you just talk about what the pipeline looks from here? And does a partnership with a player like Loadsmith maybe broaden your view on your go-to-market strategy and different players that you can target essentially?

Yes. I mean -- it's Pat. I would say that we've seen really robust appetite from folks throughout the freight ecosystem to understand how they can incorporate AV trucking in their models. Certainly, we've seen folks that you may not initially associate with having assets like a semi-truck in their operations as a principal focus. We make reservations and be part of it. So I think the applications are quite broad and people are really thinking outside of the box about what the freight ecosystem will look like as this technology proliferates. So I think it's probably one indication of the breadth of the types of discussions that we're having, and the momentum that we have with the commercial base has never been stronger. So we're really gratified for our existing partners and excited to bring on new partners.

The last piece I would just say is that we're being very disciplined, and that with our reservation partners, it's a 2-way commitment to work together. And we want to work with partners that are committed to not only making a financial commitment and having the resources to deploy this technology, but making sure that our near-term goals around testing and freight operations are aligned such that we can fulfill our obligations to the partner, too. So we're trying to be very, very disciplined about reservation partners as opposed to doing anything to just grow the reservation or anything to just grow the reservation number itself.

Your next question comes from Colin Rusch with Oppenheimer.

As you get into this next layer of mapping data and integration of that data, could you talk about the cadence of learning cycles and what you're pulling out of the system, and kind of where there might be some blind spots or some unforeseen strengths in the system so far? So the question's really about -- to the pacing and learning cycles; and secondarily, about where you need to improve or really feel pretty good about the system so far.

Can you repeat the first question? I think there's some noise. I didn't hear clearly.

Okay. Yes. I mean what I'm asking about is the pace of the incremental mapping data and that integration into the system and the learning cycle related to that integration, what you're seeing in the system, how fast is that getting integrated. And then secondarily, where you're seeing some strengths and weaknesses in the system.

Yes. So there are several things. This is pretty technical. On one hand, for example, one vehicle saw police basically having a partial lane closure of the road, then the method will be transmitted to our cloud-based data center and in a number of minutes, all of the truck will be aware of that. So this process is like a first automatic process, and then human verification and then broadcasting. So in the number of minutes for this aggregated map, aggregated real-time sharing map.

And the second is that the lesson that we learned from this is that we have all different types of maps. For example, every truck actually is -- when they're driving on the road, not only serves as the purpose of using the map, but they are also validating the map. And that's early -- that's the first step of us. But after that, we realized that if you have the map being updated, but the truck has to go back to the bay in order to update the whole thing, that takes several hours. And we're -- as I said earlier, the goal for mapping is to reduce latency as one of the most urgent or -- not urgent, the most important matter, and also maintaining high accuracy. So this is basically an effort of reducing our mapping response from hours to minutes.

That's [powerful]. Now I'll take the rest offline, just in lieu of time.

Your next question comes from Todd Fowler with KeyBanc Capital Markets.

Great. There's been a couple of comments made about improving utilization on the fleet. And I'm just kind of curious what are some of the levers you can pull to improve utilization in your model right now and how quickly you can kind of see utilization ramp up over the next couple of quarters.

Yes. I think we continue to focus on it. I would just say, Todd, that we have some different constraints than a traditional truckload carrier. We're hauling freight for strategic partners. We do that in autonomous mode wherever possible in order for our customers to see the technology progress and to be part of the development. So we are pulling some of the utilization up wherever we can. The levers are trying to -- the typical playbook of trying to be as efficient with the assets as possible.

And certainly, we've had some headwinds with COVID and some of the drivers being out during the early part of the first quarter that weighed on some of the utilization. But we were able to -- I think we're pretty happy with what we were able to do from a utilization perspective and see some path for modest utilization growth for the rest of the year.

Got it. Okay. That helps. And just as a follow-up. Pat, I think you made the comment that there was no -- there's no change to the guidance metrics. I'm assuming that, that is the case with CapEx. However, CapEx was pretty light here in 1Q, so I just wanted to confirm no change to CapEx. And then if you could just remind us what you're going to be spending from a CapEx -- what you're going to be spending the CapEx dollars on this year, that would be great.

Yes, no change to the CapEx. Obviously, the spend in the first quarter was relatively light compared to our guidance. But we do expect to see CapEx ramp up over the coming quarters, particularly as we start to make some investments in Texas around the Driver Out operations. So some of the terminal locations where we have, we'll be talking more about this. We talked about the ride locations, but there will be incremental locations that where we'll be investing some improvements on parcels of land to support the Driver Out Texas operations. That's one that I think is fairly significant. And then there will be the usual type of corporate CapEx as we expand and invest in things like office space and other facilities as well.

Your next question comes from [Scott Trout] with Wolfe Research.

Apologize if I missed this, but can you just go through the rationale again on why miles mapped are flat and why they're going to stay flat from here?

Yes, because we have a finite number of trucks hauling freight, and that freight is our current supported AFN. And we do not have a plan to like 10x or 100x that of our AFN in this year or next year, right? And by that, we have to be conscious that the goal is not to prove how quickly can we map the future routes. It's more about, okay, we map this route and we utilize on this route, we [cooperate] on this route. So the total amount of mileage on our mapping, same flat is a natural result as we do not need to prove anything further.

But on the other hand, the technology side, the lower latency and more functionalities are the things that we need to focus. So that's why you can see the miles being flattened, but I think miles is not the only metric for this game.

And [Scott], it's Pat. I'll just add on to that. In 2020 and 2021, being able to show that we could expand our network and be able to run more autonomous miles across the Sun Belt was certainly a strategic aim for us, and I think we showed that we could do that efficiently. For 2022 and 2023, as we try and expand Driver Out into Texas, it's much more about the density of the network and the penetration of full Driver Out autonomy. So from that perspective, adding new routes throughout the Southeast for autonomous operation doesn't really fit with what we're focused on over the next year or two. What we're really focused on is being able to go from doing just a single route of Driver Out in Arizona to be able to add new routes and be able to expand the ODD in the ways that Xiaodi was talking about earlier.

And so that's really where our focus is. At some point, we will start expanding the AFN again and starting to expand the reach of our Driver Out operations. But for the next year or two, it's not really a major focus for us. We're really focused on the density and the penetration of the Driver Out operations.

Okay. And so you don't think that this has an impact on the middle part of the decade revenue ramp?

No. And if we need, we can always easily map new routes without a problem.

And then just secondly, the head count was also flat. How should we think about head count and SG&A from here? Do they sort of level out? Or do we see further increases in head count and SG&A the rest of the year?

Definitely, I think you will see a steady increase, but not exponential growth of the technology team. Because on one hand, we have to be very cautious in hiring new people and be capital efficient given the general economic situation right now. And we -- but on the other hand, we do -- we're not slowing down in hiring.

Where would you like to see that head count get to by the end of the year?

So we're not providing any specific guidance around that, but I think you will see some modest growth throughout the year, [Scott]. So it won't be essentially a doubling, which is what you saw in 2021 relative to 2020. It will be a much slower pace of growth. But I think you will see some modest expansion of that.

And the head count being flat is, you'll notice we had a restructuring charge during the quarter, too. Some of this is by design as we -- as you would expect with a young company that's working on a technology like ours, we have done some workforce restructuring to better optimize the way that we go about developing the technology. So that also is partially the impact of where you saw the flat head count quarter-over-quarter this quarter.

This concludes the Q&A session. I will now turn it over to the speakers for closing remarks.

Thank you, everyone, for joining our call, and we look forward to continuing our dialogue. At this point, we'll adjourn the call.

Thank you, everyone. Goodbye.

This concludes today's conference call. You may now disconnect.