저자: Kevin · 외국인
요약 보기
테슬라 로보택시 시승 영상. 주행 중 '후행차 따라가기' 편향으로 주차장 진입 시 안전요원이 개입했고, 원격운영 지원으로 3점 턴 후 탈출. 대체로 쾌적했다고 평가하나 엣지케이스의 상용화 리스크와 과금 버그(36분 탑승에 4.17달러만 청구)를 지적. 안전요원 제거 후 재무 모델을 가정해 하루 100만 건이면 주당 약 21센트(일부 가정 시 약 7.5달러 가치), 1000만 건이면 약 75달러 가치로 추산. 현 주가가 3000만 건 수준 기대를 반영 중이라는 의견.
다듬은 글
He just had to intervene. Yeah. Said stop in lane. Oh man, did we just break a road taxi? And what did we find out? Well, in this video, I'm going to first of all shout out Ellie in space. Thank you so much for inviting me to Austin, Texas. I was in Texas for a house hack project and popped over to Austin from Dallas to check out Robo Taxi.
All thanks to Ellien Space. So, shout out to her. Of course, the Tesla team for actually creating robo taxis. That said, we got to talk about what happened on this drive and what went right, what went wrong, and then of course we get to talk about the valuation of robo taxi. What do we think this means for the stock?
Because that's what we care about. So, first things first, meet Ellie. Hey everyone, me Kevin here. I am in Austin, Texas, and we're here with Ellie in space. What's up? Hey, welcome. Thank you. Thank you so much. We are going to get into a robo taxi. Yes. And that's what we do. We book a robo taxi.
Anyway, so um you know, you can pick basically anywhere in this area or you can pick some of their pre uh you know determined destinations. So book rides $65. Oh, the price went up. Okay. Yep. And um inflation ter and uh Yep. And so if the car gets here, it'll wait 15 minutes for us. So Oh, that's actually like much longer than the five minutes.
Usually I get ordering the ride was just like ordering an Uber. Unfortunately, when it came to finding the pickup, that was when things got a little funky. Now, this could be because Austin has a lot of run single directional roads, so one-way roads, but this is something that Tesla should be aware of.
And I do think the vehicle should have taken a more convenient pickup route choice, especially since we pinned at the valet or a hotel, which is a pretty common destination. That said, let's get into the car. Let's see. What does it say? Please enter the vehicle. So, it must be here. Okay. Some It is uh around the corner.
Is it this corner? Uh, Boba Steakhouse. Oh, Bob. Oh, Bob's steakhouse. That'll work. Boba Bob, you know. Hey, a boba and steak. That sounds kind of good, actually. I don't know. Never had the combo before. So, all right. So, we did request to do the W, right? We did. All right. So, I mean, I guess the building sort of wraps around, but All right.
Yeah. Um Oh, there he is all the way over there at Bob's Steakhouse. Okay. Oh, all right. You know, I I'll get some steps in. You know, I haven't gone on the treddy yet today. So, I'll let you in first then. There you go. My challenge. Robo taxi. That's cool. All right. So, close all doors. Fasten seat belts.
Does Does it know if I'm going to like fasten or not? Let's Yes, I think it does, actually. Okay. You can try not to, but I don't think it'll go. Oh, wow. That was fast. That's intense. Cuz I I just hovered it there for a moment. So, I will let you choose which one you want to touch. Um, I want to see the latency.
That's pretty good. I That's That's impressive. All right. And so, this will be interesting cuz this was the one where it couldn't take me last time. Okay. So, well, I guess we'll see what happens. Yeah. Arrives. So, how hard is it just for giggles if we're like, you know what? We changed our mind. We want to go somewhere else.
By the way, I don't know what this is. Oh, look at this. Enter a destination. What? Okay. All right. So, let's try Oh, let's try this. That takes us back this way. Oh, and that would be right on the edge. Okay. All right. So, let's see what happens. Okay. All right. Uh, calculating. Okay. Okay. Well, you just taught me something new.
I not Okay. So, there's And then let's see how, you know, what the car does because obviously at this point we've got to flip a UI here or or it looks like it's going to make a right turn up here. But that's interesting because Yeah. So, this little um golden text, just tap that and then you can pick whatever location you want.
for much of this ride, most of which I filmed with my iPhone and then of course with the #notsponsored meta glasses. Hate them for the AI. Love them for the video and the pictures, the convenience of them. But anyway, for most of the ride, I things actually seemed pretty chill. This seemed and felt like a car that was on some form of enhanced version of hardware 4.
probably something like a hardware 4 asterisk or a hardware 4 o see the open AI borrow there or a hardware 5. Anyway, point is, it seemed very much like the comfortable FSD that I feel like now I've grown up with. From first using autopilot in 2017 in a Model X to getting to FSD on the various different Teslas and seeing it evolve and grow over time to going to endto-end neural nets and then now being in a robo taxi with, yes, a safety driver.
This felt very casual and frankly up until now the safety driver had to do nothing. And this is when we decided, well, let's try to give this app a little run for its money. And we changed destinations, which I then proceeded to do over and over and over again. Right before my last change of destination, we ended up hitting a little bit of a snag.
And this is where we made a big discovery about robo taxi. Oh, this is perfect entertainment. Oh, there was just a shooting a few days ago. Actually, read national news. Really? What got shot? Um, a four-year-old granddaughter, a grandpa, and a Target employee. Oh my gosh. Very sad. That is all right.
So, this is going to be interesting. Yeah, cuz I feel like probably strategically a Uturn here would be best. Uh, I think it's trying to follow this lead car. Yeah, it's following the car into the parking structure. Stop in. What does it stop? He just had to intervene. Yeah, it said stop in lane. It was going to follow the other car.
So, it still has the bias of a follow car. That's interesting. Yeah. How's your robot taxi? Hello. Your vehicles making uh or having troubles making progress on it? Um, I we just tried to route to Ziki and it seems to be uh confused. All right. Yeah, it looks like they have a police blockade right there.
Uh-huh. Correct. Let's put you Oh, he's moving it. Uh, maybe not. He just Yeah, it's still blockade. Still a blockade. Although he is sort of moving the car. I don't know exactly what's going on. We have accomplished the drama. I feel like I always get Tesla support on the line. What is going on? It's you.
It literally is. I'm like, jeez. And then these people, we got They're giving us the looks like, "What's your problem?" At least they can get through. They're fine. Yeah, they're fine. They're totally fine. They don't have to give us the angry look. Yeah. The whale that picked me up parked behind a car that was parked at McDonald's.
Just parked right behind it. And I was like, that's kind of a dick move. Yeah. Okay. So, so now connected to support remotely trying to figure out what do we do? Yeah. So, that tends to a U-turn. Well, we're in a three-point turn. That's Yeah, right. Back it up. Go in. I wonder what this is about. The blockade.
Gunshots. Our Austin serial killer that doesn't exist. I don't know if we've heard about that lore. Okay, so here we go. Okay. All right. That sounds fancy. Good to support. And now this guy. Oh my god, there's a bus behind us. Oh, wow. It's even doing it. I would suspect this is tea operated at this point.
That bing bing. Yeah, I that's my guess, which is fine. Um and this is this is a definitely an edge case. Yeah, but that follow car I think was something to observe. Yeah. Well, and I wonder Okay, that's the funny thing is like, go ahead, pull me over. Yeah, please. I I'd love to get pulled over in a robo taxi.
Welcome to Austin. See, the longer you stay in the car, there's bound to be a little bit of drama. Yeah, just keep rening. But, I mean, okay, it kept us safe. It did. They did have to intervene. That's okay. But you know like this is not rolled out as as Whimo is right now. So it's like these are important things to point out.
Yeah. So it seems like the same bias that we actually have in our FSD versions of preferring a follow car or in other words preferencing or biasing to follow another vehicle assuming the other vehicle is going to be a tool to help get you out of a sticky situation is the same bias that exists here in FSD.
Now, usually this isn't a problem because cars generally just follow the navigation path, but in this case, we did have a situation where the robo taxi wanted to follow another car into a parking structure, which was the wrong decision in this case. So, I don't know if there's a waiting towards bias towards a follow car.
If you get in a sticky situation, there probably should be a different bias because in this case, we got stuck with tech support. Now, I have to say getting stuck with tech support was actually kind of cool because it seemed like we heard this sort of chime of FSD turning on for a teleoperator, uh, which I'm presuming was a teleoperator since our safety driver manually paused the ride.
Uh, and then the vehicle proceeded to make a very cautious sort of three-ish point turn to U-turn out of there, which is exactly what you would expect a tele operator to do because that's what a human would do. Okay, roads blocked, threepoint turn or U-turn and get out of here. So, this was very normal and very natural.
Now, what does this mean for Robo Taxi in general? And how did the rest of the ride go? Perfectly. Frankly, the ride itself was amazing. I didn't have to deal with an Uber driver trying to talk to me. I didn't have to deal with an Uber driver chewing popcorn loudly. I had that like at the same day that I was in Austin.
It was the most miserable experience because I wasn't sure how to politely tell them like not to eat cuz it's their car, but then also like could you chew less loudly, but then is that rude? Should I just suck it up for the next minute or 10, you know, 10-minute ride or whatever it was? It It was a painful situation for me.
I don't know. Those those were some pretty loud munching sounds and and you know, not not my favorite. Anyway, I'll put that behind me. But let's understand this is a car that essentially outside of a police blockade, which was a total fluke, performed really well. I felt comfortable in the car. I didn't have to talk to anyone other than the person that I wanted to talk to, which was the passenger I was riding with.
I didn't have to engage with uh anyone for, hey, I'm going to change the destination. I could do it all through the app. I could control the climate the way I wanted it. Everything is essentially functional. Well, with the exception of the internet access, which wasn't working, so we couldn't really stream YouTube or get up.
But that's okay. We get enough of that anyway. uh and so really very pleasant experience with robo taxi overall. Now the risk factor that we run into with robo taxing is obviously that these sort of edge case scenarios become commonplace because if edge case scenarios become commonplace then what happens is you end up with the perception that this is a novel technology but it is a risky technology that if there's some edge case scenario you're going to end up getting delayed you're going to miss your Pilates class.
you're going to miss your appointment, you're going to miss your work meeting, whatever. That is a real risk to the future of robo taxi. Also, another risk would be the delayed removal of the robo taxi safety driver because frankly the safety driver is expensive. And then of course another risk would be some form of delayedwide expansion.
But outside of these risks, because I do think they are risks, but I do think that at some point they will go away. I think we will have a wide expansion of robo taxi. I think we'll get rid of the safety driver and we'll probably get to purely teaoperated. I mean really all the safety driver did in our case was take a situation where it was obviously stuck.
It could have remotely said, "Hey, I'm stuck. Should I bias and follow this lead car or should I just stop?" And a teleoperator could have taken over at that point. you really didn't need somebody in the car. Like the safety driver already in my opinion feels redundant and not necessary. Like no offense to the individual who was with us.
This is just sort of broadly I feel like they're mostly redundant. We don't already need them at all which is great because that all comes into the valuation for the actual stuff. Now this is predicated on Tesla actually fixing how they bill. Right now, if you change destinations, it will only charge you for your last ride.
So, we were in the car for 36 minutes and only got build $4.17 because our last destination was pretty close to our final destination. Ride history. Oh, so it it only charges us 4.7. Wait, it Oh, that's a bug. They only charged us the final destination even though we rode for 34 minutes, which to put it Yeah.
To put it into perspective, I think my 15minute rides for Uber were $40 each way. Yes. Or no, Whimo. Whimo. That should have been $50. $40 each way. Yeah. What? That's crazy. So, this should have been actually closer to 80 if it were San Francisco prices. 100%. That's nuts. So, all right. Well, there it is.
You don't owe me anything. No, stop. I'll make sure you get to 420. Yeah. Yeah. This is where things get really interesting because something that you need to know about robo taxis is robo taxis will never have a valuation like an Uber because the reality is with a robo taxi the revenue isn't going to the Uber driver.
That money that's being made for the drive is going to Tesla which means the Tesla shareholders are the ones who benefit from the rides. It's kind of like hiring a fleet of terminators to go work for you and then you are reaping that return on investment. That is the big difference here between Tesla and Uber.
Tesla is not another Uber. Tesla is not just creating an app and a platform for other people to go make money on your platform. That's like Upwork. Tesla instead is creating an army of terminators to go do an Uber service and replace the Uber infrastructure. That is pretty remarkable. Now, we have to make assumptions when it comes to math here.
So, obviously, everybody's going to have a different opinion on assumptions, but I actually think the numbers are pretty optimistic. So, take a look at this. If a robo taxi does 1 million trips per day, you know, maybe we use 50,000 uh vehicles doing 20 trips per day. Uh I like to assume a revenue per minute of about $1.
However, I water that down to only 30 cents uh with uh and the reason I do is I assume a 30% utilization because I think 70% of the time you're going to be either empty lagging to get to your next sort of customer waiting for a customer order because you want cars available when people click, hey, I want a ride.
You want a car ready. So, that means there's going to be some waiting time. So, there's waiting time and there's empty leg time. Uh then you also have cleaning time and charging time. So in my opinion, you've got maybe 20 minutes of utilization of these vehicles per hour. Now with that said, I think that each robo taxi trip will probably be on average about 10 minutes.
I think Uber trips are a little on average longer because they're not geoence to cities. And maybe in the future, e as we have a broad expansion, I could take a, you know, a Tesla robo taxi from LA to Vegas, but I think that'll be way down the road. I think at first we'll stay, we'll have like a broad release and we'll have robo taxis in San Francisco and LA and Chicago and New York, but they'll all individually be in their little radi, so to speak, so they can always go home and charge and they have sort of their little base.
That will change in the longer term, but anyway, that's where I'm going with 10 minutes. If you want to change the assumption, change the assumption. Okay. So then with a safety monitor, let's not even talk about this. The costs are just like really really bad if you have a safety driver, which I think is mostly redundant.
So I I don't even like talking about it because we don't even need them. But it's extremely expensive to pay somebody probably, you know, 20 bucks an hour in a wage plus workers comp plus uh the liability insuranceances and all the nonsense and the training and the paid time off and all the headache that goes with having employees.
You're probably frankly looking at 30 to 40 bucks an hour for that person plus the tea operated staff plus all the oopsies and incidentals and crap. I really think you're probably with safety monitors just purely losing money. Now, we're doing it right now because we know that if there's one oopsies with the robo taxi, then the stock's going to plummet by, you know, 10 years worth of safety monitor costs.
Elon knows that. Everybody knows that, right? So, that's why there's a safety monitor. But to actually fundamentally understand this, you have to do so without the cost of the safety monitor, which really based on my experience already, I think is mostly redundant. the tele operators can do pretty much everything.
That said, if I do not consider cap X uh as in investing into the actual vehicles and maintenance and charging, although charging would be more of probably an opex as maintenance would be, but that's fine. The numbers aren't too terribly different. Just keep this in mind. Just purely thinking about robo taxi revenue minus staff costs.
So revenue at, you know, a buck a minute, 30% utilization, costs at about 10 cents for the teleaoperated staff per minute because they can handle multiple at a time. They're not needed all the time. You get the benefits of AI and all the other nonsense. Uh, at 50 times price to earnings, you probably have with a million robo taxi rides per day, only about $10 of share impact in value.
That's nothing. That's not really worth it. You really have to get to the right columns for this to be valuable, which it gets very valuable very quickly. If you take off depreciation and some of the opex, I think then you'll be down at closer to $7.50. Not that much. Now, where you got to get to is about 10 million vehicles per day because that's when it starts looking or rides per day.
That's when you take the same numbers, you extrapolate it by 10, and you get about $75 per share in value for a robo taxi network. That's after, you know, depreciation and and some of these other expenses. That's actually pretty decent. Unfortunately, today, I think that the market is pricing in that we're going to replace pretty much all Uber rides at about 30 million rides per day, and this is in like 70 different countries.
So, I think the markets are probably over here. You know, you put about 100 bucks on the manufacturing side, and then maybe $220 for the robo taxi. Although you got to throw some Optimus in there, although that's, you know, that's that's many years out. You know, you could throw Optimus and insurance and semi-truckss and all that into sort of the future icing on the cake bucket.
But markets right now are probably over here on the right side because markets are a little just sort of euphoric and excited right now. So, you know, a lot of value is probably already priced in. But the point is, just forgetting about the share price for a moment, there's actually a lot of serious potential earnings per share that we could make with robo taxi.
And my bottom line is we don't need the safety driver. We could make 21 cents a share with a million rides a day, two bucks with 10 mil, 643 with 30 mil. Like we'll be somewhere in this range. It's going to make money as soon as we get rid of those safety drivers, which we will. Again, we don't I don't even think we need them right now.
Even with the edge case scenario I saw, at no point did I feel uncomfortable. I felt very happy and excited about this. This is actually better than I expected. So, congratulations to the Tesla team. This is very exciting. Frankly, all without liar. So, go Tesla, go Elon, go robo taxi. Have hands down better than expected.
Does that mean stock is a screaming buy right now at these valuations? Ultimately up to you. Anyway, thanks for watching. Check out House Hack, my real estate startup. And we'll see you in the next one. Very, very cool. Thank you so much. How do people follow you? Oh, they can find my YouTube channel, Ellie and Space, or ES TV on X.
I'm also on X. Awesome. Thank you so much. Thank you. And you know, I I hope you can ride in a Whimo, too. I I would recommend it. Thank you. .
Comments 0