Walt, welcome back from your winter break. I hope you had a nice relaxing time.
I did the one thing I never did. I relaxed. And we're back here on Autonomy Markets.
And to the listeners and viewers, Merry Christmas and Happy New Year. And thanks for all the great feedback you gave Walt and I last year. Look forward to producing more great episodes with you this year. Next week, you and I will be at CES. If you're there, come say hi to Walt and I. I'll be the colorful guy. Walt will be the tall guy. It's not twins with Danny DeVito and Arnold Schwarzenegger, but it will be fun. But well, let's start with the big news. When you were relaxing, the lights went out in San Francisco and Waymo, no pun intended, came to a grinding halt at intersections on the roads. What do you make of all that?
First of all, there's obviously no surprise that the lights go out. The infrastructure, the power infrastructure in our country, regardless of what the city is, you know, is challenged. Obviously, this is the exact thing that we talk about in terms of Waymo and Tesla and others scaling is the challenge of getting more power to charge all these EVs that are out there. That wasn't the surprise. The surprise was that these hundreds of Waymo cars in that market just basically stopped in place and turned their emergency flashers on. Very surprising.
And what was more surprising to me is, because things happen as we know as this product is moving forward, is like Waymo is just silent for multiple days. And I think it again goes back to something we've talked about in this podcast. You run by engineers, right? It's very tech focused. Like, hey, we just make a good product, everything else will fall into place. And the PR is just very lacking. It took them multiple days to come back with some explanation, which was like a half-assed explanation, to be honest, in terms of what happened. They need to do a better job at, you know, getting in front of this thing as this product becomes more mainstream. So I think it's just a really disappointing response from what was a pretty big event. We can get into, you know, what exactly happened.
It's crisis management 101. The vehicles, no pun intended again, they played possum and they just sat there. I did an interview with Axios about this and the key takeaway from the interview we did with Axios was public trust, no response. You're eroding public trust at a time when public trust is absolutely crucial to scaling autonomous vehicles. And when an official statement came out, and I'll be rude in PR speak, me reading between the lines, while it looked a lot like they were blaming the cellular networks. You're Mr. Telco on LightShed. You do telco corner. You've got a lot of telco relationships. What really happened? Was it truly cellular going down or what was it?
Yeah, I mean, that's not the way I read their explanation, so I don't want to put this on Waymo, but I did call the operators and I'm obviously familiar with how the network works. There's power backup on these cell sites, whether it's battery or generators. So there was connectivity there. I think the issue when you read that, you know, again, it was a very, I thought, weakly worded statement. They just, you know, just be direct, tell people what happened. And I think when I read it, it sounded like, you know, you have these remote drivers, which we've talked about before in the podcast, where as the Waymo is driving and it's about to hit something that it doesn't quite understand, it goes to an external person that's sitting in front of a computer and that person can click do this or do that. Well, they had such an overwhelming flow of requests back to these remote drivers. It was understaffed. That seemed to be the issue. So the question is, like, this is also something they don't disclose. We don't know how many remote drivers there are per car and maybe they're going to start getting asked questions for greater transparency. Do you need to allocate more humans as remote drivers that are there?
The bigger question though, the bigger issue, which is, you know, because I think you and I would both agree that's not the solution, is how are the cars not programmed, even if there was no connectivity via the cellular networks, because that could certainly be a challenge that happens in the future. If there's no connectivity, how are the cars programmed not to just pull off to the side of the road in a safe location so you're not blocking an intersection where emergency vehicles might need to get through, you know, during a power outage? I don't even understand. And why was there not a better explanation of saying like this is why that didn't happen and this is why it won't happen in the future? It was very generic wording. Just again a really poor job by the company at handling, you know, what was a pretty big PR event.
PR is not very good. Let's look at the Toyota statement, one sentence. The one thing Waymo PR is good though on announcing new cities. So I don't get this. They should be transparent with us to build trust. What is the current ratio of remote drivers to vehicles? What is the current intervention with remote drivers? Disclose that data, build trust, make it part of the Waymo safety benchmark data set, release it publicly, open source it on GitHub, and allow engineers to build on it. Because this whole incident and the response, Walt, unfortunately, it raises more questions about how technologically far Waymo is ahead as it compares to competitors.
Actually, I don't even think that's the issue. I think the product is great. You know, these things are going to happen. Just have better responses for it and, you know, and maybe be more transparent about what's going on because you're going to run into cities, which you already have, Boston, Seattle, whatever, where they're going to try and keep the product out and you're not helping yourself by how you're dealing with these situations. Be honest about what's going on. I'm sure the issue is not anything that's not solvable, but be more specific and open in talking about it.
The other, you know, the interesting thing about this is it was only a few weeks ago when again this debate surfaced about whether Tesla needs lidar or doesn't need lidar. And I forget who said it, but someone's like, 'Well, what happens when the lights go out with the Tesla?' And so the ultimate karma was this. The lights went out and the Teslas still operated, right? If you had your FSD. And by the way, that was always a ridiculous thing for someone to argue as it relates to lidar. I get that there's issues that people believe that lidar is a required thing. You know, maybe that is the case, but it's not because of the lights going out. The Tesla has headlights, right? They're powered in the headlights. That's when the lights go out, the headlights continue to work. I drive on highways where there's no street lamps. It works, quote unquote, works in the dark. But I think that was the ultimate karma that when the lights went out in San Francisco, it was the Waymos that stopped and not the Teslas.
Bingo. And there's individuals say, 'Well, this and this and this and that.' Well, I'll give you two real life examples. When I was running 13.2.9 and an intersection in town, the lights were completely out. They were dead. Very similar to San Francisco. The Tesla vehicle 13.2.9 treated it as a stop sign. Came to a complete stop and safely went on its way. Didn't run through it. And now I'm running 14.2.2.2. Yes, there's 32 twos there. And I went through an intersection the other day where a blinking red handles it perfectly. And I live in a situation where there's a lot of Secret Service activity. And I want to point out for the listeners that Tesla with 14.2.2.2 handles Secret Service checkpoints and roadblocks no problem. It's really truly amazing to the point where I'm going to say, Walt, very publicly here, it's just a matter of time until they crack full self-driving and they validate that you don't need lidar.
Well, that was a lot of numbers, Grayson. And again, you know, we'll see if they ultimately don't need lidar. And like I think, you know, the jury's still out until maybe we're three, four years down the road here and we can really have enough statistics to see which of these solutions is performing better from a safety standpoint. So I think, you know, I understand the argument that people make that use lidar in their stacks. So I think it's going to take some time for us to truly understand, you know, whether that makes a material difference. Clearly it's not like the product is not functional. You and I both use this product. I don't know how many people, hundreds, let's call it thousands of people use this product every day. It's, you know, I get enough situations where, you know, the cameras have worked functionally well. It's not like the product doesn't work. Now, does that mean it's going to be as safe as something with lidar? I don't know. You know, time will tell on that and hopefully we get more stats.
Let's go back to the remote operator thing because again, I'm not sure many people understand this. But again for any autonomy company, like if there's an issue that the car faces, like you know, you think back to the times when I think it was the Waymo going in the circle, like that car will ping back to a human who will tell the car what to do. They won't drive the car. Maybe in the case of Tesla that's what happens right now, but that's not the endgame in terms of what these remote drivers do. There is a ratio, like okay, for X number of Waymo cars there are certain number of remote drivers. What's the number? Like we have Chinese companies that we deal with that give us better disclosure in terms of what that ratio is of remote drivers to the cars that are out there. So here you had a specific example where they clearly didn't have enough remote safety drivers available in this scenario as they wrote in that letter. So what is the ratio and what are your goals? And like these are the types of things that I think if you want to build, getting back to your point of public trust, that there needs to be more disclosure when you have situations like this. That it's just that simple.
And here's the other thing. This goes back years ago when I met with individuals from Homeland Security in DC and there was a remote operations company, for privacy purposes I'll admit them, and this remote operations company who claimed they were the first to go drive route out in California was operating in a garage with the door open and Homeland said, well, something can go in there with a gun and remotely take control of the vehicles. So security issues when you have remote operations. Where are the remote operators domiciled? Are they in Mountain View? Are they in the United States? Are they overseas? To me it raises a lot of national security issues, especially what's happening in the world today with geopolitics. We need clarity on where these individuals are located. How are they trained? What is the security of those buildings? What are the restrictions around it? I remember years ago Dr. Mark Rosekind, the former NHTSA administrator, hosted me at Zoox headquarters and he showed me how they built the remote operation for safety. So somebody couldn't go rogue and drive it on a sidewalk or hit something. All the safety parameters have to become more transparent there.
Just to be clear, like the remote drivers when that is triggered for Waymo is not a person with a steering wheel, right? They're getting an option that says like, 'Okay, this car is in a situation it can't handle. Pick A or B.' And that the remote person is literally just clicking a box saying do this or do that, right? And you know, so they're not like taking control of the car. The car is always the thing that's driving in the case of Waymo. I don't know if that's the case with Tesla. It's possible in Tesla right now if they get stuck there is a person with a steering wheel. That is not going to be the endgame for Tesla as well. That's the way probably Waymo started in the earlier days. At some point, you evolve in terms of what that remote driver does.
The bigger question here though, Grayson, is, you know, again, scale. Like there are 2,500 cars. That's not scale in the context of the number of cars that Uber has, but as you get more and more cars, like these are the things that happen. Like these remote cases are not the things that necessarily are going to cause the car to crash, but it's like, are you ready to take that up to 700, 800 cars? Like if there's an expansion area of remote drivers that'll happen when you have 10,000, a million cars on the road, like what are those scale implications? So scale is not just about opening markets fast, which is what everyone wanted to define scale for in 2025. It might not even be about getting enough cars. It's can you service the hundreds of thousands of cars and deal with these situations, you know, when they occur. This is all a part of scaling a product, which is obviously important. This is nuts and bolts. These are different skill sets than someone figuring out how to get that car to operate autonomously, which is the engineers that have done such a great job at Waymo up to this point. And this raises the question which I've raised in several issues of this week in the Autonomy Economy newsletter is why does Waymo not set up Waymo Operations LLC, bring the operations in-house because you can iterate faster and learn from this and bring in experienced operational people.
Yeah, I don't know. Again, this is like, you know, Uber's got the same thing. You don't want to necessarily own, you know, in this situation. Waymo doesn't want to own. They want to partner. But the problem is when you don't own and you're partnering all the time, you're still, it's your brand that's out there, right? It's Waymo's cars that are parking in public curb spaces in between trips or like aggregating in someone's cul-de-sac. All the consumers going to see is that Waymo brand because whoever your partner was to service those cars hasn't been able to handle them and sent them somewhere else. I mean, these are real issues that come with, you know, what is true scaling. Again, we had this debate, you know, a month or two ago, like where everyone's, you know, I got that argument with that guy on Twitter about like, oh, scaling is the fact that I can open markets faster. No, it's not. Scaling is about can you operate hundreds if not thousands of cars in a market efficiently and that's not going to upset the consumers and the politicians that are in that market. These are key issues that I think they have to deal with. So whether it's the PR on this or the execution of this, like these are real issues that Waymo is going to have to deal with as they continue to ramp cars. And like, look, we're getting every day we're getting closer and closer to when they, you know, with Hyundai and perhaps other OEMs get to volume production and they're going to, you're not going to see like hundreds of incremental cars hitting the market every month. It's going to be thousands or tens of thousands. And like, you know, you got to be prepared for that.
Well, let me ask you this question and this is all publicly documented for the record. Who operates the remote drivers? Is it Waymo? Is it Moovit? Or is it an unknown partner? I think that's something we'd like to get some clarity on.
Yeah, I have no idea. There's no transparency. I would hope that these are Waymo employees, right? But I have no idea. I don't really think, you know, any of those companies, I think those companies that are doing the servicing are just charging, cleaning, maintaining. I don't think they're the ones, you know, doing the remote operation in these situations. But I don't know.
All right, Grayson, let's shift topics off of Waymo and Tesla. And, you know, it is now 2026. Like I said, Happy New Year. Bottom line is they didn't really execute on their 2025 predictions, right? It was close. Sure. Like there was a picture of a Tesla in Austin with no one in the driver or the passenger seat, right? But that is definitely not what anyone expected in terms of, well, anyone that believed that they were going to deliver on these results. Certainly there's a lot of negative people out there that didn't think that they would do it, but you know where they would be at the end of 25 in terms of cars with no safety attendant in the passenger seat. So, how are you feeling about 2026 in terms of executing on that and, you know, increasing the number of cars in Austin?
I'm sticking with my 2026 prediction and I still believe, and I'm sticking with this, that the driver-out, safety-out will be in a limited ODD to start. So, I'm sticking with the prediction and anybody who follows Tesla knows they don't necessarily keep the timelines, but at the end of the day, they do truly execute.
I mean, I guess, yeah, it just, whatever. This one seemed like it was more immediate. It's not these like, hey, these large grandiose predictions of a million robo-taxis the following year, which were never really, at least for me, fully believable. This one was like more near-term. Like you saw the ramp happening. You saw the markets. We're seeing what the software updates were like. Certainly, you know, the more recent Tesla updates for our own FSD have been much better than when we were going through that 13 phase where it was very, I thought, very, very choppy. But I don't know, this is like, you know, they were so close. You know, I guess you could argue culture of safety of kind of not moving it forward. Like what changes in a week? Like I don't know. When do you think, how much more time do you think it's going to take for them, even if you believe that by middle of the year they scale to a couple hundred in Austin, like how many more weeks until we even get like 10 cars in Austin taking rides with no drivers out? I don't even think, I mean, I haven't checked this, but have you seen any YouTubers out there posting videos? Like how many rides are they even doing right now in Austin?
We don't know. We don't have data on it. We have not seen a YouTube video of a Tesla influencer, nor have we seen a post on X or a mention on Reddit. We did have a video where a cat went in it, but a cat's not necessarily a human. I'll give you a prediction. Commercial service, no safety attendant, no safety driver, complete full driverless, end of February, early March, and I will say sub-10 vehicles in a smaller ODD than the vastly available robo-taxi ODD today.
I mean, that's not good. In my view, that's not good. We went from thinking that was going to happen towards the end of the year to now you're saying like end of Q1, pushing it back that far. I don't know. Again, not a company that you've got a lot of information. You know, we just kind of ripped on Waymo for PR, but like, you know, maybe someone from the company should provide more updates like, hey, it's end of the year. We acknowledge that we're not there. Here's kind of the latest update. I guess we can ask that on the upcoming earnings call and see what the update is there. But, you know, certainly, and I think there were supposed to be multiple markets as well. So let's push this out to end of the year. Let's forget about being too precise in terms of weeks and months. End of 2026, you know, we already know what you think about Austin. Like, how many markets do you think have, let's call it, you know, 50 cars operating, taking paying customers by the end of 26?
With safety attendant or no safety attendant?
I'm going to tell you a number that's going to upset our Tesla bulls. Two to three.
I mean, that's pretty bad though, right? I mean, if you believe that they can get Austin there, right? We know that this is like a true generalizable system. Why should it only be two or three markets? Like something doesn't fit there in terms of like how special. And look, I acknowledge like I think there is some specialization that should be required for a market. But when I drive from New York to Philly, right, or you know, there was this video or at least a claim someone went, which I'm a little skeptical about, coast to coast. This guy David Moss, who's ironically a lidar salesman and claims have no interventions. Like if that's true, then like two, three markets, like that makes no sense, right? Because there's not a supply issue. There shouldn't be a local market issue. That makes no sense. So Tesla themselves, I think, is going to have to provide, and which they do, you know, at most, I guess, on the earnings calls in terms of what's going on there.
By the way, on the coast-to-coast thing, like in my experience in yesterday, New York to Philly and back, largely was great. My interventions were, I was coming up to a toll booth and it was going to go in the cash only and I'm not going to pay cash. And I think in Philly, like it's truly cash only. It's not like you can go through E-ZPass wherever you go. So I had to intervene. There was a situation where I was coming off the GW Bridge and it was about to miss. It gets very confusing in terms of the lines. It was about to miss the lines. It was going to send me south as opposed to north, which would have been problematic. And then also what I, when you know, I had this trip, but it had a scheduling a recharge. It took me to a off the Jersey Turnpike, a Jersey recharge station and went in and it started to park where there was 25 chargers there, maybe 15. I forget. It was a lot. Plenty of open spots. Drove right by them and then tried to back in where there was literally no charger. So, like, how is it that this person went coast to coast and didn't have one random one-off thing? Like, again, those are not like deal breakers. Like, these are small things that are clearly going to get fixed, but like how did that person really go coast to coast where they had to have multiple charges and not one time have a situation like that? I'm a little, no offense to this person. I know the data has been analyzed, I suppose, but like I'm just going to say I'm skeptical given my own experience, which by the way has generally been positive, but those are the interventions that I've had.
In Jersey, are you able to charge your own car? Because you can't pump your own gas. Can you charge your own car?
You can in fact charge your car. And by the way, the charging was super fast. It was great, right? You know, in terms of the powering. But like getting back to the point, like I don't know, there's still stuff to work out. I think you're setting a low bar for them. It doesn't logically make sense that it should only be a couple of markets because what you're really saying, what you've described, should have, if I had asked you what you described for the end of 2026, if I had asked you this in February of 25, you would have said that that would have been the case at the end of 25, I think more or less, maybe give or take a quarter. So, you're basically saying it's another year of delay in terms of getting the tech healthy enough to move forward.
It's not just the tech. You're harping on the tech. And to really go back to Mr. Moss' drive, it'd be great if we had time-lapse video like Whole Mars Blog has released the whole video, but we don't have it. So, there's really no way to truly prove and verify that. Yes, we have data, but data could be fudged. We all know, I'm not saying that happened, but we'd really truly need more details and more data.
It's not just technology, Walt, it's also the regulatory environment. And last night down the road at Mar-a-Lago, Elon had dinner with Trump. And when I speak to individuals in the policy world, they say certain individuals and certain organizations have gotten it all wrong. They said, 'Oh, they're pushing for an act of Congress, an act of Congress to get this through.' That right now there is really no path to pass a national autonomous vehicle framework through Congress. It's gridlocked. However, if the industry was smart enough, and maybe it's perhaps what Mr. Musk had dinner with President Trump last night, an executive order. Look what the president did with an AI executive order. Why don't we have an autonomous vehicle executive order? Mr. Musk has the political clout to get that through. And if there is an executive order legalizing autonomous vehicles in all 50 states, and there's a little thing, I'm sorry, state folks, preemption that would allow that to go, then I might raise the bar six to seven.
That's a cop-out because you know you said your prediction was specifically two to three markets, right? Or whatever it was. It was not lame in terms of your, lame, but it's a lame development. I think I would characterize that for 2025-26, excuse me. The regulatory environment does not limit you to three markets. It certainly hasn't limited Waymo to three markets. There are plenty of cities out there that you do not need national preemption of any sort for Waymo to hit, you know, more markets than I think what you're predicting. So I don't, that's a cop-out. I think it is a technology issue and again to give them credit, focus on safety. We can feel the progress in driving it every day, but let's be honest, like it's, you know, we would have hoped and liked and expected them to be further along at the end of 2026 than what you're laying out right now.
