Latest / Freightvine / Olivia Hu, Uber Freight | Inside Uber Freight's Autonomous Vision

Transcript

Olivia Hu, Uber Freight | Inside Uber Freight's Autonomous Vision · Freightvine

• Chris Caplice: Hello and welcome to the Freightvine Podcast, your source for all things freight transportation. I'm Chris Caplice Chief Scientist at DAT Freight and Analytics, and in this episode, we're continuing our focus on autonomous freight transportation with my guest, Olivia Hu. She's the head of autonomous trucking and electrification at Uber Freight. Uber and Uber Freight have always played a leading role in the adoption of autonomous vehicles. Rather than develop their own technology, however, Uber Freight has partnered With both Aurora and Wabi to provide a variety of potential options for their shipper customers. Olivia had explains how each AV technology vendor has different service capabilities, or what they call operational design domains. These help determine where in a shipper's or carrier's network that technology best fits. We also discuss shipper skepticism of autonomous trucks, the logistical differences between AVs and EVs, electric vehicles, and the technical challenges of scaling these operations. Beyond just specific corridors in the southwest of the United States. Ultimately, Olivia predicts that there will be significant autonomous adoption within the next decade, driven by increased road safety and improved supply chain economics. Now following my conversation with Olivia, I'll provide a truckload market update. So let's get started. Hi Olivia, welcome to the Freightvine Podcast. Olivia Hu: Thank you for having me, Chris. Great to be here. Chris Caplice: Yeah, it's I've had b other people from Uber Freight on before, but I I don't think I've had you on before. And I think we've probably met at some delivery things. And I know you've been up here at MIT and have been on different panels up here, so I'm sure our our paths have crossed. Olivia Hu: Yes, definitely. It's a small industry, small ecosystem of players. So I Chris Caplice: Isn't it it? Olivia Hu: you know your name very well here. Chris Caplice: so before we get into the details of what Uber Freight is doing with autonomous vehicles, because that's the real focus of the of the why I wanted to have you on, tell us about your career path. How did someone from Stanford with a MPP, a master's in public policy, end up in trucking? So tell t Olivia Hu: Yeah. Chris Caplice: talk us through that path. Olivia Hu: Yes. So funnily enough, you know as a naive young college student, I believe that the public sector was where you can have the most impact on society for creating a better Chris Caplice: Okay. Olivia Hu: world. and that's why I studied public policy. and I still believe that that is true, but it's not Chris Caplice: Mm-hmm. Olivia Hu: as easy as I thought it was at one point. And so I entered the private sector as an internship first shortly after when I was in grad school, and then right after I actually worked at one of the largest asset managers at BlackRock on Chris Caplice: Mm-hmm. Olivia Hu: their risk analy analytics team looking into mortgage backed securities, looking into their portfolio risk for various clients. And from there I learned a lot. I mean we worked with the Fannie Mays and the Freddie Max of the world, so still very policy oriented. Yes, exactly. Chris Caplice: And this was post two thousand eight, right? The collapse had already happened, okay. Olivia Hu: Right. And so risk assessment was more important than ever. and there were lots of regulations around it. and I I I loved working in that space, but I wanted something that would move a lot faster. And so I moved from New York over to Silicon Valley, back to San Francisco, back to the Bay Area where I went to school, and decided that let's go smaller, let's go work at a startup where you roll up your sleeves and you can really make a big impact on the business that you're building. And so So I started working at Open Door and we were in around two markets at the time. we were in Arizona and Texas, and then when I Chris Caplice: An open door is more of a real estate play? Olivia Hu: Exactly. So open door is a real estate startup where you get an offer for your house site unseen. And that leverages a lot of the technology, the data science, the models that we were building at the time to give you an offer on your home site unseen, so you can go and buy your second home and quickly get an all-cash offer, for example, and not have to go through that process of, you know, having people visit your house. And so we built that company. but I had Friends actually, when I was at Stanford, who started working in the self driving industry, and I caught wind of it and started to learn more. And I realized if I really wanted to have a big impact on the next technology that's going to change the way we live, where we work, etc., it will be in self driving. It is a revolutionary technology, I would say almost as revolutionary as the internet. And so I decided to join a self driving startup, Drive AI. That was started by one of the Stanford professors, Andrew Ng, and his research students, graduate students at the time, and learned a ton. And we worked with OEMs, we worked with Tier One partners, we worked with government entities, commercial leads as well, and really tried to bring the offering to market as well as deciding on kind of a scale and OEM strategy. Chris Caplice: Right, right. Olivia Hu: Unfortunately, you know, back then, You kind of see the peak of the technology, but you can also hit the valley where once people realize, especially investors, realize that this technology is going to take a lot longer than originally anticipated to scale and to actually we were seeing social adoption at the time, but it was very limited. we were essentially providing you know incentives for people to be more open minded to actually take the the cars that we've Provided in Texas. And we had two deployments, one in Arlington and one in Frisco, Texas. And adoption really varied. I will say that the technology was there. It was very limited. It had specific routes. However, people were skeptical. There were still a lot of questions back then. I will say it's vastly different now, and we can go into how that has evolved over time as well. But from there, unfortunately, we hit a valley. We sold the business to Apple. Apple, and that's when I was brought over to Uber Freight and started the autonomous trucking business here back in 2019. Chris Caplice: Okay. So the trucking's a little different, but let me you you raised a bunch of interesting points. I kind of hit something similar when I was trying to so I got out of the army and decided what did I want to do? I was a civil engineer and all that kind of stuff. AI didn't exist, supply chains didn't exist at that time. And so my choice was do I want to get an MBA or do I want to get a master's degree? Do I want to work for the government? And I decided, kind of like you, I yeah, I'd like to change the world and do all this, but I think the faster path is in the private sector. because the government's it takes too long and it's too slow. And I think you can do more in the private sector than you can do with the government has a lot of good intentions, but it just moves a little slow. But I had this I made the same kind of choice. I don't want to work for the government and build roads anymore, so let's let's look at trucks and see what we can do there to make it operate better. and you talked about something really interesting. You so you yeah, the acceptance, social acceptance. Has really changed for for autonomous vehicles, especially for passenger, but also the technology has improved dramatically. So which do you think was the thing that unlocked the amount, the amount of upload for the autonomous driver vehicles, not trucks yet, for but for the POVs? Was it was Olivia Hu: Yeah, for the passenger Chris Caplice: it the technology that got so much better? Was it social acceptance? Were it combinations, something else? What do you think happened? What unlocked that? Olivia Hu: Yeah, I think it's the chicken or the egg. definitely the technology helps a lot. So the fact that the technology can go anywhere, the more the ODD expands, the more people can trust that this is a real technology that fits into their lives. And I I think the key that I'm learning is that it needs to meet the customer where they are. So if you tell me there's a Waymo here that only goes from this street to this street, I might never take it. But if it can go anywhere I need to go, which is how my life options. Operates, then it's a nice substitute to my usual mode of transportation. And so I think the technology caught up. And then the usual pace of adoption where that you have your early adopters and then you have your followers. And now it's so common that for example, when I talk to our shippers who used to not believe in this, now they see it in their cities. You know, wherever Waymo is unrolled, all the headlines, getting to experience it for themselves, getting your kids in the car and and experiencing it, I think that really changes people's emotions and beliefs around it, and that's where we see adoption really start to penetrate. Chris Caplice: Yeah. One of the best phrases I heard for that is that it going into a driverless vehicle, it goes from miraculous to mundane in like three minutes. You know, it's amazing at first and then after a while you're just like, Whatever, I'll read my paper or do something else. So Olivia Hu: Yes. Right. Chris Caplice: So that's so you took a circuitous path, you weren't in college saying I wanna work in trucking, but Olivia Hu: Yeah. Chris Caplice: you ended up there through the autonomous side. So what have you found that's d different in the trucking industry that surprised you versus when you're doing passenger? Olivia Hu: Yeah, so I will say my dad was a truck driver, actually. Chris Caplice: really? Olivia Hu: so yes, so he owns a C he has a C DL, he started his own flea, he was operating in the northeast corridor, so had to deal with all the traffic, and you know, snow and all Chris Caplice: When when you say northeast corridor are you talking New England? Olivia Hu: yes, I'm talking about New Jersey, New York, Pennsylvania. Yep. Chris Caplice: Wow, that's that's not easy roads. Olivia Hu: Exactly. And to this day my dad refuses to use Google Maps. He knows the roads so well. He does not use Google Maps. And sometimes we go in a circuitous route, but still, we get there. Yes, exactly. Chris Caplice: He gets there. That's funny. father in law is the same way. He doesn't use it. He thinks that Google lies. It lies. She lies. It's so funny. Olivia Hu: Right. I was exposed to trucking when I was young. but I never thought it was very cool. I I wasn't planning on following my dad's footsteps or or anything along those lines. And and so I, you know, I I went to Stanford, it was very driven by the tech mentality of like making a difference in the world and how do you do that and not being complacent with the status quo. and as I got more and more into the trucking industry, it's like they say, it's really hard to leave. I see it everywhere now and how it is actually the backbone of our economy, how it shapes everything about our lives and the goods that we have in front of us. You know, I get Amazon deliveries, I Half, you know, like I have a a can of a beverage here in front of me, you know, everything comes from a truck. And so I'm really I I I see how impactful it is and how big the total addressable market is in trucking. And so I think the big difference is passenger vehicles, it is all about adoption. It is all about how do you change the beliefs and the emotions of people. and that can be marketing. It can be through social it content creators and influencers. Right? It's a very different go-to-market than working with an established industry where there is a playbook where our shippers have networks that need to loads that need to get shipped every single day, and we cannot see any kind of disruption. There's so much more at risk. You know, these are products that people depend on, that other businesses depend on. And so I think there's a lot higher stakes, but also a much bigger opportunity as a result on the trekking side. Chris Caplice: No, that makes that makes a that makes a ton of sense. so give us an overview of Uber Freights dipping their toe into autonomous trucking. I have to tell you though, the very first Uber freight conference was in twenty sixteen, maybe? Back when you guys were on Mission Street. I went to it back when Bill Driegert was there, Lior was there. I actually I don't know if Lior was there yet. And they had the auto truck that was there and we drove on that. can you give us a background or overview of what the autonomous vehicle side of Uber Freight looks like? Olivia Hu: Yeah. And and so when Uber Freight first started, you know, our journey started with a core belief that this industry can be transformed with technology, especially with autonomous Chris Caplice: Mm-hmm. Olivia Hu: technology. It's not a matter of if, but a matter of when. And so we leaned into it. we were building autonomous technology for a while on the Uber side, but ultimately decided actually the partnership strategy, given how capital intensive it is, and that is not our core business. Business, right? Uber is a marketplace that is not our core business of building assets. And so we ended up choosing to go towards the partnership strategy. And when I was brought in in 2019, I built out the strategy, the operations, the day-to-day that we see today, and the partnerships. And my goal is not to have headlines because I was at Drive AI and I've seen that you know show before. And headlines don't work, they don't stick and things change. And you move on. But what was really important was actually to have long term commitments where both parties, you and your partner, are both committed to the same North Star. And how do you build that out over the course of five, 10, 15 years together? This is a marriage, and this is not easy, but we are going into this early because we know it's not easy and we need to start building this together. And so that included commercial operations, technical streams. etcetera to build out this network of autonomous capacity on Uber freight. And I wanted to be the network of choice for all of this autonomous capacity in the future, but how do I start building that today back in twenty nineteen? And here we are seven years later, I mean time flies, so Chris Caplice: Yeah. So so the first partnership that I that I could see was with Wabi. Is that correct? In twenty twenty three? And so you're building out the the driver as a service. Can you tell us a little bit about that? How'd the Wabi relationship come up and what does driver as a service mean? Olivia Hu: Yeah, so actually w our first partnership was with Waymo. when Waymo had a truck Chris Caplice: okay. Olivia Hu: yes, with when Waymo had a trucking unit with Waymo Via, and we had announced that and and we had started working on on loads and deployments, and we had a long-term strategic partnership with Waymo. And unfortunately their leadership decided to change course and really focus on passenger vehicles. And they paused the program. And then our next partnership was actually with Aurora. And so with Aurora, we've been working for many years as well. Yep. Starting first on a very small corridor down in Dallas, and then expanding to Dallas to Houston, and then from there expanding to Dallas to El Paso as well. And Aurora, similarly, we applied a very strategic long-term mindset, and we signed a multi-year agreement with them across all the various work streets. I mentioned technical integrations, having operations running daily on our network with their capacity, and then on commercial fronts as well. And then WABI is the newest entrant, and as the newest entrant, they have many advantages, like the latest AI research. And so we've partnered also with Wabi on understanding how does that their go-to-market work? How does their product change? They have a very strong simulation engine. How can we leverage Chris Caplice: Yes. Olivia Hu: that for for understanding how networks will change, for example, when we incorporate A V capacity, this new mode of transportation into our customers' networks. Chris Caplice: So so d how do you differentiate between the two? Is this more of a you're diversifying your risk because you have two suppliers, or do you see them as complementing each other? How do you view the two partnerships? Because they Olivia Hu: Yeah, I Chris Caplice: are technically competitors, yes? Olivia Hu: Right, right. And and when I think about it, for example, we are a network that's looking to bring and facilitate loads and bring autonomous capacity into our network. So for us, it's what can the trucks actually do? and while on the surface it may seem, you know, they're all the same, it's just a truck that goes from A to B, it's not necessarily true. For example, in the industry, one might think Schneider, Werner, you know, Covenant, Hirschbach are all Similar, but if you look under the hood, they all have their differences. and similarly, here I will say that Aurora Wild. You know, Tor Kodiak, they all have slight differences, and we work very closely with their teams to understand what's your ODD, where can the trucks haul, you know, and get really into the nitty-gritty details of well, how much can you haul? what can your sensors tell us? What information do we receive back, for example? Or if something doesn't go according to plan, because that's just how trucking works, what can we do to respond? The response plan can be different. Maybe your safety framework. So we look really deep with the teams into understanding how they differentiate. But for us, we are a marketplace that brings and aggregates all of that autonomous capacity here to bring to our customers, our shippers. And so for us, on the surface it looks the same, but we really go deep in understanding the differences. Chris Caplice: makes that makes sense. Now you used an acronym, O D D. Did I hear that correctly? Olivia Hu: Yeah. Chris Caplice: What does that stand for? Olivia Hu: So that stands for operational design domain, which is specifically what can the AV do? It is, can it go anywhere and everywhere? Can it go within a certain, you know, boundary on the map? Can it haul a certain type of commodity? Can it only haul a certain type of trailer? so that's what ODD means. It's like what are the requirements and the constraints of this AV? Chris Caplice: Because as I as I talk to more I'm talking to all these different com companies and understanding they have very different approaches how they came in and how they developed, are you finding that they are diverging or are they starting to converge in what they can do? Olivia Hu: That's a really good question. So I would say that right now they're all slightly different, like I mentioned, slightly different flavors, Chris Caplice: Yep, sure. Olivia Hu: different approaches. but are they converging? I mean, at the end of the day, I think that we are offering the same service to our customers, which is how can we fit into your network and how can we offer the best service, how can we offer the best on-time performance at the lowest cost possible to our customers? and so the North Star, I will say. is the same. we see how it can help the industry as a whole, but how they each approach it is slightly different. Yeah. Chris Caplice: Yeah, I think where they all started from is the big difference. And and so we'll see how that converges. Now you're also in charge of EVs, electrification of trucks. so which do you think will come first? Strong A V adoption in trucking or strong EV adoption in trucking? And let me say with middle mile for the most part. Olivia Hu: Yes, right. That's a great caveat. I am only focused on class eight, middle mile, long haul. So so I I will say with EVs, there's a lot of infrastructure requirements. Chris Caplice: Yeah, sure. Olivia Hu: as you know, with charging and now there's a competition for power on the grid with data centers as well. So that further exacerbates the problem. And maybe I'm biased, and given that I used to work in self-driving, but I I do think that A Vs Will be faster to adopt and accelerate. There are so many tailwinds that are accelerating adoption with AVs right now. Everything from regulatory, Chris Caplice: Yeah, yeah. Olivia Hu: the technology is ready. You know, we see this AI race with China. Like there are lots of tailwinds that are helping accelerate the adoption of AVs. So so I do see more potential with A-Vs than AVs. Chris Caplice: Definitely for the middle mile. I I would argue for first mile, last mile, local delivery EVs are already there and and dominating. Olivia Hu: Absolutely. yeah. Chris Caplice: Because they the circadian rhythm they can do and it it just makes more sense. So it's kind of funny how AV and EV are adopting in different segments of of transportation. That makes that makes a ton of sense. but let's talk about geography. are there any pilots or any use of A V trucks outside of Texas, New Mexico? Southern California, Arizona. Olivia Hu: You know, there there's testing. There's definitely lots of testing everywhere. You know, there's testing in Pittsburgh. so maybe tru Mm-hmm. Yep. Mm-hmm. Chris Caplice: Really? For trucks? For class eight? Cause that's not that's tricky. Olivia Hu: Yeah, with snow and all the winter conditions. But there's definitely testing happening. I know some that are going to Florida, to Atlanta, trying to go to the East Coast and understand how hurricane season is going to impact the sensors. But I I think that when it's ready to scale is the question that we're all trying to answer. the testing is definitely happening, but when it's ready to scale nationwide across every single freight corridor, every highway, still still a question. Yeah. Chris Caplice: Sure. There's gonna it's gonna be state by state, because a lot of the regulations are statewide. I know that Texas is is pretty consistent, but I know there was trying some legislation they're trying to pass to say any autonomous vehicle has to have a person in it. and luckily I don't think that passed, but it kind of shows the sentiment. so we'll see how that how Olivia Hu: Right. Chris Caplice: that changes. But are there so there are different stages. One you have a pilot and there's pilots happening different places, then it's kind of commercialization. And then it hits scale. Where is the Texas that Dallas was it Dallas Houston and and Olivia Hu: Mm-hmm. Yeah. Chris Caplice: El Paso? Is that where is that? It's a past pilot stage. Is it commercialized now? Olivia Hu: yeah. it's been commercial, I will say, since we started back in twenty twenty one. we've been hauling commercial freight, real customer freight, they're all paying, loads at pay. Chris Caplice: Truly no human in the cab? Olivia Hu: Not truly no human in the cab. Chris Caplice: Okay. Olivia Hu: they all have a supervised vehicle operator who is supervising the vehicle at all times. so I I think you if you're talking specifically about fully driverless and no human in Chris Caplice: Yes. Olivia Hu: the cab, you know, we we will have to wait and see when that is fully ready. first from a technology perspective, then from an OEM perspective when they're building the redundant platform, and then last but not least from a shipper adoption perspective as well. I I think there's a lot to be to we're still waiting to figure out what that means. Is it a hub to hub model? Is it an end-to-end model? how does that work? Chris Caplice: That's one of the big distinctions as I talk to different companies, the the end to end versus hub to hub. And then I was just talking to the founder of Humble, and so they're doing more of the drayage side, and it's EV AV at the same time, totally changing the format. So it's interesting to see with a new technology, like you described, it's like when the internet came or when cell phones first came. There's so much innovation, no one knows what it looks like. But at some point we'll hit a dominant design. It'll come, whether who knows when, ten years from now, twenty years from now. But it's fascinating to be here in a time when there is this much diversity and experimentation to see what what happens. Do you Olivia Hu: Yeah, and I think we I think we have a great opportunity, you know, you and I as being part of this industry and watching this happen, is we get to shape what it will look like ten, twenty years from now and what it will look like for our kids and our grandkids and future generations. So it makes it very exciting. And that's why I love it and that's why I've been in it for so long. Chris Caplice: Was your father happy or got involved in trucking or did he was he upset? Olivia Hu: He was very skeptical. you know, as someone who knows ro the roads and driving, he was very skeptical that it will ever handle all the conditions that he has seen and every single scenario that he's replayed, he's like, but what about this accident that happened on the bridge where you know I'm like, that's a great question, Dad, and we'll find out. But I think so. Yeah. Yes, exactly. Chris Caplice: Yeah, we always focus on the focus on the corner cases. Yeah. so I've got to ask, do shippers care if it's an autonomous or not? Is that the is the is the idea eventually where the shipper really doesn't care as long as it meets the characteristics they need for that load, whether it's a drop and hook at both ends, live load, whatever, do you find resistance from the shippers? What what is what does it take to get them over the the hump to accept an A V? Olivia Hu: That's a great question. So I do think that shippers, as table stakes, it needs to meet their service performance, it needs to be cost effective, competitive with other alternatives in their transportation networks. that's just table stakes. Now, whether or not they care, it is the cherry on top when they have additional benefits like fuel savings. right now, with high fuel surcharge costs. We see a new heightened interest from our customer base asking about autonomous capacity. Or in a tight market, we also see interest in autonomous capacity, right? The the the driver shortage, the delta between demand and supply, that's what's driving a lot of it. But as we know, markets are constantly cyclical. So things will change again. Chris Caplice: Sure, sure, sure. Olivia Hu: Will this last and how long will this sustain for? We don't know. but shippers, I think the conversation is now different from what it was before. I think part it's also the development on the passenger vehicle side has created more believers in understanding that this technology is here and it's real and it's about to start scaling and so we are seeing a lot of interest across the board from various customers. Chris Caplice: Yeah, I think I think consumer acceptance is really challenging and I'm curious if do you see that as Waymo goes and autonomous vehicles goes, the robo taxis go out, that's where the next trucks will go. It's kind of like paving the path. but you mentioned something else about fuel. So what drives the fuel savings for A V versus a regular seasoned driver? Olivia Hu: Yeah, so fuel savings are Estimated to be somewhere between I think 10% to 20 plus percent. and these are all estimates because the technology is constantly changing, Chris Caplice: Sure, sure. Olivia Hu: you know, the base vehicle is going to be different eventually, but we see a couple things. Operational efficiency. so these trucks are more operationally efficient. they're not suddenly accelerating. There's no sudden harsh breaking because it can see further and potentially react faster as well. And so as a result. That level of fuel consumption gets eliminated from like a driver's lead foot. And so we anticipate that there will be lots of fuel emissions savings and fuel just surcharge savings as well. And for the the time being, we see that all autonomous trucks will be on diesel engine on ice vehicles for now. And especially since the use case that they're solving is for long-haul trucking. So EV trucks today can't really do a thousand miles for. Chris Caplice: Right. And so th it's interesting 'cause I know that some of the fuel savings comes 'cause a an A V can go a slower speed and it can drive longer because it is not a subject to hours of service. that's one thing. But it it's curious to see how much that's gonna change. But it is very important now. It's just looking up the price of fuel, it's cresting at six dollars a gallon national number two diesel, which is crazy for that. And the and because the shippers pick up the majority of that cost through the fuel surcharge program. so that it is a driving force and it's driving some to intermodal, obviously too. You do a mode choice if you can. So what is the big selling point then when you 'cause you have to get the shipper's permission, right, to Olivia Hu: Mm-hmm. Chris Caplice: say it's gonna go on autonomous or not. What where is the resistance? Do they are they worried about insurance? Olivia Hu: Some some ask about insurance, some ask about liability. I think the the biggest question I get oftentimes from shippers is can it do what I need it to do? And every shipper has their own unique needs. You know, some need, for example, specific type of trailers like welded doors, or some need specific types of security and locks, or some have, you know, appointment times. that are not very favorable or they have long loading times. So it's all very operation specific, I will say, for each customer. And even at on the customer level, on like the lane level, it depends on their warehouse and and what they're ready to do. And and so the resistance isn't really something specific to autonomous trucking. It's more are these autonomous trucks able to meet my supply chain needs where they are today? And every customer has a different need. Chris Caplice: Yeah, and because there's more to a driver than driving. Right? And so the autonomous that Olivia Hu: Right. Exactly. Chris Caplice: yeah. So let me let's just step through it. Let's say I'm a shipper. Are you are you selling mainly to shippers or to carriers? These services? Olivia Hu: So Chris Caplice: Who who's your primary customer? Olivia Hu: Right, so at Uber Freight, our customers are shippers and we sell transportation Chris Caplice: Shipper. Okay. Olivia Hu: services to shippers. Right. on the managed transportation side, we manage our shippers' networks. and so we we have both pieces, but when we're working with the autonomous capacity, we are specifically working with our customers who are shippers and understanding does an autonomous truck fit into your network's needs? Does it help improve supply chain and transportation needs if Chris Caplice: Got it. Olivia Hu: so then we will deploy our autonomous capacity wherever it fits and depending on the use case. Yeah, great question. We don't have to address Chris Caplice: So so who owns the truck? Who own so so the w pick pick Aurora or Wabi or we can do one or the other, let's say it's an Aurora truck. Who owns it and how do you coordinate with that? Are they their own carrier? Olivia Hu: Yes, so today Aurora offers transportation as a service and so today they own and operate the truck when they're hauling up to work. Today. Mm-hmm. Eventually, Chris Caplice: They Okay. Olivia Hu: starting next year, they're going to offer driver as a service and they will no longer own Chris Caplice: Okay. Olivia Hu: the truck. And Depending on what type of asset owner wants to purchase these trucks, it will come with a driver already built in, which is the Aurora driver, or the Wabi driver, or whomever it may be. And so these trucks will be offered on our network, but the carrier on file may be different from where what it looks like today. Chris Caplice: But it will probably be an established carrier with authority. Olivia Hu: Right. It could be an established carrier. We can see new up and coming carriers that are A V specific, for example. We already see that today with EVs, for example. Some of our carrier partners are purely EV fleets, and we might even see that in Chris Caplice: That makes sense to me. EV makes sense. Do you see middle mile AV only fleets happening? Do you think that's a Olivia Hu: I o Chris Caplice: short term, a long term, or what do you think? Olivia Hu: I think in the long term, that will be a possibility where there will be pure play A V fleets. just the model is so different. For example, today our fleet partners, their job is to recruit and retain drivers and to manage a driver's preferences, their schedules, etc. In the future, with an A V truck, you no longer have a driver, and so their new job as a as you know a fleet is to maintain these trucks, is to make sure they're ready to go. to hire people to do the all the other parts of of that it's not related to driving, you know, like you check the trailer, you check the vehicle condition, maintenance, et cetera. But the driving and the recruiting and retaining drivers no longer becomes the core business. And so how does that change? And I think new players will emerge as a result. Chris Caplice: Yeah, one of the interesting things that the that makes truckload middle mile so different is because it has economies of scope, right? You need the backhauls. You need empty miles are a big thing. If I go from Dallas to to El Paso, you know, well you gotta make sure something's coming back or something's going somewhere else. And so right now you're on two corridors and eventually you need to have a network as you expand that out. Do you think that having AVs will minimize the amount of empty miles? Or will it just make them not as painful? Olivia Hu: my goal is definitely to make sure we have a network that is dense enough that we create those types of efficiencies where we minimize empty miles across the board. Course, you know, that means obvious which is why I'm starting now and not waiting for to be a Chris Caplice: Right. Olivia Hu: follower. because I want to start building that network now, that density, fatil facilitating those loads that our customers need, so bringing the demand for the autonomous capacity while simultaneously also bringing the supply of autonomous capacity. So yeah. Chris Caplice: D do you see a s a certain length of hall for a sweet spot for this? Is it like over five hundred miles? So you you cross you can do in one day what a usual human driver would take two? Do you look at it that way or is it are there other factors? What makes you decide that this lane makes sense? Olivia Hu: I I think we're I'm very open to all the different ways that autonomous capacity can add value. Sometimes it's just we need capacity, right? Sometimes it it's not a length of all question. It's you know, on a certain lane that no one wants to go to, or maybe for example, going up from the Mexican border, we need more capacity there as we import shipments. And and so it's just where do we need the most capacity? I I will say that we do see a lot of tangible benefits when we can when we no longer have hours of service limitations, when it can go and have faster delivery, and our customers are very excited about that. Faster shorter transit times and faster deliveries. Chris Caplice: So can Yeah. Can you give us any sense? I don't know if you have these numbers or can release them at all, the volume and what the typical length of haul is for these autonomous trucks. Are they within two, three hundred miles? Are they like a a back and forth between a DC and a a plant? Or are they longer? Can you give us any sense of volume and the types of lanes that you currently have moving? Olivia Hu: Yeah. Yeah, I I will say the volume will vary. you know, we haul on a daily basis in the in the Texas corridors that I named. and the length of haul Chris Caplice: You don't do the triangle, do you? Do you go up to Dallas, Houston, was it San Antonio and back up? Olivia Hu: So no San Antonio today. Yep, just Dallas to Houston or Dallas to El Paso Chris Caplice: Okay. Yeah. Olivia Hu: a and the and vice versa, El Paso to Dallas and Chris Caplice: Got it. Olivia Hu: and up to Dallas. Yep. that's where most of the hubs are and and so that's where we're hauling today. Once they're ready to expand, we're happy to go and expand further into the Texas triangle. but we're really waiting to see when that opens up and we go back to the O D D conversation we mentioned earlier. So Chris Caplice: Yeah. And so right now the cost having the driver there, even though he's not driving, the what you called him something else. You didn't call him a driver. Is this an observer, a safety what what is their role? A what? Olivia Hu: A vehicle operator. Vehicle operator. Vehicle operator. Chris Caplice: Vehicle operator. When that eventually goes away, which hopefully it will, the economics should improve dramatically. Is that a fair statement? Or d will it just be an incremental? Olivia Hu: That's what we anticipate is that the economics will improve, higher utilization for fleets, and then we also see potentially lower insurance costs as well, lower risks, less damage, cargo damage, et cetera. so across the board, I anticipate better economics and who reaps Chris Caplice: Yeah. Olivia Hu: the values. That that includes the carrier, I think lower cost for them, that includes the shipper, and then hopefully for us as well. Chris Caplice: So we talked about what shippers why they might hesitate and the things that get them over the heap to accept this. What about carriers? Do carriers have objections to this? Is it the cost? Is it the safety and insurance? Is it the loss of jobs? What what where do you see the biggest challenge for carrier adoption? Olivia Hu: Yeah, I I see carriers are very interested and I've talked to a number of fleets on our network but also across the board. Carriers are asking more about how does this contribute to my business? what is the TCO? How does this add value to my business? Are there shippers that are interested? they're doing the math of like what are the margins? What does this look like? And that's what we're all trying to figure out as well. It's not just before it was, let me understand the technology better, etc. Now I think people really see it coming and believe in it, and now they're starting to plan well how do I incorporate ABs into my business and what does it look like from a margin profile? Chris Caplice: And so we talked a little bit about the market cycles and right now we are on a pretty tight cycle. We'll see how long it extends, but we're coming out of a three year trough where spot rates certainly declined, but contract rates were kind of flat. do you see the tight capacity accelerating or slowing down the adoption of A Vs? Olivia Hu: Yeah. like I mentioned before, I think the tide capacity is definitely helping. the A V demand and the need for capacity rates going up. I will say because markets are always cyclical, I anticipate that will change as well. but overall we are in this going up and to the right in the same direction, but like this. Yeah. Chris Caplice: And so right now we're operating how many loads per week on average, rough figure, is it less than ten, a hundred, a thousand along the El Paso, Dallas, Houston? What what's the order of magnitude? Olivia Hu: Yeah, I will say we're in the dozens. Chris Caplice: Okay. Olivia Hu: we haven't I mean the real question is how much capacity is available out there. And the reality is the capacity is very limited right now. both from an OEM perspective, from an A V provider perspective. And so that's what we're limited by. If we had more capacity, we would definitely be moving more, but unfortunately today it's it's dependent on what fleets have available. Chris Caplice: And does Aurora have you said it's a long tenure agreement? Is did I hear that correctly? Multi-year agreement. Okay. And do Olivia Hu: A multi year agreement. Mm-hmm. Chris Caplice: they have is there an expectation by the end of 26 or end of 2030 to have a certain amount of, is it autonomous miles or autonomous vehicles under your system? Do you have a goal that you can share with that? Olivia Hu: Yeah, well we have many goals. and one of them being we want to have as many autonomous miles as possible on our network that we fill with our customers demand. And so it obviously all depends on what the truck can do. So it depends on if what use cases and what applications are best suited for autonomous technology. And from there we're gonna match it and pair it with our customers' networks, where it best suits our customers. And at the end of the day, we're looking out for our customers' freight and we want to only use autonomous capacity if it actually improves their supply chain. And so so our goal is to maximize autonomous capacity on our network, but this assumes that the autonomous capacity can meet our customers' needs. Chris Caplice: Yeah, so as you know, being in trucking, there's two types of loads out there. There's spot and contract, right? And contract are where they you reserve the you set a price up and that carrier quote owns it for the year at a certain price, and that's eighty percent, ninety percent of all truckload moves. But the spot is the dynamic piece. Where do you see AVs fitting in? The consistent contract late rates, those those lanes that are kind of power lanes moving back and forth. Or those irregular lanes, the spot lanes that come up, the new plant, the new new kind of things happening, or when the routing guide fails. Do you see the place in contracts long term consistent or the dynamic spot side? And you can't say both. Olivia Hu: I would no, you stole my answer. Yeah, Chris Caplice: Yeah. Olivia Hu: well, you know, I think it needs to be compared to like a regular truck today where Trucks you know today can do both and it depends on how the fleets run their business. of course certain fleets have you know a portfolio is part of their portfolio is spot and part of their portfolio is contract. and I anticipate it will be the same and every asset owner will run their A V trucks differently. maybe some will only do dedicated contract and maybe some will choose spot, but it's on the same lane every day. So so I I I think it really varies. Chris Caplice: But it's interesting because it it seems like if I have a steady lane, it seems to me the big sweet spot would be a 750 mile lane that I can now do in one day and just let that go back and forth. Because there's a lot of those lanes from large manufacturing plants to DCs that happen on a regular basis. to me that'd be a massive savings because then you don't have you know the overnight stay and it have to rest. Now do the The observers to the the operators, do they are they subject to hours of service rules? Olivia Hu: Yes, yes. Chris Caplice: Okay. Olivia Hu: the vehicle observers today that our partners use are all still subject to hours of service, exactly. Chris Caplice: Okay. So so we haven't even unlocked that major piece yet. That's the real thing that'll unlock the A V adoption 'cause then the price the cost of it will will certainly go down, I would think. Olivia Hu: Exactly. Once you remove that vehicle observer and it's fully driver out, there's no human inside the cab, that's when this will get unlocked. And the anticipation is that s for some of our partners, it's this year or next year where this will start to happen. and then Chris Caplice: Nice. Olivia Hu: when are these trucks ready to actually start scaling? You know, when do they receive trucks that are ready to go? I think that's what we're waiting for right now. Chris Caplice: Yeah, 'cause I know there's some FMCSA issues like with the s silly things like the safety triangles. If no one's in the cab, who's gonna put the triangle out? And there's there's issues with that and whether you can do something on the trailer. So that yeah, there's th that's the big thing as I talk to more A V companies and people affiliated with this is that it's more than driving. That's just that's the big piece of it, but there's all the ancillary pieces to it. well, Olivia I've had you here for a long time, so I really appreciate it, but let me ask you one more question. what year will fifty percent or more of long haul over the road full truckload miles be autonomous? Olivia Hu: Oof. I like the caveats. Chris Caplice: Over half. Olivia Hu: Right. Over half, long haul. I like those caveats. Right, dry air. Right. No Chris Caplice: Yeah. Full truckload, class eight, middle mile. How when will the middle mile tip to fifty percent? Olivia Hu: no hazardous materials. yeah, I like all of those. Yes, Chris Caplice: yeah, okay. Take the hazmat out. Okay. Olivia Hu: exactly. over fifty percent. Oof, that's tough. Man, I would say give me another At least ten years. Chris Caplice: Wow, that's fast. That's faster now. Olivia Hu: I know, I know that I I'm being I'm being aggressive and and and optimistic only because I I see so many benefits from autonomous trucking and potential lives Chris Caplice: Yeah, yeah. Olivia Hu: saved and like the safety benefits as well. So I you know, I talked to my my dad about this, like I mentioned, like how much safer it is and he agrees actually. He that one he did not fight me on. He was skeptical about what it could do, but he was not skeptical about potential lives. Saved. So there you go. Mm-hmm. Chris Caplice: Makes sense. Makes sense. Well, Olivia, thank you for joining me today. I really appreciate it. Olivia Hu: It was a great conversation. Thanks for having me, Chris. Chris Caplice: All right, everyone stay tuned for the truckload market update. This is the truckload market update for 15 October 2026. In the dry van market, we saw the change in active contract rates increase 1.5%. Spot rates actually stayed flat over the last two weeks. The current level of replacement rates for dry van is positive 10.1%, which means on average, new rates coming in, new contract rates coming in are about 10% higher than the rates they're replacing. We saw the market gap between spot and contract, eight cents a mile, positive, where spots above contract for that dry van market. Year-over-year change in spot rates for dry van is 35.2% and 13.3% for contract. For temp control, we saw contract rates increase slightly, 0.3%. Spot rates actually dropped a little, 1.1%. The current level of replacement rates for temp control is a positive 9.3%. The gap between spot and contract for temp control is strangely enough negative one cent a mile. spot is below contract. I think it's a little blip in the data. We'll see if that irons out, but that's where the data is this time period. For temp control, we saw a year-vear change in spot rates, increase 21%, and increase 12.4% for contract. Intermodal saw contract rates drop 0.5%, spot rates dropped 2%. Current level of replacement rates for intermodals 5.2%. And the gap between spot and contract is negative two cents a mile. But again, as I say, every every one of these updates, there's not that much spot and intermodal. the overyear change in spot rates for intermodal is 9.2% and 10.4% for contract. For flatbed, we saw change in active contract rates drop 0.3%, spot rates increased 0.7%, so just like kind of fluctuations. Current level of replacement rates is about 10.5%. And the gap between spot and contract for flatbed is two cents a mile. change year over year for spot rates is 29%, and 15.1% for contract. Okay, what do we see? The active contract rates were mixed. Drive in and attempt control increased slightly, flatbed and intermodal dropped slightly, like one less than one percent, give or take. spot rates turned mostly flat or slightly negative, with the exception of flatbed, which rose less than one percent. Most other ones stayed flat or went negative. Replacement rates remain high and are within a few percentage points of the last period. They're almost all double digits, right right there, low double digits except intermodal, which is about half that, about five percent. This means RFP is coming in, you're gonna see an increase. And it really depends on when your last bid happened. If it was in the spring, you're not gonna see as big of a jump as if you had it in last November. So just be prepared for that. Look at your year over year and when your rates were put into place. The gap between spot and contract contracted for all but flatbed and temp control and turning negative for the first time in nine months. That's that was kind of a surprise. And year-over-year changes in contract rates range from 10 to 15%, and for spots between nine and thirty-five percent. interesting, the spot year-over-year rates are slowing down, right? While the contract rates year-very-year change are still increasing. It's that lag, right? Canary in the coal mine, that's what spot rates are. the leading indicator, we're seeing what's happening to contract. But anyway, that's the status of the market in fifteen October twenty twenty six.