What Elon Musk and Ashok Elluswamy Actually Said

So, most people watched Tesla's earnings call. They heard the headline numbers and moved on. But when you take a closer look at what Ashok and Elon actually said line by line, it tells you way more than you realize. Today, we're going to walk you through three deep dives that do exactly that.

The first breaks down Tesla's vice president of AI, Ashokwami's comments, and what you'll see is he doesn't view robo taxi as an autonomy problem anymore. It's a scaling problem. 380,000 unsupervised miles, zero notable incidents, and those miles are compounding double digits every single week.

The second pulls 13 takeaways from Elon's comments. And the big nugget in his mind, demand is no longer the constraint. Growth now comes down to how fast Tesla can build cars, robots, chips, and electricity.

And the third asked the critical question, is Tesla being too cautious with robo taxi? The answer might surprise you. All three come from CERN Basher, one of the best analysts in the Tesla community. CERN is a charter financial analyst running his own investment advisory firm called Brilliant Device providing wealth management services. Welcome CERN.

>> Hi Herbert, it's good to be here.

>> First of all, thank you very much. Uh the big debate amongst uh Tesla investors right now is they're trying to understand what exactly was said during call and two camps immediately formed mostly on the negative side. They're hearing things like uh robo taxi won't roll out much at all this year and uh and things are much more delayed and there's there's some truth to that. But I appreciate what you've done here is you dabbed three deep dive articles where you parsed together exactly what was said and you want to go walk us through it and let's all together try to decipher and I think you found some key nuggets right?

>> Well, I think so. I think one of the things that happens uh in these earnings calls, I feel that people maybe pick out a few words and then run with that. And what I like to do is is take as unemotional approach to it as I can and actually look at the comments and really try to dissect them. Um I find that that leads me to much better um observations than if I just take some of the headlines.

>> Exactly. I appreciate that. So, let's start off with Ashok's uh Q2 comments and then we'll go through your article on is Tesla being too cautious on robo taxi and then after that we're going to go through 13 takeaways from Elon's Q2 comments. So, let's go through Ashock's Q2 comments. Um, so four developments happening. The existing system is increasingly safe to operate without an onboard supervisor. Unsupervised mileage is compounding rapidly. The engineing work required for each new city is declining. Cyber cap production is being synchronized with expected expansion of autonomous mileage. You say taking together Tesla is describing the beginning of the transition from a small carefully managed pilot into geographically scalable transportation network. Let's start off with the first one.

Um so first Travis uh kind of as instead of ask answering the say questions he just combined them together and let Ashok come out and give his full review. Uh so he Ashok first says first of all I'd like to state that robo taxi program has been operating extremely well especially in terms of safety. The program has an impeccable safety record. We have driven more than 380,000 miles of unsupervised robo taxi now across six cities in two different states. We've had zero notable incidents. Any reports have been of other actors impacting us when we were stationary. I'd like to emphasize how safe the operation has been so far. Zero notable incidents. Historically, the so-called experts have always claimed you need lightars, radars, AG maps, entire kitchen site to drive safely. We've proven otherwise. Basically, it's what he's saying. What's your thoughts on that?

>> Yeah, he covers a number of I think important points here, right? Certainly um the comments about zero notable incidents I think is is very encouraging. Um now it's a small number of miles 380,000 miles is small but again you've got to start somewhere and they've started somewhere and it's it's growing quite rapidly. Zero notable incidents though I guess the question is you know what exactly does that mean? Now they did say that that that is when the car is stopped and some other vehicle or something has impacted the vehicle in some way. Um it would be nice over time if we got more sort of uh data on that in terms of you know is it zero traffic violations, is it zero remote interventions, is it zero aborted trips, uh zero passenger complaints, zero operational issues th those are all things I think that over the longer term would be good to know. Uh but for now I I take what he says at face value. It it seems like that's that's a good place to be. Now it's also very important to say that that perfection here is not the objective. You are if you're running a transportation service and we'll get to this when we talk about are they being too cautious that perfection is not the objective here. It is impossible to be perfect but for now at a small scale they want to minimize all the issues that they are seeing. So while it's near zero notable incidents maybe there are some operational issues that they've uncovered. Maybe there are some remote interventions that they have to deal with. Right? So they have to work through those issues first.

>> Yeah. Well, I you and I have done many shows. I've done a deep dive on every single incident that's been reported uh for Tesla and uh anyways, we we can refer to those shows. But yeah, I we're very clear what exactly happened for every one of those.

>> and there's minor things too, Herbert, in terms of when the car picks you up, does it pick you up where you're standing or does it make you run across the parking lot? Right. Those kinds of things are issues that would not be reported anywhere. Yeah.

>> So that may be more of the type of stuff that they're looking at right now.

>> Exactly. Not not in terms of safety incidents though. Okay. So then he continues um here we show that that such is not true. You can have safe, comfortable, affordable autonomy with just cameras. This record should be a huge validation of Tesla's entire AI approach. So when I heard that it's like people gloss over it. Yeah, it's a given. It wasn't a given just a year, two years ago. People did not really believe that just vision only is going to work. They're basically now saying, "Yeah, we've put this nail in the head." Um, hopefully nobody will start talking about liars anymore.

Second, I'd like to discuss scaling. We started the robo taxi program, bre a year ago. We had safety monitors. The end of last year, we had fully unsupervised robo taxis. Beginning of this year, we have grown at doubledigit growth rates in the number of unsupervised miles that the fleet drives every week. I repeat, we've grown at such a high compounding rate on a week-over-week basis over the last several months. Not only that, we expect to continue growing at such a large rate through the rest of this year. So his words were doubledigit growth last few months and then every week and we're going to see that the rest of the year. Elon said he thinks he goes I think it's at least over 10%. So we don't know what double digit growth means but um people then discounted this saying yeah small numbers of course you're going to see double digit growth and then double digit growth for the rest of the year per week. It sounds like a lot but it may not be. What's your evaluation of what he said here?

>> Well, there's a few things here I think that are important. So, back to the roll out in Austin. You and I were there that first day. Um, a lot of people are saying, "Well, they've they've been getting all those miles since then." Yes, they were supervised. And all those miles are not throwaway miles. That those vehicles were essentially unsupervised operation. The car was doing the driving. There was just somebody sitting there that could stop it if necessary. That's really all they could do. So, those miles are part of this as well. So, that's the first thing to point out. Um, the other thing is a lot of people, excuse me, a lot of people say that Elon's completely lost credibility in terms of his comments. So, here we now have Ashok putting his credibility on the line and setting expectations about the growth of unsupervised miles, right? And he's saying that the double-digit growth rates to the number of unsupervised miles the fleet drives every week. I repeat, he says he he wants to emphasize this. I repeat, we have grown at such a high compounding rate on a week-over-week basis over the last several months. And then he says, not only that, we expect to continue growing at such a large rate through the rest of this year. He's very clearly here telling us what his expectations are. Elon didn't say this. Ashoke said this. And a lot of Tesla investors now for years have been saying, "Okay, discount Elon. He's he's optimistic." But when Ashoke and Lars speak, we can rely on what these guys say. So if that is still true, this is a very strong statement.

>> Mhm.

>> Now, I wouldn't take this to mean that they're going to grow at 10% every week after week after week. It could be more one week. It could be less. And also, this is not a linear exponential. This could be, excuse me, this could be a double exponential where all of a sudden they get to a point where they have a certain level of confidence in what they're seeing and they add a lot more vehicles to the fleet and all of a sudden you've get you get a major expansion in the number of miles. So, this is actually an extremely encouraging comment from a shoke in my opinion. This is one of the strongest comments in the entire earnings call. Um so what's amazing about this is even at the lowest possible interpretation of what he said it results in enormous growth in terms of the number of unsupervised miles.

>> Yeah. I think the vast majority of people who heard that statement they go 10 first of all they locked into the 10% number that Elon said and then they just locked into this idea that it's 10% per week for the rest of the year and that's it. And if you calculate that 1,500 cars, two somebody did the calculation 2,000 cars maybe it's very small. That's the vast majority of people. But that's not what he said. So when you listen re read what he said it could be like you know this compact this growth could could be 10%. But it could also be like you said eventually even higher right. So

>> yeah. Okay. Uh the other thing then he continues and he says the other thing to note is that for expanding to new cities it has been relatively less effort on our front. We expect that the time to launch to a new city will continue to trend towards zero. Wow. Towards an end where we operate in entire states as a whole instead of going city by city. Such growth was only possible because we have a worldclass engineering team on the frontier AI development here at Tesla. The currently operating robo taxi fleet is already running early versions of version 15 we referred to in the past. That's a 10x parameter one. For V15, we had planned roughly about seven major improvement tracks and they're all happening in parallel. The early V15 build that are running on Robbo taxi have already 40% of these tracks merged together. That's what's running in the fleet right now. Okay, so there's two parts to this, right? The first is that when they launch a city there less effort on the front and then separately in a separate part of the call and the analyst asked the question, why are you guys uh launching in city by city? Why not uh or or why not just scale Austin? Four of the Tesla executives went in and answered that question of what they're doing. But they all basically said something similar to this, which is the goal of why we're doing multiple cities right now is not because we plan to go city by city forever, which is what you know, Whimo has to do. No, it's like we just need a handful. Maybe you know these four these extra cities they just added, but maybe a few more. Once they nail that down, their goal is to get to zero trend towards zero uh towards a zero time to launch, which means then you're going to we're going to go to state by state. So it's like boom, at some point it's no longer going to be city by city. It will just be state by state, all state, anywhere because this thing works anywhere.

>> that's right. The geographic expansion as I wrote here becomes nonlinear and I think that's a very important point. Um, this is the the comment about, you know, a switch being turned on. It's going to feel like that in some places where you go into the whole state of Arizona, the whole state of Nevada or something at once or at least in the in the cities, for example. So, that that's really promising. Um, I think the second thing here, Herbert, is the the version 15.

>> Yep.

>> Um, this is something that they talked about, I think, six months ago in terms of what that would do. And so it's encouraging now that the vehicles are already running that uh at least to some degree and the question is now is how long will it take for the remaining tracks of V15 to be implemented. Um they didn't give us any sense of that unfortunately so we don't know. Um but I would guess that's probably relatively quickly maybe the next 3 months or so maybe sooner. Um and that's also a very encouraging thing. Uh Elon's talked about many times how the car is going to appear sentient. He's talked recently about how the car is going to remember where you want to park and things like that. Those are all massively favorable developments.

>> Yeah. So this whole thing about version 15. Uh so first of all I think it was a surprise to all of us that version 15 is actually being used in robo taxi now. That's great. Uh we don't know. It's a early version though, right? So we don't know. Sure. They're just trying it out but it's still early. They haven't completed. It could take months, could take a year to complete the whole thing. We do not know. They haven't told us. Um, but Elon did say on a separate uh comments, and again it's Elon comment, so you can take that with a grain of salt, but he did say that in the next few releases, we'll have memory and it'll have destination um reasoning where it'll remember where you want to park at the very, you know, like in my garage, it'll remember how I drive through my windy driveway. It'll know that it's this garage on this far right, and it'll know how I typically do it. That to me is is is the last bit of version 15. Remember the reasoning part um destination reasoning, but then there's all these other reasoning and other intersections. That's what they're waiting for because once that's complete, then the big complaint I have about uh like I took a robo taxi when I was in uh Fremont, California recently and it I was a hotel and I had to walk to the parking lot to get in. And I'm like, dude, that's just not the same experience as taxis or Ubers. It's got to pick me up at the lobby or further down. That's fine. What's going to happen now is every single hotel the every cyberc will know that that hotel go over here that hotel go over there intersections and that's how they'll be able to launch cities instantly. They basically go to the city and they go okay every one of these places this is how you you behave it's got memory good to go. So when that happens they might be able to do city by city so or state.

>> Yeah.

>> Yeah. And the second part of this is is the scaling of production of of the cyber cab. So there's a software component that we've just been talking about V15 particularly. And then if you look at the next uh section of quotes Herbert, he says, "As we continue to complete our work on V15, we'll see that the car is going to be ridiculously safe and capable. We're already seeing internal signs that all of these bets that we have made for technology improvements are all starting to work." That's a strong statement. We expect that the payoff will be huge and that's what's going to help us sustain the same growth rate through the rest of this year and so on.

>> Maybe comment about the same growth rate might mean that that's where people lack onto the 10% growth.

>> Well, again, but he said double digit. It's it could be between 10 and 20, right?

>> Mhm.

>> Um, finally, he says regarding the cyber cab, and this is a really important statement. He says first of all, it's a phenomenal product. Anyone who rides in it instantly falls in love with the experience. We have aligned our manufacturing targets to roughly match the projected growth rate of unsupervised miles. Now, aligning the production with that precisely is going to be a difficult task, but that tells me that that we'll see cyber cab ramp as as we start to see more cyber cabs come out of the fat tree and hopefully we get some some sign of that with Joe Tech's videos and others. Um, that'll be a good first hint that that the growth of unsupervised miles is is continuing. I also read this to think that okay what's the point of making more cyber cabs even if unsupervised miles are increasing on let's say Model Y's if you're not putting cyber cab in the robo taxi network. So I think we're going to see cyber cab being included in the robo taxi network fairly soon and call it the next 1 to 2 months I think is what we can read from the statement because what's the point of aligning the production with cyber cabab with unsupervised miles from model Y that doesn't make any sense unless you're using cyber cabab and robo taxi right that's the reason then to align the production because you're putting them in service and it's one thing I haven't heard really anybody talk about so much in terms of yeah cyber we're actually going to see cyber cabs deployed customers is being able to use cyberc cabs. Right now it's employees. That's great. Good test. But if you're going to produce more cyber cabs and if you're going to align that production with the growth and unsupervised miles, then that means you're putting them in service.

>> Yeah. Elon confirmed in this call u why the cyber cab hasn't been put into service yet. And I've been saying this and I did a whole show about two months ago now and I was seems to be correct. People were saying a cyber cap should be imminent. We're going to see them roll it out any day now. See how many they're out there. I said, "No, I think you need thousands of cyber cabs to be driven all over the the US and maybe even other countries to track FSD how good it works because it's a totally different form factor and and that's exactly what Elon said. So when that happens, when they feel comfortable that the cyber cab is behaving as a Model Y is because it's impeccable safety. That's why I think you still need about thousand cuz remember with the Model Y, we had millions. All of us were driving as supervised. And you can't just do it with just hundreds. You need thousands of cyber cabs out there driving as many miles as possible to accumulate FSD miles to then get the confidence that you've got it dialed down to that form factor. Then they'll release it. Because the whole point of robo taxi is cyber cap. It is cybercap. 90 plus% of robo taxi will be cyber cap. So there's there's only so much you can learn with the model wise. It's got to be proven with the cyber cap. So I think you're right. But I I agree it could be a couple months it was still away or longer. So don't hold your breath on that one.

>> Yeah. I don't know how long that sort of validation procedure will take with cybercap. We're seeing you know a whole fleet of cyber cabs being driven around the test track. We're seeing a whole fleet of uh engineering test vehicles in in dozens and dozens of cities all around the United States in all kinds of crazy places. They're they're driving them around testing them. Um those are all things that they have to do. And the question is how long that will take. I don't know.

>> We don't know.

>> Um, but probably over the next few months, I would imagine.

>> Yeah.

>> All right. Then you say later in the call, Ashok made these comments. The reason we've been expanding across different cities instead of just doubling down on a single city is that we want to make sure that our stack is very general one. It is a general one. We just want to prove to ourselves and to other folks that is working across a lot of different cities without too much effort per city. That's what we see internally. All right. So basically they're already seeing that they can just take what they're doing, move it to another city, and it doesn't take that much effort. Like it it really is a general purpose. So everybody's complaining and they're going, I thought they're not supposed to go city by city. The whole point is a generalized FSE. Well, yeah, of course, but you want to first do city by city to confirm that it is generalized. As Elon mentioned earlier, the growth rate has been literally exponential. Just it's in the early part of the exponential. That's why it's hard for others to comprehend in terms of miles versus vehicles. Since these vehicles operating the world basically continuously as opposed to human drivers who use vehicles for maybe a couple hours a day or something like that, these vehicles are mostly continuous operation which means that even for a few vehicles you can get a lot of miles out of them. That's why we're tracking the amount of unsupervised miles or just the number of vehicles or something like that. So the the whole point of what he's saying here is so critical. This is the part that um again I'm doing a whole show on this. The whole thing people don't understand like people go okay you need to have a thousand cyber cabs thousand vehicles out there why don't we see that already today no what they're trying to do is have a single vehicle and show that they can take that single vehicle to as high utilization as possible not just 50%. They want to figure out how to you know maximize the utilization of that one vehicle maybe 10 vehicles because once they do that then they'll do a thousand you know then they'll grow it. You don't need to have a thousand to try to maximize one vehicle's operation.

>> That's what he's saying there, right?

>> Yeah, I think so. It's it's the idea of this initial fleet is not to establish a commercial like a meaningful commercial service. Uh it's really testing to make sure that this autonomy approach is generalizable to to many areas um with you know without as I point out here extensive retraining mapping or local engineering. Now they mentioned a few months ago that there would be some of that. Absolutely. there's different ways that cities and states design their roadways and signage and so on. So the question is then can FSD, you know, adapt to that and how difficult is that going to be? And that's what they're testing now. Um, and I, by the way, I've noticed that I I moved from Florida to Texas. These are the two states right now where where Tesla has these robo taxes in operation. There are differences in how those two states design the roadways and signage. I've noticed that. Mhm.

>> So, this is a good test for them.

>> Well, I mean, the reality is uh there's millions of us driving FSD all over the US, different countries, Canada, now in Europe, right?

>> And basically, everyone is reporting safety has been solved to the to a certain extent. Again, it's all about large numbers, so we don't know exactly to how many nines of safety they're at yet. That's what part of the comment. But basically, here I am in Washington, you're in Texas, Jeff is in Chicago, Alexandra's in California. You know, this drives you anywhere.

>> No. Now, here's what I would say. I think is to the extent that the V15 resolves the issues that they've been working on, I think you could see more model Y's being introduced to the fleet. to the extent that CyberCab is ready, that they've done all the validation work and they've verified that the camera placement and everything works just as well as what they're seeing with Model Y, that's when we start to see the Cyber Cab being deployed.

>> Yeah.

>> So, there's there's some different milestones here that could happen.

>> All right, you we're going to talk about is Tesla being too cautious with robo taxi, but let's let's conclude this section here. Um, it is delayed. A lot of the things that was promised did not happen. And I don't know if they they tried to explain some of it, but it is delayed. Uh the people who think that nothing's much is going to happen for robo taxi might be true, might be correct that they're probably true that it won't happen in 2026. That the actual scaling will happen mid 2027, but the good news is that safety seems to they're very confident safety solved. They're confident that their solution works and they're they're confident that they're it's going to be generalized and they can go state by state. So, it's painfully slow right now and a lot of investors are pretty pissed. But,

>> yeah, I would just add that when you look at the previous statements and I've gone through and and examined everything that Elon said over the last five earnings calls about expectations about scaling and being in, you know, 50% of the population, cities were accounts for 50% of the population and so on. He always hedged those comments with things like, you know, pending regulatory approval, pending, you know, making sure it was safe. He never just flat out said, "Here's what we're going to do." Now, a lot of investors just took that comment at face value and don't recall the context around what was said. So, I think we need to recognize that. At the same time, they kept saying that. Now, things have been adjusted. So, clearly, it's been more difficult than they anticipated. And I wish that they perhaps would have been maybe a little bit more forthcoming about the changes in their thinking. Um, on a quarterly earnings call, you know, you get that four times a year. That's not much communication. It would be nice on something like that where investors had such expectations that maybe there would be more frequent updates about where they're at. That would be the one thing I would wish for. But I don't think that what they had said in the past was bad or misleading or or reckless. I I think that people are forgetting the context in which it was said.

>> True. I get it. But yes, you can't dance around it too much. They've missed milestone significantly.

>> Sure.

>> Yeah. And and I wish they would they give us more frequent updates perhaps on that.

>> And the people who are saying that uh you know what they heard at the earnings call means that Robboaxi is not going to scale this year. Uh there's more truth than that. But I think when you reread that we just did that does not not mean that there's there's a they're they're solving things to go massive scale and so things and they're doing it at a small number of vehicles first they're doing it at a city by city first they're doing it at FSD you know the final bits but all of them looks like it's coming and then when it does we're talking state by state that's the first time he talked about state by state that's a huge boom but then Of course people are pissed because they go into well I wanted it to happen like tomorrow

>> they pro they kind of promised it was going to you know we kind of got led but I agree with you context matters. Uh any final conclusions you want to talk about this Ashok's comments?

>> No I think we've covered it mostly Herbert. I think we we can go on to to the question of you know should they be rolling out uh robo taxi sooner? Are they being too cautious? Um, I think this is a good thing to think about and particularly people compare this to what Whimo's doing and I think there are some important important differences. Um, so Elon on the call talked about, you know, causing no harm even even to pets and that's a nice a nice goal. It's a nice thing to say. They want this to be as safe as as humanly possible, but it's not it's not something that that long-term, you know, perfection is not possible. You can't have perfection here. Um, the key is making sure it's materially safer than human driving. That's that's the issue. And that's also what Whimo's focused on that we'll talk about in a second. The other thing I think I would say is that Elon is is managing political risk here as much as he's managing safety risk. And I think for Tesla, which seems to be under the microscope more than Whimo might be, although I think that's changing for Whimo, some of their supporters uh politically uh lately have kind of turned against them a bit more as there's been some high-profile issues with Whimo in terms of blocking traffic and so on. So even Whimo is having, I think, a bit of a political backlash right now, even from some of their supporters. So this is an important thing for the company to manage through.

>> People forget what's happened in China. in China. Everybody goes, "Did you not know all of these uh robo taxi services are already out there? They've got way more like 2,000 3,000 cars. Forget the fact that many of them are remote remote operator, remote monitored, but China had to shut them down or had to put more regulations to slow them down because of these issues that Whimo was having. Whimo and these other BYU and other services, they all have a lot of issues. They will stop at certain areas. Whimos going through puddle uh flooded streets they they will just stop uh crash or you know or of course accidents but overall despite the safety level these final pieces is being an issue such that China shut them down right now in Europe there's a vote that's going to happen and it's it may turn out that despite the fact that five countries have approved FSD B, if the vote if fails, uh, France has already said that they're going to vote no. If Germany votes no and just another country votes no, all of it votes no. And so if Tesla had come out and did something that had some major incident, all of Europe would be that it will be done. That's for sure that's done. Like and and by the way, when they say no, even the five countries that have said yes have those are just temporary provisional approval. they would have to stop. So that's part of what he's working on too is like, okay, do we just let this go and see what happens? That's not but also this. So is is Tesla being too cautious with robo taxing? Now you say right off the bat, no. Um can you just why you say that? Because obviously the the two the big debate that people have right now is that they're a lot of people are pissed. They're going it's already safe now. We know this now. Supervised FSE is safer than humans. We know that it's safe now. You just heard Elon um Ashok say that. Why are you slowing this down when you could be saving lives now? The longer you wait to roll this out, Tesla, you're you are now not saving lives. But then of course the other argument is that if that one incident occurs, the whole program can be shut down for a long period of time. You've just hurt yourself. There's that balance between the two. But you say no, they're not being too cautious.

>> Yeah. uh if you look at it through an investor lens, a short-term investor lens where you want revenue and you want linear progression, then you are frustrated at this. If you put your longerterm hat on and say, "Okay, what what is the desired end state?" The desired end state is to have as safe as possible autonomous robo taxi network in as many cities and as many countries worldwide. What's the best pathway to get there? Well, the one of the best pathways is the one with least friction. And the least friction is the least amount of political issues as you just mentioned. So if you're rolling out an autonomous robo taxi operation and the European Union's having a vote about supervised FSD and they look at your robo taxi operation, they say, "Look at all the issues you're having. We're not going to approve the supervised FSD. Your robo taxi operation is having all these issues. So you might want to lay low a little bit while that that procedure is going through." That that that's part of it. I mean, I'm not suggesting that Tesla has to wait for the EU to to to approve supervised FSD before they start ramping autonomous robo taxis. But it's part of the equation. It's got to be part of their thinking in terms of minimalizing the political backlash. Also, to create a little bit of daylight between what Whimo is experiencing and what Tesla is experiencing also is not a bad thing. There's a lot of frustrations about Whimo. blocking traffic, people getting trapped in cars, all kinds of issues, right? Tesla doesn't want any of that. So, that's not a bad thing to have. The other thing is Tesla is on the cusp of potentially a big jump in FSD version 15, 10x parameters, right? This car is supposed to be more sentient, have memory, etc., etc. Why not wait until you know you've got something that's even better coming out just around the corner? What's the point of rushing that today? What what's to be gained? A few dollars in revenue maybe, but so what? Right? And I think that's the disconnect between investors who are short-term focused that want to see that linear progression and a company that's more focused on an endstate. That's the the most desirable one.

>> Yeah. I saw somebody on X uh do an analysis. I don't know how accurate this is, but it was at least somebody did the analysis. They said that, you know, Elon talked about March 9 and we don't know what that means. According to this person's analysis, they said that uh the 38,000 miles uns unservised miles driven with zero notable incidents, that is 99.999%. To get to 99.999%, that's 69s, 49's dots, uh you need to get a million miles unsupervised without any incidents. And their calculation is that we might get there by the end of this year kind of thing. But once you get there, that is in comparison to human drivers, a human driver would have a uh not a not only death, but actually every 1 million there is a mass a serious injury. So if Tesla is able to say in every 1 million we don't even have any incidence at all. not only just death but also no no serious injury. That is completely like beyond proof at that point that this is much safer. You they might actually already claim that now because as you had said earlier the 380,000 is unsupervised driving with nobody in the car but they actually have already 1.2 million driven in robo taxi with a person on the you know unsupervised like they're not touching the wheels. It can probably kind of like show that they're not even touching wheels. The car is driving itself. They could already prove that, but they want to prove it with robo taxi with no one in the car.

>> They have two and a half million miles of that.

>> Okay.

>> Right. With no fatality.

>> Yeah.

>> So that that's something that that data is still worth something for sure.

>> But you're right that that data speaks volumes. The other thing too that I highlighted and Herbert if you maybe want to bring up the the article um they talked about on the call about deploying Starlink uh in the

>> me now here

>> just scroll down until you see the Starlink um keep going um and so they talked about this being an important part of robo taxi itself that they don't really want to lose you know contact with the vehicle and there are certain situations even in places like Silicon where you might lose communication with the vehicle. And so from a fleet management standpoint, this is important. From a customer experience inside the car, if you're use if you're using kind of entertainment, this might be also important. So if they've figured this out and they're saying this is important for us, then again, why not take the time with respect to cyber cab?

>> Yeah. to put this in all the vehicles and have this rolled out as part of the customer service. Um, that's that's another reason. I mean, it it's not necessarily by itself why you might want to slow roll the roll out, but it's another factor. It's like, well, okay, we've got this thing that makes this vehicle better and more reliable. Let's wait until we can deploy this. Well, that is the main reason that BYU was stopped by the Chinese regulation uh regulators uh and that's why Whimo was stopped and they're being investigated now too because every once in a while there's a power outage in San Francisco. Hundreds of Whimos stop mid just stop there's abandon causing havoc throughout the city. Imagine with the uproar and that's why you get a lot of people hating Whimo now. They you know the the vigilantes who will come out there and they'll like you know and and um so that's with Whimo having 3,500 cars if your plan is to have 10,000. Why aren't why don't you roll out a million over the tomorrow? Well, imagine what would happen if if that was an issue. And so if you have the Starlink integrated into the cyber cap, you don't need Starlink to make it safe, but you need Starlink in order to um make sure that the communic you don't lose communication so that it doesn't um it's like it's like you don't know like I guess what happened was shut down like doesn't the car doesn't know what to do. I don't know. Or for for robo taxi management of where the cars are and what what to do where they should go next. The interesting thing too, Herbert, is just on the topic of the growth in the unsupported miles. You mentioned I think that you might we might get to a million by the end of the year. Actually, I think it might be a million and maybe 10 weeks. Uh again, depending on

>> I don't know that guy I don't know. I I didn't look closely at his calculation, but what what makes what's your calculation? How did you come up with

>> Well, at 10% weekly growth, you double every seven weeks.

>> Yes.

>> Okay. And it would grow roughly 12 times in six months and over a year, you'd be 140x. Again, I don't think it's going to be a smooth 10% growth. There may be actually some stair steps in that as well that that would actually potentially accelerate that. So, we I think we need to be careful modeling this out. I mean, it's tempting to put a spreadsheet together and model it out and get a beautiful curve and you can look at 10% growth versus, you know, 14% growth versus 17% growth and map it all out. And those are good exercises and uh some people have done that and it's been it's been helpful. But I I do think that we need to be careful making those kinds of projections. Um, but I think what we're seeing is rather than slow growth, I think it's a controlled exponential and a shock made the point of mentioning that in his comments.

>> Uh people don't realize that a a Whimo a a cyber cab will drive three times uh more miles than a Whimo does just because of the battery. Just that alone. And if you add all the other operational efficiencies, it can drive more. So they once they start using the cyber caps, they can and then they have full utilization of it, not just 50% utilization. Like they actually are able to utilize. That's what they're trying to get to. So one car can continue driving forever. Basically 24 hours a day is what their goal is. Charge really quickly and keep going. And then they can accumulate, accumulate, accumulate. So they might be able to get to the million sooner, we think. Agreed. Mhm. It's also interesting just on the topic of Whimo versus Tesla, Whimo has announced that they've got 220 million fully autonomous miles.

>> Yeah.

>> So that's 580 times more than Tesla has. Now again,

>> yeah,

>> Tesla's 12 billion miles FSD miles are not junk in this regard.

>> But just in terms of where Whimo is at in terms of fully autonomous miles with no one in the car, um, there's there's some differences there. The other thing too I would say is that Whimo's approach is um when the car is uncertain about something, their default is inaction

>> until a remote operator tells them what to do.

>> Yeah. And again, as you mentioned, if there's a power outage in the city, that remote operation may not be available.

>> And that's why there is because there's a remote operator, just not enough of them. And so many of them just sit there waiting for somebody to tell them what to do. And there's just not enough. And that's why Tesla goes, if that's 3,500 cars and you have that equip, we can't do that. We got to remove the need for remote operators. We need to have this all AI driven and managed and then we can go beyond 3,000 cars. We don't even want to get to 30. If if if Tesla does exactly what Whimo does, they can do 3,000 cars tomorrow and then they can have the same number of incidents that Tesla Whimo has. What would happen? It's worse. It would be everything shut down. It would be no one would

>> I'm just not sure where it gets. Ultimately, you want a car that can figure its way out, that can reason out of any situation. And that's what they think they have in their software. Again, V15 10x the parameters sentient has memory, etc. That's what they're driving towards. That's what they're heading towards. Uh Whimo hasn't resolved that yet.

>> All right. So, you you say this though, you Tesla's roll out may be faster than it looks. What what's uh what do you mean?

>> Well, we're all measuring progress by the number of vehicles, and that's kind of incomplete as it is. Um, there's registrations, there's number of vehicles actually in use, and it's kind of hard to measure that. Some of the data, I'm not sure anymore is accurate. Um, what is more important is the unsupervised miles. And unfortunately, we don't have a real-time look at that. So, I wish we did. I wish there was a way to do that. Um, but this is ultimately the metric. And of course, if you look at what the growth is, even at 10% growth, as I mentioned, you get to some pretty astronomical numbers pretty fast, right? So Tesla can can, you know, really get the number of unsupervised miles that they feel that they need to in order to have the data to be able to go to regulators and look and say, "This is so safe." You know, we haven't had any fatalities in 10 million miles or whatever the number is. You know, look how safe this is. It also, you know, again, coincides with the validation of CyberCab, installing Starlinks, all that stuff. It just makes it a better service out of the gate when they're rolling out into new cities. So, take the time to do this, right? Um, so the the growth can come very quickly and again, this is just a 10% weekly growth. I I wouldn't anchor on this number too much. Tesla can actually have a stair step above that. So, this article is quite lengthy. Uh if you want to skip down to the to the versus the Whimo section, I think there's some interesting data there from Whimo. Um, so I encourage anybody that's more interested in this to to read the full article, but uh here's a discussion about margin of nines and so on.

>> Um

>> wow, you really covered a lot of ground here. Okay.

>> Yeah.

>> You want me to go right to

>> just go to the Whimo section, I think, because again, I think this is the the big issue that people are are focused on. You know, I don't I wouldn't say that Whimo is necessarily being less

cautious than Tesla because they have a different architecture. Their cautions are in a different place. And as I mentioned, when there's a problem, they default to inaction. Tesla doesn't want to do that. They don't want to default to inaction. They want the car to reason its way out of a problem. And as you mentioned, at scale, that's critical. Okay.

Now, hopefully Waymo, you know, for their sake, is working on a solution to this too, so that their default is not inaction. But right now, that's what it is. Um, they have, you can see quite a few miles, uh, rider-only miles in Los Angeles, 52 million, San Francisco, 67 million, Phoenix, 80 million, Austin, 15, and Atlanta, now 5 million. So they're racking up these rider-only miles, you know, to a nice degree. Um, and you can look at their data: 94% fewer serious or fatal injury crashes, 82% fewer injury-causing crashes than comparable human driving. So they're, they're really getting some nice data to show regulators how safe their service is.

Now, they're creating other issues again. Cars sitting there and not doing something is creating its own set of issues. Tesla wants to avoid that. They have a different solution than what Waymo is offering here. But I wouldn't necessarily say that Waymo is being less cautious than Tesla. It's just, given their different architecture, this is the way that they have to do it. They are doing a nice job, I think, in getting the number, you know, rider-only autonomous miles. They're racking up a good number, and there's lots of data, and this is from their website, and there's a link in here that people can go to to see all this data. Um, but they're proving that their car is safer than human drivers. Okay.

Again, it's causing other unique problems. You know, it's causing traffic jams and so on. An inactive car in the middle of the street is an issue, particularly when, you know, uh, first responders have to get somewhere and you've got five Waymos blocking an intersection. That is not, not a good situation.

>> Okay, investor interpretation.

>> Yeah. So I would just say again, my comment here is that Tesla is not being too cautious. If you look at what the end state is, they're taking the steps I think that they need to take. Do we, as investors, want them to go faster? Well, of course, we want it yesterday. Are they taking a prudent, smart approach? For whether at, I think so. Again, we don't have all the information that they have, so we can't independently validate that. But when a guy like Ashok comes on the call, somebody I think that has a pristine reputation for the things that he has said, there's a lot of people that say, again, when Elon talks, we disregard it. When Ashok talks, we listen. Well, here's a guy telling you what, what they're doing, what their plans are, and it's, it's very, very encouraging. So my conclusion is that Tesla is being appropriately cautious, but I do expect to see an acceleration in the coming, you know, months, quarters. Again, I can't predict that because I don't know where they're at with the V15 development. So I'm not making a prediction here, but just as they make progress on those things, we should see an acceleration, and it might be a bigger acceleration than many, many understand.

>> Okay. Well, I appreciate you, you came up with all that information. Is Tesla being too cautious with robo taxi? The answer you believe is no. But the, the result though, even if they are not being too cautious, is that it's, uh, the time frame for a rollout is a lot slower than people were hoping for and slower than what Tesla themselves had initially, you know, communicated.

>> What's interesting about that though, Herbert, is again, I think most people think in a linear way. And I think Elon, even in maybe the Q4 conference call, was talking about doubling the number of vehicles.

>> Here we are.

>> Right.

>> Right. And and so if you're starting at at a low point and you're expecting there to be a, you know, to be here at the end of the year, and and today you're still at this low point, you feel like they're way behind. But on an exponential curve,

>> you can catch up very quickly. Like

>> in the month of November, for example, you could catch up completely to where people thought they would be at the end of the year on a linear curve.

>> Yes.

>> Right. Just because of the step change that can happen in an exponential situation. So I just caution people to to not be so focused on linear progress. FSD improvement is nonlinear. Adding vehicles to a city, adding cities to the network also could be nonlinear. And they mentioned that again, talking about state-by-state activation. So we could be very surprised at some point, pleasantly surprised about this the nonlinear aspect of this.

I think the signal for us to be really watching for right now is version 15. And I think you said something like, if you take a look at the, you know, one chart in the earnings call, few even talked about, which is that they doubled their Colossus to uh, Cortex to uh, AI in the last several months. And everybody goes, "Well, that's cool." Moved on. Well, what's that being used for? You asked me. We had that discussion. I'm going, "Well, it's not for Optimus yet because there's no Optimus. They need a thousand of Optimus before they can start using that for that." So, they launched version 15. That is a 10x parameter one. If they are able to, and now we don't know this, it could take them a year, take them six months, we don't know how long. But if they can solve the final, you know, seven parallel tracks, the version 15, then maybe they can then do the exponential. Because Ashok has said that version 15 is, I don't know what words, remember, we just covered it, ridiculously safe or whatever.

>> This is now the po, this is it. This is the ridiculous safe one. This is the one that we're waiting for. So when that happens, and then it's got the final, uh, and destination memory and parking. So I think that's the one we need to really focus on. All right.

>> And V15 is again, already running in robot taxi cars to some degree.

>> That's what he said. Yeah. So that's, that's I think the surprise. And so now it's okay, maybe.

>> So it's not, it's not about us waiting for V15 in our vehicles. That, that's not the metric we're looking, just to be clear. That's not what we're looking for. That's, that's great. I can't wait to get it myself. But for Tesla, it's V15 and the robo taxis.

>> Yeah. Yeah. All right, Elon, 13 takeaways from Elon's Q2 comments. Um, and in your central message, what you got out of it, the Elon central message is Tesla's entering a much broader and more capital-intensive growth phase. I think we knew that. Uh, FSD is a demand engine for the automotive business. Yeah, we're here.

>> 55% of the new vehicles sold in North America had FSD attached at the time of delivery. Uh, it was 42% I believe, uh, for the entire world. If you look at the global deliveries versus the number of new FSD subscriptions and purchases, um, that is absolutely phenomenal. It is a massive demand driver. We're seeing it more and more, and I think this will only accelerate from here.

>> And, um, and this is why we, many of us think that Tesla should be advertising and telling everybody FSD now to to drive vehicle sales. But I think Elon thinks that the best way to do this is have robo taxi out there, and then people realize, uh, cars can drive themselves. And I think Loris, when I asked him that question, he answered something similar, which is, you can advertise all you want. You could tell people till you're blue in the face. Nobody's going to care until they actually get sit in the car and actually test drive it themselves. That's the only way to convince somebody. Like you see compared it to flying airlines, like until somebody gets into an airplane, flies once, do they realize it works. Otherwise, you can tell them blue face, there's these planes now that could fly you. No, yeah, sure. Won't believe you.

I think one of the things that you, if I think about their lack of willingness to advertise FSD broadly,

>> not that they don't advertise FSD, but just broadly to the general public, is it's still beta software. It's still got some rough edges to it. So the people that find it naturally and want it,

>> they're more likely to deal with the rough edges and the fact that the car doesn't park where they want it to park every single time or whatever the issue is.

>> If you advertise it more broadly, then you run the risk of people being disappointed that this, okay, it's not full self-driving. It can't do everything I want it to do. Then you've potentially got some problems. So I don't necessarily mind their approach here. Again, as a short-term shareholder, it's frustrating. You want them to advertise. You want them to save as many lives as possible. Absolutely. But from an engineering standpoint, it's not quite there yet where you could say it's solved. It's complete. You don't have to do anything. You can fall asleep at that point.

I think we're there. I'm point to that point. To me, there's no no reason then not to advertise at that point.

>> True. I I get you. But I think that they should be advertising it now. Um, they made a huge mistake, by the way, when they gave away free trials of FSE 12, whatever it was, early versions of it. I have my brother and my cousins who I told them, but yeah, my FSE is driving me like crazy. Uh, and they're like, "Yeah, I tried it myself. Doesn't work." And I said, "Well, wait, wait. You got to have hardware 4 and you have to have the latest version."

>> And then like, but in their minds is already set. This didn't work. It was a complete backfire. It was too early. I think I think I think it's we're listen, what we would just cover FSE is now a demand driver. So people who even rewards and all it's still today safer now, and you got to get it out there. So I think that they should be pushing it.

Um, okay, robo taxi will transform Tesla from a manufacturer into a transportation operator.

>> Yeah. And the most important thing here, I think, is the, the chart that goes with it. If you scroll down, um, we've talked about this before, the nonlinear nature of this progress. So, if you look at the, the blue line, and I've mapped this out kind of in a linear way, and I don't think it is, but the technical capability, the reliability of FSD is improving all the time. Again, not smoothly like this. There's stair steps here. There comes a point in the middle of the chart that I've labeled deployment threshold. That is the point where Tesla is comfortable in saying, okay, this is good enough. We still got some things to work on. We we can always improve that. That blue line's always going to go up over time. Maybe less rapidly, but they're always going to make improvements in FSD. But we've reached a threshold now where it is reliable. It is safe. It is great. Let's roll this out. And so then you get that beginning of the yes curve in the revenue. Right now, the difference between the blue line and the green line is what I call investor anxiety.

>> We all in our mind see how great it is. But what we don't see is the operational issues and the other smaller stuff that Tesla is dealing with. And we don't think about what that would do at scale. As you mentioned about Waymo, if you've got, you know, five cars in an intersection that's stuck, fine. You can handle that. But if you've got 3,000 in a city all stuck at the same time, good luck, right? So

>> yeah,

>> there is a point here. I think that that it's almost like that switch-on moment,

>> the deployment threshold, peak investor anxiety. I don't know if this is it right now, but it sure feels like it. Yeah, nice chart. Good explanation.

Okay, the third thing that Elon said during the earnings call, Optimus has a large opportunity but the least certain timing. I mean, he said this many times before already, right? Anything new here?

>> Well, it's funny because in contrast to the the things that people find fault with Elon about robo taxi expectations, 50% of the population by the end of the year, he's not saying anything about that with Optimus. In fact, he is slow-walking this to the nth degree in my opinion. Right. And he talks about how every component is new and there's no supply chain and and all this stuff, and and he's right, he's right to be cautious about this. But at the same time, he's he's reminding us that this is going to be the biggest product for Tesla that's that we currently see. Um, so I like the fact that he's he's he's constantly reminding us about that. It's kind of like again, Nvidia, when Jensen Wong was saying how big AI was going to be, he had to remind people for years about this before anybody believed him.

>> Yeah. But again, it's Elon shooting himself in the foot because just what a year, two years ago, he was the one that said we'll have 5,000 Optimus this year. We're going to have, he replied to your chart or what I remember where he said each year will be 10x magnitude. All the CEOs of robots, robot robots were basically saying the same thing that we see each year, we will 10x what we did the fi final year. And people agree with that. We were expecting in 2025, 2026, and now everything is like, whoa, way later. I again, doing a show on this. If you take a look what they did with Robo Taxi, that's what they're going to do with um Optimus, which is they're building for scale. When you're building for scale, that means that you can go, if they wanted to, they can build 2,000 Optimus today. They can go ahead and put 2,000 Optimus in factories doing pick and place work today. They can do that. That's not what they're doing. What they're going to do is build 2,000 Optimus, robo taxis in the next several months, put them into university because they're training them for generalized ro intelligence and and and tasks. So that the goal, just like up um, just like robo taxi, is put the car anywhere on Earth and it'll be able to be a robo taxi. It's to put a robot anywhere, show it a video once, show it a demo once, it'll be able to do that task, any task. So to do that, their their goal is not to try to do just this one task and just this one task. It's to be generalized. And so you're not going to see every robot that they make is for training until they got the generalized. And so it's going to be painfully slow. And it won't be until version four of Optimus that you're going to start seeing that on the wild and being used. The encouraging thing for me here is that we're not only seeing the Model S next line being taken down and being made way for a million capacity Optimus line, and they shared photos of that in the slide deck this quarter. They're making progress on that. We're not only seeing that, but we're seeing a new factory in Texas that's almost as big as the car factory going up. Steel is being erected. That thing is going to be enormous. As great as Elon is with capital efficiency, do we all now think he's making a mistake in building this factory and spending the billions of dollars on it and he won't be able to produce robots at scale out of this factory for five more years?

>> right?

>> No, I don't think that Elon is making that kind of mistake. Could he be off by a few months or a year? Maybe. Sure. Okay.

>> He he will be. He he's always off by few months and a year. But you don't go building a factory 10 million units and not to say that things

>> be off by a year. It's very painful. It'll be very costly, and that'll be it could take the company say a serious hit. Yeah.

>> And particularly right now when everybody seems to be so focused on capex and their free cash flow negative and all this. Well, if if there's any chance then that Optimus is going to be delayed, then then why build that factory now? Why spend that money when everybody's so focused on companies spending so much on capex? Put that elsewhere.

>> So, to me, that's a, that's a strong long-term sign. Again, it's not going to affect the numbers in the next couple years, but I'm I'm pleased to see that they're moving ahead with that in a major way.

Fourth thing that Elon said, terafab signals that Tesla sees chips as a strategic production constraint. Again, nothing new, right?

>> Nothing new, but again, he keeps talking about it.

This is something that if Tesla can solve for this, they will put them in a position that no other company will have. So, we shouldn't lose sight of this. I I said that when the Tera Fab was announced that this was a trillion dollars in present value for Tesla. Of course, the stock does not reflect that. But if you believe that there's even a 50% chance that this works, the value created from Tera Fab is absolutely enormous, and I and I stand by my my statement on that. The market is not going to value this at all until the the the factories build and chips start coming out. I get that. But just from a present value standpoint, there's a huge amount of present value attached to Tera Fab.

Um,

>> well, when they do the Tera Fab announcement and then maybe do a do a full deep dive road show on it, then I think people might get excited once they understand what this truly is. Digital Optimus could become a business independent of the physical robot.

This is the biggest sleeper for me. This is something that they've talked about. Maybe SpaceX has talked about it a bit more with Grok, but this is something that you don't need to build a factory for other than than the AI compute. You can roll out, you know, a dig a digital optimist without having to build, you know, the physical robot, right? And so this is something that they've talked about. We haven't seen anything yet. I think Elon, when he first talked about this, said maybe by the end of this year, they'd be ready to show something. Um, but again, he talked about this on the call. He also mentioned that the training digital Optimus is actually also good for training Optimus itself, the physical robot, that a useful humanoid robot will need to operate touchscreens, terminals, machinery, interfaces, and computers designed for humans. So not all is lost here in terms of also helping with Optimus. But this could be a huge revenue generator for Tesla by itself, and this is a a non-physical, non-manufactured, you know, product.

>> Yeah.

>> But we haven't heard much about this yet, so everybody's completely ignored it.

>> Well, I I think it's going to be part of the SpaceX Tesla road show when they come out. They're going to share Grok, digital Optimus, and what it can do. But if it works, we don't know. It could be a complete dud. But if it works like you're saying, it could be instantly what um, what Anthropic has with their cloud work and cloud code. They can come out and start selling and being charged to customers, especially with Cursor already, and 80% 70% of all Fortune 500 companies use Cursor, which Tesla, SpaceX owns. SpaceX owns. They can distribute this and that can generate revenue.

>> And the important thing to mention is this is without any APIs.

>> Yeah.

>> This is basically sitting an AI down at your computer screen as though it was you, computer. Any computer screen.

>> Yeah, but I think uh Claude just rolled out that feature too. So, which is, uh, you just, was it OpenAI?

>> Yeah, you just give it access to your browser, your any app, and you just show it, and you verbally tell it, like you would an employee, here's how I use my app, and then you verbally tell it, and you're just moving the mouse around, clicking here, and then you'll be able to replicate it, and then you don't need APIs. And that's where I'm like, still questioning whether or not what they say that vision in action out FSD, but now with what did they what they call full self computer instead of full self-driving, full self computer, um, if that's true, if that is real advantage versus what Anthropic can do and others can do.

>> Well, we'll see. But if you think about the technology, if you think about if it works for cars, and again, those of us

>> I get the logic.

>> right?

>> But on a computer screen, computers can do this. You show a demo and you know, like what's the difference?

>> Honestly, again, that's not clear yet in terms of what the advantage of applying FSD to it is versus what maybe OpenAI has without all the FSD development that Tesla's done.

>> right?

>> Yeah. That that remains to be seen. We haven't seen it.

Somebody has to explain to me in detail the advantage of a full self computer versus just a computer looking at your demo and replicating it and and maybe understanding context. But so can you know, these LLMs are so smart now, they can understand context, especially when it's digital context. Don't know. We'll see. You need to understand. Megapod could turn Tesla's physical footprint into distributed AI infrastructure.

Again, this is another product that's been announced and we haven't actually seen anything physical yet. Um, but he's talking about, continues to talk about this. So, you know, just a reminder, if if again, Elon's focused on the longer term and building for the longer term, um, this is not to be lost. U, this would allow Tesla to monetize their their infrastructure, their supercharging network, etc., um, and so on. So this, this is also a very important aspect of Tesla's future. Again, none of this is for short-term investors. This is longer-term investors. Um, and Elon continues to talk about this. Um, number seven, Herbert.

>> Wait, before you move on there. First of all, he talked about Megapod in his opening statement. And then two, I don't know what you mean by it's a long-term thing. I mean, the AI data centers is a thing right now. They can put up an AI data center in like three months. They've got that experience now. Both Tesla and SpaceX Megapods already existing, you know? So I don't know. Uh, this could actually be a sleeper thing that can happen sooner than we realize, as these are micro data centers distributed at multiple locations. Could you not, you know, why, how hard is it to create a Megapod in a supercharger station?

>> Well, we'll see. And this could also be an interesting sort of Intel partnership with the x86 computers that are part of it. Um,

>> uh, so energy, you know, obviously, we know energy is is an important growth vehicle, and we've seen how SpaceX has used the um, the Megapack as a way to smooth out the energy going into their their data centers. If you have any energy spikes, that's really harmful for running these models. So now, you know, Mega is is critical infrastructure for AI. It's part of building a data center, and SpaceX has proved that out. That's probably going to be a demand driver for Mega. Um, the good thing about that is that their customers are probably less price sensitive. Um, we're seeing that the price per kilowatt hour for Megapacks have come way down, and that's to be expected. As as the costs come down, those two curves are coming down together. But from a margin perspective, going forward, to the extent that that data centers begin to use this technology to smooth out their energy, um, that's a good thing, and that's a favorable development, I think, for Tesla.

>> Yep.

>> Okay, solar manufacturing. They mentioned it, but you know, we haven't really seen anything. Any there hasn't been a solar day or anything like that. So, this is kind of interesting. Um, I feel like there's a few set of product demos and so on that that we're due for here. Um, and maybe later on the summer, we'll start to see some of the stuff. Um, but they're talking about, you know, building, uh, solar manufacturing. Uh, do we even know, Herbert, where this is going to be built?

>> Well, I think, uh, both Tesla has their own um efforts, and so does SpaceX, and I think it is going to be out in New Jersey or

>> Yeah.

>> building a battery.

>> I don't know.

>> Yeah. So, this is interesting. I need to do some more work on this, but again, Elon mentioned it on the call.

>> Uh, autonomous semi. Yeah, he talked about that, too.

>> He talked about that. Uh, not a priority right now. There's other things that they're focusing on, but this will come. He said maybe end of 2026 or early 2027. Okay, add a year to that if you want to. Um, but this is also in Tesla's future that they'll work on. Um, SpaceX is becoming a more and more part of Tesla's uh, whole product suite. You know, Grok helping or relating to digital Optimus, Starlink connectivity in the cyber cab, um, potential Starlink installation and other Tesla vehicles, Tera Fab, Megapack use in X, SpaceX's AI data centers, and then also using Starlink terminals as ground connectivity relays was also mentioned.

>> vehicle.

>> And the vehicle.

>> Vehicle. Yeah, all the vehicles will have Starlink. He mentioned that first the star, then eventually all Tesla vehicles.

>> And each of these Starlink terminals can be Wi-Fi enabled, and so somebody walking by can get Wi-Fi too.

>> Yeah.

>> So again, another basically, you know, the whole point about, uh, AT&T and T-Mobile, they spend all these years building their satellites and all over the US and so forth.

Now you've got every car, you know, is a terminal. So you got this distributed connection.

Instead of going to a conference call for your video call, you jump in a cyber cab and have your call there because that's the where the best connection is.

Um, yeah, he spent a lot of time there. Existing Tesla vehicles may receive an important second life.

>> Yeah, I mean, this is, you know, upgrading the vehicles, upgrading maybe cameras, upgrading computers. Uh, with the with FSD V14 light, there's some evidence that they, you know, they're taking this seriously. They've rolled out this new version.

Um, you know, a lot of people worry about, uh, the risk, for example, if China takes over Taiwan and we lose, you know, semiconductor manufacturing ability. Well, Tesla has this massive installed fleet. Tesla could still grow its business if there's a massive supply chain disruption on the semiconductor side. Um, and here's one way they can do it, right? So that's kind of interesting. Um, although, of course, if you're upgrading the vehicles, you may need additional chips and so on as well. So it's not definitely doesn't totally solve the issue, but it's part of it. Um, his comments about, uh, speed over maximum capital efficiency. Uh, Elon and Tesla are well known for how efficient they are spending money. They spend a lot less than other companies and they get a lot more for it. Uh, he's saying that they're relaxing that, maybe those stringent requirements right now because speed is more important than getting, you know, all the value out out of every last dollar. Whereas other companies, maybe, you know, I'm sure Tesla's still going to be very focused on capital efficiency, but maybe just a little bit less so than many other companies are. And then last but not least, um, Micron, uh, is making room for Tesla in the years to come. So Elon highlighted back in October at the annual meeting how, uh, you know, the situation with Micron and memory. Um, and now we know that the situation is very tight. Micron is basically sold out for for next year or maybe even the following year. Um, and Elon has been able to get an allocation that he appears to be happy with. He's not complaining here about, uh, the memory allocation. So given his growth plans, Micro has made room for him. And that's that's nice because Micron has some other pretty big customers, like we've seen from Apple, that they're struggling to get the memory that they need, and yet Elon's being able to get what he needs. That's that's impressive.

>> This is huge. And he repeated it three times.

>> Uh, two times at least.

>> Okay. Overall interpretation of what Elon was saying.

>> Well, again, he was not focused on near-term earnings. He was giving us, I think, pretty clear view of their longer-term business scale. Um, and as a long-term investor, I think that that is really nice. Uh, again, people have been very critical about Elon for maybe some of his short-term, uh, prognostications that have not come true. Again, I think that they're maybe missing out some of the context that he gave on those on those statements, but even so, he's he's clearly now stepped away from that. Uh, he's letting some of his other lieutenants maybe give some of those more specific expectations, but Elon, I think, has painted a very nice picture for many of the different areas that Tesla's focused on for the long term. And I think some of these are really exciting and absolutely massive. So, as a Tesla investor, I think at the end of the day, you have to ask yourself, what's my time frame? And if your time frame is sufficiently long enough, um, the this company has never been better positioned. If your time frame is shorter and you're looking for nearer-term results, then I think you may be disappointed, and that's why the stock is reacting the way it has over the last few days.

>> Yeah.

>> But I'm reminded, Herbert, of a Peter Lynch quote.

And he says that the real key to making money in stocks is not to get scared out of them. And the first thing is is know what you own. Those were the two big statements that he made for investors is that when you experience volatility like this, if you don't truly know what you own, you're more likely to get scared out of the company because everybody else is reacting emotionally and they're selling it, and you're like, "Oh my gosh, what am I missing? Should I be selling too?" And many people do. But if you, if you know a company really well, and you see a company like Tesla investing for the future, investing its capital and building factories, multiple factories all around the world simultaneously, that's actually really encouraging.

>> Yeah.

>> Because in order to generate future revenue and cash flow, you've got to invest the money up front to do so. And Tesla is doing that now in an aggressive and in an accelerated way. So for me, looking at long-term, I've never been more excited about Tesla, which is interesting because if you look at the stock price, it's kind of like, well, what, where's the disconnect? And I think it's just short-term versus longer-term.

>> I mean, there's two things that can be true at the same time, right? So one is it's absolutely frustrating that things are taking a lot longer than we expected. There's no debate that, uh, Tesla themselves made announcements of certain milestones that they haven't hit. So that is true. It's frustrating as a Tesla investor. On the other side, it's also true, you kind of mentioned this earlier in the in the show here, where you said that it's also true now, though, that scaling of robotaxi is going to be a lot faster than we actually even had guessed earlier.

>> Yeah.

>> Because of how slow they're taking it now, their plan is to scale like crazy, and they've got everything in place. So it's frustrating this year where it's not as fast, but when they do scale next year and two years from now and three years from now, that part is going to be a lot, you, you know, your exponential curve needs to go faster, then you got to change your curve, move it over, but then it goes like that faster. So the the point of us getting global robo taxi is a lot faster today than it was, you know, six months ago of what we expected. So, it's frustrating, but that is what I heard scaling wise, right?

>> Yeah. Elon has talked about double exponential in previous earnings calls.

>> And I still think that that is what's ahead of us. He he has stopped talking about that. Ashok has kind of picked that up.

>> And I think that lends again more credibility to what the plans are going forward.

Okay. Again, thank you so much, CERN. Check out his website at uh brilliantadvice.net. You put in a lot of work here. You actually dove deep into three articles that you wrote and um, you, they're deep articles. We all watch the same earnings call, but, uh, you parse them out to what is signal, what isn't. People may not agree with some of the things you said there, but at least you laid it out and you know, let's review exactly what Ashok said, exactly what Elon said, and then let's try to figure out what exactly, because otherwise, what we're hearing from X is somebody just took a piece of what they said and then they run with it, and then, yeah, they're true. There's some truth to it, but it's also there's nuance. Let's dive a little closer. Thank you so much, sir. Appreciate you.

>> Thanks, Herbert.