Walt, this week it was GTC in the Valley. And what does Jensen do? He gives a speech, holds court, and you get the famous Nvidia bounce. While Jensen was busy holding court in the Valley, Uber had some news of their own. They got a new robo taxi partner. And our good friends over at Waymo, they released some interesting safety data, which we'll get into a little bit later. But let's start with the news of the week, GTC. What did you uncover?
Well, Grayson, I think I deserve a clap for this one because I think we talked about this on last week's podcast, previewing GTC and expecting autonomy announcements specifically with Uber, and that that would drive Uber stock. And sure enough, it did. We'll get to that later, but before we even go into the Nvidia news, let's try and dissect the alphabet soup of terms that they're using. You've got Thor, they're talking about compute. Hyperion, which I thought was the sensor stack, but maybe it means the full driver. Alpamayo, which I thought was training, but again Jensen seems to be using it interchangeably with now, you know, the kind of AV stack. You got full stack, you got half stack, you got open stack. I'm a bit lost in all of the words. And I think even Jensen when he was, you know, on the All-In podcast was getting tripped up on describing some of this stuff.
Well, first and foremost, happy clap for you. So, there you go. You got the clap, not the open claw. Maybe at some point we'll get the open claw to happen. Right now, we have the happy cat clap. Well, really what we need, and don't laugh, it's almost that you need a guide to Nvidia. Hyperion means this. Alpamayo means this. And basically, do we need the Nvidia chart to truly understand what all these things mean? Because it seems to me from reading the press releases and going through the technical blogs that there is a lot of technical overlap between all these various different services and functions and technologies that Nvidia is offering to the market.
I agree it's confusing and I swear that they've described some of these things differently. I don't have anything to back that up at the moment. Bottom line is I think they want to be full stack, right? And you know, as a result, that really dates back to CES, like Uber first moved on them talking about effectively being, and when I say full stack, meaning competing with other autonomy providers, whether it's Waymo or Nuro or WeRide, maybe more specifically than Nuro and WeRide. Those full stack companies are trying to sign up OEMs just like I think Nvidia is trying to sign. And by the way, for longtime listeners of this podcast, like we didn't, you know, we talked about Nvidia wanting to be full stack all, you know, well before CES. In fact, you know, I think we speculated on maybe that they want to jump start into this area because of doing acquisitions. And I think, you know, you put the Grayson inspector hat on and, you know, you I think identified a document that showed that they actually looked at purchasing one of these companies.
Yes, there was a court document from the Delaware Chancery Court where Nvidia technically didn't look at purchasing. They looked at entering into a licensing deal with Zoox for a $100 million licensing deal where Nvidia would inject $100 million into Zoox in exchange for the exclusive rights to license the technology. However, in the court documents, there was a very clear caveat: not acquire, pure licensing injection. This is the WeRide versus Zoox lawsuit which is still currently going on. This is prior to the Amazon acquisition in 2020. But to me, interesting sign that they wanted to exclusively license that technology. And in the court documents, it says that, and quote, Nvidia wanted to use this as a showcase for their technology. So to me, based on these court filings, based on what we've seen in GTC in the Valley, what we saw in GTC and DC, Nvidia has very big ambitions. But how big are those ambitions? Because on the All-In podcast, Jensen said they don't want to build the car. They want to enable it. And let's roll that clip and get your commentary on it.
We believe that everything that moves will be autonomous completely or partly someday. Number one. Number two, we don't want to build self-driving cars, but we want to enable every car company in the world to build self-driving cars. And so, we built all three computers: the training computer, the simulation computer, the evaluation computer, as well as the car computer. We developed the world's safest driving operating system. We also created the world's first reasoning autonomous vehicle so that it could decompose complicated scenarios into simpler scenarios that it knows how to navigate through, just like us reasoning systems. And so that reasoning system called Alpamayo has enabled us to achieve incredible results. We open this. We vertical optimize. We horizontally innovate and we let everybody decide. Do you want to buy one computer from us? In the case of Elon and Tesla, they buy our training computers. Do they want to buy our training computer and our simulation computers, or do you want to let us work with us to do all three and even put the car computer in your car? So, our attitude is we want to solve the problem. We're not the solution provider and we're delighted however you work with us.
So the words that kind of pop out to me is we are not the solutions provider. So I don't know what he defines as solution provider. Is he talking about Uber? Is he talking about the OEM? Is he talking about like Nuro as the solutions provider? So like that answer to the question, and I think Jason did a great job setting him up saying like hey, and this was by the way the thematic, the narrative that came out of CES which was like you know Nvidia can be the Android, you know, in competition with iOS, which is iOS would be Waymo or Tesla. So if they're the Android, then they're like the third player and all these other guys, I guess, go away in that theory. And like I don't think Jensen really took the bait on that based on that clip that we just played.
It's interesting. So I don't mean to go back to Encyclopedia Britannica. I don't mean to get into the weeds, but how do you define the solutions provider when you're providing the compute, you're providing some of the self-driving technology, and yet you state that you want to enable self-driving, you don't want to own the vehicle? I just think that I don't want to sound like the AP style guy, but we need references for what all these terms mean.
Look, the bottom line is let's just put all the words aside. Let's go back to what we've said on this podcast for again well before CES. We think they want to be the full stack provider. We think they want to sell to OEMs. They're obviously showing videos with Mercedes and talking about other OEMs. You don't see, you don't hear any mention, you know, of Nuro or WeRide, which are partners of Nvidia, in that those companies use their compute and none of this other stuff. None of the sensor stacks or the drivers. So, let's just put all that aside even though there's, you know, kind of a jumble of terms and I think, you know, you can again, you know, decide for yourself after listening to that Jensen quote what you think all that meant. We're going to assume that they are going forward and going to try and sign up OEMs. And I think the implication then for these other companies, these many many companies that Uber has partnered with, is it has to make it more challenging for them to raise additional capital. So I think to a certain extent, you know, while Uber wants this fragmentation by endorsing, you know, and kind of pushing to the forefront Nvidia, it's going to make it harder for competition to drive, you know, competitive innovation across these multiple players just because it's going to challenge financing.
It is. And if you look at the numbers, the Nvidia chips are very, very expensive. And the question is at some point another chip provider could compete with them. We're seeing moves out of Qualcomm. We could see custom fabs, what we're seeing with Tesla and Rivian. So, there's competition on the horizon.
Let's move on to where we think that Nvidia is compared to, you know, Nuro and WeRide and some of the others. We had a video as evidence, Jensen and or Nvidia put it up there with Jensen in a Mercedes with another Nvidia employee and you know, obviously CEO of the company and he's like very excited. He's like, 'Oh, the car made a, you know, went into the left lane and went into the right lane.' I'm like, 'Okay, I've done that for months, years, whatever it is on my Tesla FSD and obviously the many other cars that are out there.' And it's like, I don't know why this was so exciting. This looked like just a level two implementation in that Mercedes.
It clearly looked like a level two. I would compare it to GM Super Cruise, Ford's Blue Cruise, Tesla's original autopilot. The technology seemed very clear there what it looked like to me and I'm not an engineer but it looked very clear like it was an L2 system. And on the All-In podcast, Jensen made a very interesting remark when Jason, who asked a really great question, asked Jensen about what is the role of the driver? Where do drivers go? And then Jensen offered this remark of how the car will drive itself. It'll be an L4 car in that case and the chauffeur will sit behind the wheel. But I put on my regulatory hat. You're opening up a massive liability if that is indeed the case. But if you're going to operate an L2 service, that's 100% correct. If you're going to operate as an L4 service, the liability, the benefits of L4 go away because legally, the way that the laws are written today, the individual sitting behind the wheel is legally responsible for all interactions. When you're at L4 legally and there is no steering wheel, no pedals, we don't know who is responsible, but we do know right now as the way that the laws are written, it will not be the passenger or the individuals riding in those vehicles. So again, it goes back to I'm going to call it the great Nvidia chart of terms.
Yeah. And look, and the bottom line is the business case here is to not have anyone, you know, behind the wheel in the passenger seat. I mean, obviously, we're going through the same issue with Tesla and their launch. You know, they have a driver now in the passenger seat and obviously we took rides where there's no one in the car. So, these are all just kind of steps along the way and there the process is again them being at level two similar to I think WeRide in some of their OEM relationships, but it wasn't, you know, one of the big companies that was going after the OEMs again with level two. And yet Mobileye, I don't think, was down on this announcement. And it would seem to me like they would get more impacted than anyone if you're effectively having someone like Nvidia coming in and competing to deliver level two service at least initially, you know, with a path to level four, which again is the Mobileye game plan as well. It is, I think this is my perspective, that there's a lot of disconnect in the market between reality of the technology, but Nvidia, if you go on their website you can see all the logos or if you watch a Jensen presentation, I don't know the exact number, but they have a significant OEM relationship and that relationship, that Nvidia technology could easily be integrated on the seven-year cycles for L2 in theory, this is in theory, which would replace Mobileye if that's the route that they chose to go.
So, look, at the end of the day, the keynote, the video, that's not the thing that was really that interesting. I mean, the new news was the announcement with Uber. I mean, they had a ton of announcements as Nvidia always does, but the Uber announcement which pushed the stock up 5% plus. I think Lyft was also up a similar amount. We'll get to them later. And I think let's break this into two parts. You have first you have this milestone, the first milestone which they're saying by the first half of 2027. So that's about a year away from now, they're going to launch in San Fran and LA. So this is important right because this is the near-term timeline. This is, you know, effectively Nvidia with whatever OEM not announced demonstrating that they can deliver what we believe to be the full stack. This is on Waymo's home turf. So, we got signal on, you know, what the relationship is like between Waymo and Uber. So like, you know, this was the thing. They're laying down the gauntlet, which I thought made it feel more real. I mean, who do you think the OEM is going to be? And what did you make of this first milestone?
I mean, they give who I think it is, and this is based on a press release. It will most likely be a Stellantis brand based on the press release. However, I will caveat and also say there was a company mentioned in the press release, Foxconn. Foxconn unsuccessfully did not acquire the Nissan plant in Japan, but Foxconn has the engineering resources to do it. So perhaps it could be Foxconn, but then you're going to have to go through the FMVSS and all the regulatory stuff to do that. So I'm going to give you a copout and say it's most likely a Stellantis brand. Which brand in Stellantis? I don't know, but most likely a Stellantis brand. I mean, I didn't see anything in the press release that suggested that. So, I'd have to go back and take another look. I mean, the fact that Jensen is in a Mercedes and that seems to be their high-profile, you know, car, my guess is it's probably going to be that. And at the end of the day, I'm not even sure it matters because when you say launch, you know, I put my inspector hat on to try and determine like what exactly does quote unquote a launch mean? And in the press release, I did see this in the press release, it said, quote, 'An operator-led launch before transitioning to fully driverless level four deployment.' So this first half of 2027, to be clear, is not going to be a Waymo type thing where you've got cars, you know, driving around with no one in there. To me, that means you have a safety driver and that's, you know, by definition not level four, whatever. Maybe it's a level four in terms of the technology and it's just like hey the safety guy's there just to watch, Aurora style, but like you know there will be a safety driver there and you know so I checked with Uber. Uber's definition of the launch is it's commercially available on the app which again confirms kind of my view why because if you look at AV ride in Dallas or what May Mobility is going to be doing in Arlington, they will have safety drivers and it will be available commercially on Uber's app. So I think that is what we should expect in the first half of 2027. What are your beliefs?
It will be a safety driver. Same thing that we saw in a press release a few weeks ago that the Lucid launch with Nuro and Uber will be safety drivers. But the thing to watch in all of this removing the safety drivers is the ODD. How large is the service area? Is it running in complex urban areas? Is it running in suburbia areas? Are Uber and their partners going to have to have their Chandler moment? And when I say their Chandler moment, Waymo very famously launched when they were the Google self-driving car project in rural Arizona because the streets were wide, the infrastructure was good to learn the technology and that did very well for Waymo. So I think that's something that we're going to have to watch, the service area and where inside of that service area or the geographical region that this is launched.
I got to give you credit because you've always been very consistent on pointing this out and that's something I left out in the note that I wrote for our investors earlier in the week. The ODD, and you know you talked about this in Austin and have been consistently right and yes that's a big part of it meaning that like what is the broad investor viewpoint of when Nvidia says SF and LA they're thinking probably what Waymo is delivering today and the reality is like if it's you know maybe all of San Francisco maybe it's even less than that. Remember when Amazon launched Zoox it was like in a very small even subset of San Francisco so that will also be a part of it. Technically is it a launch? Yeah it'll be commercially available on the app but like where's it going to actually go. So I think that's a great call out and like again if you think about how investors react to this stock and the thing I don't think they have that understanding. So when people end up going in the first half of 27 and checking this out is their view going to be as positive in terms of how that positions Uber and Nvidia, you know, based on what they're delivering there and then more importantly like shouldn't we get data points prior to that? Every other autonomy stack company we talked to even prior to whatever the launch is, we or someone has been able to get into cars and the only evidence that we have of anyone getting into an Nvidia full stack car I guess if they call it that is this Mercedes and all of that feedback although generally positive has pointed out the interventions and the fact that it's really just a level two type of experience.
I was seven, eight years ago I was in one of the original Nvidia Drive vehicles. We did it in a parking lot. This is before we went to open roads and it was marketed and built for an L4 system, but in all reality it was operated in a level two system in a controlled environment. I have not been in the latest version of the vehicle. However, we do have multiple reports in YouTube videos and I believe it might, and talking about the interventions, and if I'm an investor or I'm an analyst, if I go out to San Francisco to experience this technology, I experience the Nvidia technology. Okay. And then immediately I order the Waymo. There's no driver in there. I mean, I'm just thinking about this. I don't run money, but that's like, wait a second. What am I missing here?
No, that's my point. And I think hopefully we'll get data points prior to that other than, you know, the hand selected media people that they've done. And I'm not, you know, there's nothing wrong with that. Like I think those people did a great job at reviewing their experience in the Mercedes. Obviously, they were doing test rides at CES. And look, you know, I'll throw it out there. I know we have got a good set of listeners. I'll fly anywhere to get into one of those cars or anyone else that has an autonomy stack that they want me to try. Get on a plane no matter how long the TSA lines are right now to try it out. But again, those data points, I don't know, relative to what people are going to expect from Nvidia, there may be a disconnect there.
I think there's a disconnect. But the question to you is when the disconnect becomes a reality, how does the market react? They say, 'Oh, that's less than 2% of the revenue. It's okay. They still sell GPUs and data centers.' Or how does the market react? Well, this podcast there is no investment advice on it by definition, but I think over the course of the year there's going to be a lot of puts and takes in terms of, you know, who's in the lead or not. And I think look, you know, I would expect Nvidia to provide more OEMs in terms of the partnerships. I don't know what the various, you know, connection will be with those OEMs. Waymo will do the same. WeRide and Nuro will likely do the same. Find additional OEMs, expand existing relationships. So, there's going to be that kind of back and forth. In the meantime, you know, over at Waymo, there's going to be cranking out more and more cars, and Tesla, you know, hopefully will finally lock down the software to start to ramp up their own volume of, you know, unsupervised cyber cabs and robo taxis. So, I think there's just a lot of noise that we will see over the course of the year, but again, it's just where expectations are today.
I'll tell you what you and I saw at Giga was absolutely impressive. And I have a whole new found respect for manufacturing. And that makes me put on my inspector hat where I'm watching press releases, groundbreakings for factories. Does an OEM spin up a robo taxi line for contract manufacturing? I think that's a sign that we need to watch over the next 12 to 18 months because that'll be a really good indicator of how fast the technology is going to scale.
So speaking of the long term, you know, there was another phase of the Uber press release with Nvidia that talked about 28 markets by 28. Now look, why 28 markets? That sounds very specific. It's pure marketing, right? If next year happened to be 2030, Uber and Nvidia would have said 30 markets by 2030, right? So that is the number. Doesn't mean they can't get there. That's not the target. But like there are, from what I can tell, and I've done the research on this, there's no specific OEMs tied to those 28 markets. There's no specific commitments. This is not necessarily a real road map. They're just saying like, look, it's Nvidia. We're comfortable with where their technology development is. We think they're going to hit this in the first half of 27 as we've talked about. They're going to sign more OEM relationships. Getting to 28 markets, you know, seems likely. I don't know which OEM, like they've listed Hyundai, Kia, Stellantis, which you talked about earlier, BYD, Polestar, JLR. I mean, there's a list of OEMs. I don't know how many of them are going to actually go there. But again, now we're getting caught in this trap, which I've talked about with the truckers, of working with OEMs and expecting OEMs to move quickly to deliver products to the market. And we've seen the challenges that Waymo has had on that front.
