Walt, this week you went from partying down on Bourbon Street to testing autonomy on Market Street, right during the middle of earnings reports from Lyft and Aurora. First of all, how the heck are you awake doing this? And second of all, what did you learn from the capital of autonomy development in San Francisco?
Well, thank you to our listeners for enduring getting this thing on Sunday. I know I like to advertise this as every Saturday, but having some extra sleep after this long week of not a lot of sleep was definitely helpful. You know, look, it was great again to integrate myself within the autonomy community. What a great group of people, very smart, you know, always trying to help to advance my understanding of what's going on because obviously this is like complex stuff and everyone has very strong opinions, and it's certainly very helpful.
Now, look, in terms of Uber, who's kind of, in my world, the largest public manifestation of this — yes, Google is larger, but Waymo isn't necessarily having as large of an impact on Google as autonomy is having on Uber. I think in this past podcast I've talked about how all of these headlines are potentially negative for Uber in terms of autonomy, but I think I'm coming to the kind of realization that there actually, in the middle of these, there are possible positive headlines as it relates to Uber and autonomy in terms of partnerships and some substantiation that fragmentation can happen in the market and OEMs might have a path for success, you know, obviously under a lot of pressure, which could benefit from Uber. And I think those will kind of intermingle, which is in our world probably good for traders to kind of understand the back and forth and the battles that will exist and people understanding, you know, what the future is for autonomy. So I thought that was kind of a pretty big takeaway for me.
More on LiDAR. My understanding of LiDAR continues to grow, although, and I think what I'm a little worried about is, you know, you've been kind of anti-Hesai and believing that, you know, it will get banned. I'm not sure that there is a legit replacement for Hesai in a lot of the companies that are using them as a LiDAR sensor. So I'm just curious from your standpoint, like what are you seeing in terms of the industry trying to work with government to say, like, like in telco, and I think I've talked about this before, when they got rid of Huawei, they gave operators a chance of a timeline to pull this stuff out and actually offered funding to rip and replace Huawei equipment. How do you think this Hesai thing is going to play out?
First of all, we're very honored, we have the CEO of Hesai coming on the Road to Autonomy podcast in March. That'll be a really special episode.
That's great. So we'll get some insight on that. That'll be after Hesai reports their earnings.
How I see it playing out is they're probably going to have to split the company in two, have a US division and a Chinese division, because Hesai is such a critical LiDAR component for both the autonomous vehicle side and the autonomous truck side. There's really no way forward, there's no alternative in the market. Companies have tried alternative LiDARs, nothing is nearly as good as Hesai, and Hesai is on every autonomous vehicle not named Waymo today. So that tells you about the quality of the product they're producing. Look, I thought TikTok was going to be gone, and just this week TikTok is now, you know, it obviously never really went away, it went away for a day, but it wasn't available in the app store until this week. So I think there could be a willingness on the part of the government, but I think that necessitates the industry sending the message to the Trump administration pointing out the importance of Hesai here in the near term to get a lot of these trials and early implementations off the ground, even if there is a longer-term goal of getting the manufacturing and, as you say, maybe splitting the company up.
Do you think that's feasible, or is it possible that in the very quick actions of our government that we're seeing in other implementations, as far as what DOGE is doing, that Hesai could fall victim of kind of a quick decision by someone in the administration?
There are political risks, but if we go back in history, Hesai was founded at Stanford University, the first office in Silicon Valley before they moved operations overseas. So to me, there is a very clear path to separate the company and have a US division, say have a ex-US division. The politics where that goes, I am unsure, but I will say and state for the record that Hesai is a critical component to the future of autonomous driving, and right now it's needed because, as I said earlier, there is no alternative. And look, I'm not going to commit to say that they're important for the future, but my takeaway is they are very important for the now, and the now meaning like the next 12 to 18 months.
If Sean O'Brien gets wind of this and uses this as a lever to try and throw a wrench in the works of autonomy, which, you know, might be perceived in our view — I know we share this — wrongly as far as bad for, you know, labor and the economy in the US, is that a risk that we have? Has O'Brien chirped up about this at all?
Sean hasn't chirped up about it, no, but it wouldn't surprise me any tactics that he would try. What really needs to happen, to cut to the chase, is that there needs to be an alternative. The engineers have to agree on an alternative. There is an opportunity in the market for somebody to advance technology to achieve Hesai's size, but nobody's been able to do that yet. And I have the CEO of Veo coming up on an upcoming episode, and they're moving much more into infrastructure, not really focusing on autonomous vehicles, autonomous trucks, moving towards infrastructure. So there's a lot of change coming in the LiDAR market.
Grayson, one of the reasons I was excited to partner with you in terms of Autonomy Markets was the library of people that you got, and even on this trip I met with some people, people that you had previous interviews in your library. It's great to see you've got now H and Oster lined up, so appreciate that and keep doing it. That stuff's awesome. And for those that maybe only listen to this podcast, you should look into Grayson's library, a lot of good background stuff there.
Let me move on to my personal experience with Waymo. I can't even remember the last time I was in San Francisco. I think we talked about it a little bit on this podcast a couple of months ago. It feels like the cars have gotten more aggressive in how they drive. I mean, I was making a left turn, there was a little old lady with a caretaker crossing. I thought the Waymo was going to try and cut in front of her. The caretaker gave the car a dirty look and then saw that there's no one driving it. Just overall, it felt more like a human in a very short period of time. And then in addition to that, Grayson, and probably more relevant, is I noticed what I would call an E-ZPass, like one of those RF pieces of equipment that enable you to get on highways. So it feels like that's a pretty big signal that they're getting the fleet ready to start sending those cars out on the highways, isn't it? The highway unlock's coming soon?
I would say the highway unlock's probably going to come in the next three months. It's coming. And if you look at every announcement they've made with a road trip, city, really good. But then if you go read the local newspapers, every comment, the San Diego local papers, the Las Vegas local papers, they all mention highway testing. It is clearly coming. That unlock is coming sooner rather than later.
My last — I had many takeaways, but the last I want to comment on here is Zoox. I saw a bunch of Zoox. I tweeted a picture of a Zoox testing car, but my sense from the industry just doesn't necessarily feel positive. I know you've said this before and I've pushed back in terms of, you know, kind of their progress and kind of a culture of moving towards quick progress here. You know, again, a lot of these people are probably competitors, so everyone likes to talk their book. Should we be concerned about Zoox? Like, isn't there — it feels like, by the way, when Elon, when Tesla started talking about autonomy, it helped everyone, right? It put pressure on perhaps Waymo and Uber to get closer together. Everyone seemed like they upped their game, which is great. That's exactly how a free market and competition should work. Is it happening at Zoox? Like, where is the sense of urgency at that company?
There is no sense of urgency, and I will direct this to Mr. Andy Jassy, the CEO of Amazon. Mr. Jassy, I have tweeted at you, sir. I have sent notes to your IR team as an Amazon shareholder, and a lot of the individuals that listen to this podcast are Amazon shareholders. We would really love clarity on the Zoox business, sir. Where are you going with Zoox? There is no clarity in the market, sir. All we hear is about the culture internally at Zoox. We had the Verge article that came out and talked about how the suspension was horrendous. That was The Verge said that, not me questioning. And The Verge further questioned the technology. So now there's questions about the suspension of the vehicle, there's questions about the autonomous driving that's out there in the public, publicly sourced to The Verge, not just industry gossip anymore. What is the path forward? How are these cars going to scale? And here's the big question: it's a lot more comfortable riding in a Highlander, and I've ridden in the Zoox Highlanders, than it is riding in the toaster. Zoox have to abandon the toaster and go to the Highlanders and follow the lead of the market.
Okay, so I'm going to push back on many, many things that you just said. First, I just think maybe there needs to be more aggression in terms of how they're dealing with things. Second, Verge — I mean, you know, there's plenty of things that Verge writes about a lot of different topics that I'm not sure that we necessarily need to assume is gospel by any stretch. And thirdly, I've ridden in a Zoox toaster. I mean, granted it was on kind of limited roads, maybe it wasn't like New York or the New Orleans roads that I just experienced, which are horrific. I thought the suspension was fine. I thought it was, again, a very comfortable ride.
Although as I do more work on this industry, I'm increasingly moving away from the concept of a purpose-driven vehicle being the answer, meaning, you know, either partnerships or otherwise using product that, especially when you get into financing and you talk about residual values of vehicles, having like an actual car as opposed to like a vehicle — what the hell are you going to do with it after the fact? I think these, you know, maybe not the best strategy for the company. So maybe they should take those testing vehicles, as you're pointing out, and use them as opposed to this, you know, this kind of belligerent and sticking with the toaster.
I don't know what causes them to stick with the toaster. We have no data on how much it costs to manufacture the toaster. We have no data on what it costs to maintain the toaster. We have no data on how many miles or how many years or rides the toaster is going to do. And second of all, outside of Las Vegas, which is a tourist attraction destination, we don't know if the consumer wants to face one another. I've said this many podcasts, I don't want to ride with anybody, I don't want to face anybody. And I did a podcast with my good friend Pete Bigelow about this and Sergey from Cruise. If somebody gets drunk and tries to open those doors, would they open? Because people drink in Vegas, let's be honest. No, they'll just puke in there and they'll send it to a service center to clean it up.
But Grayson, the cost of a Jaguar and the amount of equipment, all self-developed by Waymo, was also very expensive. And look, they're behind. I mean, they're where Google started how many years ago in that market. So starting with the expensive equipment, I don't think is terrible. But at some point, again, I'm coming more into your camp that it has to evolve into, you know, either some integration or using kits on existing vehicles, you know, to kind of get that going.
What if Mr. Jassy and the board of Amazon said we're not worth this investment and they decided to license the Zoox stack? What impact does that have?
I mean, I think that's great, assuming that they've got something that's licensable, that's something someone's interested in taking, right? I mean, obviously Nuro is in the market today with their technology targeting a Q2 hopeful announcement with an OEM and a platform partner. So I don't know, is it ready? I think my guess would be that they need more development work before you get to a point where someone would want to take it and develop it, or maybe you spin it out to the existing employees with a certain amount of money and just force them to do it. But again, that's going to cost money, and that goes back to this culture of urgency to getting to a product that is licensable. You know, when you're struggling for money in part, you know, and with small teams, you can accomplish a lot. It doesn't feel like they're in that position. Again, I could be wrong. Always look for inputs. So anyone, any of our Zoox listeners that want to convince us otherwise or provide us with data points that suggest otherwise, please reach out to us. Please do.
Now on to earnings. Let's start with Lyft. That was the biggest earnings of the week. What did you think?
Well, hold on, Grayson. I'm not sure what you mean by biggest. Lyft has a market cap of $5 billion, but there's another company, Aurora, that might not be perceived by investors as big, but after the 50% move, that stock is now worth $17 billion. So three times the market cap of Lyft, who's in the market today, obviously not a ton of revenue yet at Aurora. And for our investors out there that are listening, saying like, can I even play around in this name? It trades 14 million shares a day and it's a $10 stock, so it is tradable and it's big. So I guess let's start with Aurora.
Yeah, I mean, think about that, $17 billion market cap versus Lyft's $5 billion market cap. Let that in for a minute. That's pretty intense there. And Aurora stock for the traders, shut up, 30% this week because Aurora announced that their Dallas to Houston safety case is 99% complete. As we know from previous quarters, they have told the market that they are going to go driver out by the end of April, Dallas to Houston. I fully believe that Aurora is going to accomplish that major milestone. When they accomplish that milestone, how does the market react? And are we going through a $20 billion market cap on that announcement?
Yeah, so I don't — I mean, I looked at that call, they didn't really say anything new, but maybe it was just the relief of them not kicking the can down the road. And, you know, again, my takeaway from being out in the Valley this week, this past week, I think they are going to deliver on that milestone. But the question is, like, what's next? And I guess the what next, what's next is driving in the dark and also driving in weather. Less exciting milestones, I guess, but certainly something achievable. Getting to tens of trucks, you know, maybe that stuff is good. I think just showing, maybe doing investor demos where people could see this stuff. I'm guessing, like, I'm totally guessing, but are they going to just follow this first truck? Like, they're going to have 10 trucks with the customer, they'll follow the one truck without a driver to make sure nothing happens. Who knows? Maybe I'll follow it. I'll find this thing, I'll follow it. I'll have you in the passenger seat, we'll video it.
But I think the other thing that I kind of pulled out from this earnings call, which I think you can help me with, is understanding implementation. So Chris Urmson said this, I'm going to quote from the call. He says, 'Since our founding, our objective has been to deploy self-driving technology at scale. Our OEM and Tier 1 partnerships with Volvo, Paccar — is it Paccar or Packar? Paccar. Paccar, sorry, Paccar — and Continental are unmatched in the industry.' Then he goes on to talk about NVIDIA's Thor system. He says, 'Production samples of Drive Thor are coming in the first half of 2025 to start testing. Drive Thor is going to be the core, so this is the compute, the primary computer for the Aurora Driver, which we're developing with Continental, who will manufacture it. We also continue to make meaningful progress on other aspects of this generation of the hardware kit.' And then later in the call, Dave M — I think that's how you pronounce it, sorry Dave, I know you're probably listening — says, 'At the very latest, when we have our hardware kit that's being produced by Continental, that is expected to be a line-side install with their OEM partners, that will go directly to customers.'
So when I read those words, Grayson, this is what my perception is of what's happening. A kit is like a third-party, like that's an aftermarket add-on. Like if we think about Waymo and those parking lots where they've got the Jaguars and they're putting the kits on. And in this case, I think what David is describing is just taking a kit to a line. So you're having maybe the OEM deploy it there, but it's not part of the actual manufacturing process. Meaning that like at Waymo, when we looked at the Hyundai car, that doesn't look like a kit that's installed after the fact. That is integrated into, I think, the manufacturing line of that product. So is Aurora basically just doing kits? And if it's really just a kit and they're kind of the leader, how important are even these relationships with these trucking companies? I mean, I think they've been positioned in trucking as like the kingmakers, but I mean, if even Aurora is doing kits, doesn't that take away the power of, you know, Volvo, Daimler, others of being these kingmakers in the market?
It does. And what it further does, it further validates Don Burnett's strategy at Kodiak. Mr. Burnett gets beat up all the time, 'Oh, you don't have an OEM partner, you don't have an OEM partner.' But Mr. Burnett has the sensor pod, and with the hardware kits, which was clearly stated on Aurora's earnings call, is the same thing as Kodiak sensor pods. It validates that, because what Kodiak's doing today, they're going to upfitters and they're installing them. Aurora is just going to an OEM and installing them. Pretty much the same strategy at the end of the day. It just further validates Kodiak's strategy, that Kodiak was right not being dependent on one fail point with the OEM. They can go anywhere.
Are we missing, like when they talk about a kit, does that include the redundancy that they're adding? Or is it possible that the kit is just the autonomy technology and there's still a benefit to what Aurora is doing in terms of the quote-unquote partnerships of them implementing whatever redundancy that is required? Is there a fine difference there?
There's a fine difference. Aurora is delivering the self-driving technology in the kit. The OEM is delivering the redundant chassis. At some point, the redundant chassis is going to be available to buy on the market through a dealership because the dealers are going to have a cow. And so hypothetically in the future, Kodiak could go to a dealer or their partner, some we saw with Atlas Energy, can go buy a thousand trucks. Kodiak can go to an upfitter, put them on.
So should our lens on this industry change? Or at least my lens, maybe I was always looking at it wrong in terms of like locking in the OEM? Or I mean, is it still, the OEMs are still going to be important because they ultimately, depending on your business model, if you're going to sell the truck to the end customer, whether that customer is Amazon, Walmart, whoever, you want their maybe sales and distribution to sell it directly. So I think those partnerships are still important. Or with autonomy, does that change altogether? And can, you know, the technology companies effectively end up selling directly to, you know, the large buyers of these trucks?
They're still important, but I want to put this into context. Waymo does not have a quote-unquote OEM partnership with Jaguar Land Rover. They buy the vehicles and they are upfitted by Magna. Kodiak's partners, let's use Atlas Energy, public there, they buy the Paccar trucks, they're upfitted by a partner who we do not know. Same strategy that Waymo's taking. The days of the OEMs controlling all the power and being the make-or-break scenario, that's coming to an end. What Aurora has clearly said on the earnings call, what Mr. Burnett has said many times, that's the future. Having the OEM partnership is great, but it's no longer the kingmaker. Those days have sailed us by.
Yeah, they were maybe the kingmaker for a company that needed money, you know, or, you know, in Daimler's case, providing torque with the capital required, you know, to kind of get to the next step. Look, what the timeline is there though, they're still whatever, two years out. But I guess I kind of agree with you. Even if they were the kingmaker historically, at the end of the day, like it is a race at this point to see who can get to driver out first, then in tens and thousands. And if you're doing it by a kit one way or another, I think the OEM just doesn't play a material role in that. And frankly, if the tech, if someone gets in the lead in terms of the actual technology and it working, those guys are going to be coming to their doorsteps on hand and knee saying, you know, can we use the technology in our trucks? And that power dynamic I think will change very quickly, if it even existed in the first place.
It's clearly changing, and I want to highlight for our listeners and viewers a very big caveat that's not really discussed publicly, but it's all publicly available data. Historically, the OEMs have charged the autonomous driving developer for a quote-unquote partnership. The AV companies had to pay. That dynamic is about to get flipped on its head thanks to the acceleration of the kits.
All right, well, it should be an interesting time. The other, you know, part of that was Thor and just kind of as tech analysts like to say, double-clicking down into the role NVIDIA will play. I mean, Thor as compute is going to be critical to Nuro in terms of what they're doing. The Thor is like, talk about a supercomputer. I mean, I don't know, we don't have much information on exactly how much compute is in HW5 or where the newest Teslas are going to come out. Is this overkill? I mean, everything we've learned about is that things, models should get smaller and use less compute. And I think Jay-Chi at Nuro has the belief that these models are only going to get bigger, not smaller, in terms of in the world of autonomy. Where do you come out on this?
The models are going to become more specialized at the end of the day, and as the models evolve, the power requirements will drive down and be less power requirements. Does NVIDIA always have a monopoly for cars? No. I have a lot of really good feelings in doing research where the Qualcomm Snapdragon digital chassis could play a role. You could have somebody that could develop something on an ARM chip structure. So there's a lot of change coming to that market. I do not believe that the NVIDIA DLA monopoly will hold over the long term. We'll see. But in the meantime, NVIDIA I think probably wants to do more than just compute and add to the stack, go to modules, do sensor suites or whatever it is.
So if that's the case, if what you're suggesting is the case, you know, do we come back to a world where maybe NVIDIA does more than just invest in autonomy technology, but starts to own some of that as well?
My honest opinion, NVIDIA is going to have to go vertical at some point if they really want to conquer a big chunk of the autonomous driving market.
So if they go vertical, then does that negatively impact their ability to sell, you know, some of this product? I mean, I guess at one point Qualcomm made their own phones and then they sold their own chips and then people didn't want to use them because, you know, they had the, they were getting the licensing on the chips, they were selling the chips and then they were selling phones. So there has been places where you've been vertically integrated and still are able to sell the chips.
If your chips are still the best on the market, I believe they still will be able to sell them. Obviously, NVIDIA really has a clear desire for autonomous driving. They have a division inside of Automotive. They just invested in Wayve. They've made several other investments in there, which raises the question, if they do want to vertically integrate, does NVIDIA buy Nuro? Does NVIDIA buy Waymo and then license that out and leverage their chip structure to try and drive revenue there? Because let's be honest, NVIDIA needs a new source of revenue growth after data centers.
I know we have a lot of people that love Elon out there, and Grayson and I, I think, are indifferent. We're not lovers, we're not haters, but there is a lot of love for Jensen. So if they actually did buy Nuro, Waymo, whatever, they increased the vertical stack, I would liken that to like whenever Elon gets into a space, everything, every stock in that space and the ability to get financing from any competitor to Elon gets put immediately in doubt. So if, you know, if NVIDIA went that route, would that be a huge negative in terms of the ability of the remaining players to get financing, you know, to do the funding that's required for them to get their technology off the ground and on OEMs?
