We're currently making it on the order of thirty five thousand autonomous vehicles a week, competed to say way most entire fleet if they have less than a thousand corners thirty five k a week. The fsty is actually getting so good that it takes us a while to actually find mistakes. I feel confidence saying that we have most advanced human right robot by longshot, and we're moreover the only company that really has all of the ingredients necessary
to scale human right robots. We have a kind of ridiculous demand for the semi roaster is not diet, not just the icing on the cake.
It's the cherry on the icing on the cake. We think that we'll be.
Able to have drival as Tesla's during paid riots next year. One of the clock companies a world class that a team like Tesla does zero. If you look at eed companies worldwide, it's the best my knowledge. No eb company is even profitable, and it's the best my knowledge. There's no easy division of any company of any existing order company that profitable. So it is notable that Tela is
profitable despite very challenging one about the environment. And this quarter actually is a rec world Q three for US. So we produced to our seven million vehicle actually just yesterday, So congratulations to the teams that made it happen in Tesla. That's the straggling, the events, amount of work to make seven million cars.
So to see.
We also have enterry storage business is growing like wildfire with strong demand.
Will both make it.
Back and powerbol and as people know, on October tenth, we laid out a vision for an autonomous and looked for future that I think is very compelling something you know that the Tesla team did a phenomenal job there with actually giving people an opportunity to experience the future where you have humanoid robots walking among the crowd, not can video presentation or anything, but picially walking among the crowd, so many drinks and whatnot. And we had fifty autonomous vehicles.
There were no twenty cyber caps, but there were an additional thirty model wise operating fully autonomously the entire night and carrying thousands who was no incident the entire night for those who went there that it were empsizing that the Cybercap had no steering wheel or break or accelerated pedals. I mean there was no way for anyone to intervene manually even if they wanted to, and the whole night
went very smoothly. So regarding the vehicle business, we are still contracted to deliver affordable models starting in the first half of twenty twenty five.
But this is I think probably people are wanting to who what should they assume for vehicle sales growth next year, And at the risk of to take a bit of risk here, I do want to give some rough estimate, which is I think twenty to thirty percent vehicle growth next year, notwithstanding negative external events like if there's some force masure events like some big war breaks out or interest rates to go sky high or something like that,
then we can't overcome massive force masure events. But I think with other cost hicles with the advent of autonomy, something like a twenty to thirty percent growth next year is my best guess.
And then Cybercamp reaching buying production in twenty six. Do you feel confident of cyber reaching buying production twenty six A starting production reaching buling production in twenty six, that should be substantial. Both we're aiming for at least two million units a year of Cybercamp. That will be in more than one factory, but I think it's at least two million units a year, maybe four million ultimately. So yeah, these are just my best guesses. But if you ask
me my best guesses, those are my best guesses. The twenty six to sell forty six to eighty lines, the team is actually doing great work there. For forty six to eighty is rapidly approaching the point where it is
the most competitive sell. I think that the fully landed the cost of a battery pack fully landed in the US net of incentives and duties, the forty six eighty is tracking to be the most competitive lower costic than what are fully considered than any other alternative is We're not quite there yet, but we're close to being there, which I think is extremely exciting, and we've got several
a lot of ideas to go well beyond that. So if I think this, if we execute well, the forty six eight will have the the internally produced cell will be the most cost competitive sell in to the North America, the testament to a tremendous amount of hard work there by the team. So that's say, we will continue to buy a lot of seals from our competitors, so ours not to make to provide to make cells just eternally, so I don't want to set off any law mails here.
We're alsously increasing substantially our big output and our stationary storage output, so we need a lot of cells, and most of them will still come from suppliers. But I think it is some good news that the Tesla internal cell is likely tray needs to be the.
Most competitive in the users.
So with respect to autonomy, as people are experiencing in the cars, really from week to week, there are significant improvements in the miles between interventions. So with the new version twelve poind five, the release of fullstof driving and cyber truck combining the code into a single stack. The city driving and the engine and highway driving are one stack, which is a big improvement for the highway driving. So it's just all neural nets and the release of actually
smart summon. We try to have a sense of humor
here at test. That's twelve point five. Version thirteen of FSD is going out soon show we will elaborate more on that setator in the coal we expect to see SUT roughly a five or six fold improvement in miles between interventions compared to twelve point five, and actually looking at the year's whole, the improvement in miles between the interventions, we think we'll bring it at least creors bang to so that's a very dramatic improvement in the course of
a year. We expect that to continue next year. The characters internal expectation or the chessea FSD having longer miles between inventional than human is the second quarter of next year, which means it made up being the third quarter, but it's it's next It seems extremely likely to be next year.
Do you want to well.
Yeah, interestioning miles between critical interventions like you mentioned it on we already made that one hundred next improvement twelve or five from running of this year, and then with the thirteen release, we expect to be one thousand next from the beginning from January of this year on my
production release software. And this came in because of uge improvements going to end to end having higher frame rate, partly also helped by Harvard Force more capabilities, and we hope that like Q two next year we should cross over the average human might spe critical intervention corision that case, I mean that is just our internal estimate. Yes, aternal inestimate is Q two next year to be safe than human and then to continue with rapid improvements thereafter.
The vast majority of humanity has no idea that doesn't ye themselves. So, especially for something like a Model three or Model ye.
It looks like a normal car, so you don't expect normal car to be able to be intelligent enough to drive itself. If cybocab looks to fix different, cyber truck looks different, but bottle wine at all? Three are They're good looking cards but look look fairly normal. You don't expect a very normal looking car to have the intelligence enough AI to be able to drive itself, but it does.
So we do want to expose that to more people.
Every time we have a significant improvement in the software, we'll roll out another sort of third day trial to encourage people to try it again. And we are seeing significant improvement and adoption. It will take great for FST is improved substantially, especially after the ten ten evment. Yeah, so if there's no need to wait for robot taxi or stab acab for to experience the full autonomy, we expect to achieve that next year with our existing big line.
The car ab dedrive itself to the user within paired parking lots. Currently it is speed limited, but then it's kind of quickly we increase. We already had more than a million US which we have Fortesla employees in the Bay Area. You're right healing your body, so you can actually use with the development app. You can request a ride and it'll take you anywhere in the Bay Area. We do have a safety driver for now, but a
software required to do that we've developed. David, do you want to collaborate on the Yeah, sure, Stavid.
We shared different screenshots of this in the Key one shareholder deck, but this is real. We've been testing it for the part of the year. The building blocks that we needed to build this functionality in delivery and production
we've been thinking about working on for years. It just so happened is that we've used those building blocks to deliver great features for our customers in the meantime, such as sharing your profile, synchronizing and across cars so that every single car that you jump into, whether it's another car that you own or some a car that somebody's loaned to you, or a rental car that you jump into,
it looks exactly like yours. Everything synchronized, seat mirror positions, media navigation, everything is the same, just what you would expect from one of our robotaxis. We gave that functionality to our customers right now because we've built it intending for it to be used.
In the future.
Releasing that functionality now all the cybersecurity that we knew we were going to need to deliver that functionality. Sending a navigation to destination from your phone to the vehicle. You're doing that now with the right Hailing app, but it's something that we've made available customers for years. Seeing the progress on a route in the mobile app. That's
something you'll need for the right Haling app. But again, we released it in the meantime, so it's not like we're just starting to think about this enough right now while we're building out the early stages of our right Haling network. We've been thinking about this for quaitely a long time, and we're excited to get the function out.
There, and we do expect to roll out right Heiling in California and Texas next year to the public. The California is somewhere that's quite a long regulatory approval process, it shouldn't. I think we should get approval next year,
but it's continued upon a regulatory approval. Texas is a lot faster, so I'd say we'll definitely have available in Texas and probably haven't available in California subject torctigulatory approval, and maybe some other states actually next year as well, but at least California and Texas, so I think that'd be very exciting. But it's really a profound change. The Tesli becomes more than a sort of vehicle and a
free main battering company. At that point, we published Q three Bibicle Safety Report which shows one faster for every seven miles of water pilot that compressed the US average water crash roughly every seven hundred thousand miles. So it's currently showing a tanx safety improvement relative to the US average. And we continue to expand our a training capacity to accommodate the needs of both the FSTY and optimists. We're
currently not a training compute constraint. It's probably the big slimity factors that the EPSTY is actually getting so good that it takes us a while to actually find mistakes when you start getting to where it can take ten thousand miles to find a it's a mistake.
It takes a while to actually figure out which it is.
Is this well, it's a software bolt better than the software bolt, a bet than software bolet B. It actually takes a while to figure it out because neither one of them will making mistakes. It takes a long time to make mistakes. How long does it take us to figure out which version is better? Sort of high class problem. Obviously having a giant fleet is very helpful for breaking
this out. I think with Optimists we start of a mass of improvement in Optimist to exterity movement on or two a tenth and our next gen handed form, which is twenty two degrees of freedom double the prior handed worong it's extremely human like, and will say it's much
better attacked ale sensing. I feel confiden saying that we have the most advanced human right robot by long shot, and we're moreover the only company that really has all of the ingredients necessary to scale human right robots, because the things that that what other companies missing is that they're missing the AI brain, they're missing the blot really scale to very high volume production, so you sometimes see
some impressive video demos. But what they like the they like the localized AI and the polite to scale volume to very high nverse. As I've said on a few occasions before, I think optimists will ultimately be most it's a good chance of being the most valuable product ever
made for the energy business that's doing extremely well. The lad Through megapack factory reached two hundred megapacks a week, which is now a forty gigle what hour year, and we have a second factory in Shanghai that will begin with the twenty gig what our year run rates Q one next year so just the next quarter, and that'll pop also scale up. It won't be long before we're shipping one hundred gig white hours a year stationary storage, that's all, and that will ultimately grow, I think to
multiple turor whatte hours per years. It has to actually in order to have a sustainable energy future. If you're not at the terraboid scale, you're you're not really moving the needle.
So if if you look.
At our very complicated last master plan, which I think actually is too much, email, maybe least to analyze it and shorten enough yet give us the tail VR on the last master plan. But we showed that mass plan that it is possible to take all of OOSE to a fully sustainable energy situation using sustainable energy power generation in batteries and electric transport. And there are no fundamental material limitations, like there's not some very rare material that
we don't have enough blog on it. We actually have enough with the raw materials to take all of human civilization make it fully sustainable. And even if civilization dramatically increased its trust, the usage could still be fully sustainable. You know, one way to think of the progress of a civilization is percentage completion of cutdoship scale Cardashi of scale one would be you're using all the power of a planet. We're currently less than one percent on cudship
level one. Level two would be you're using all the power of the sun. Now level three all the power of the galaxy long ago. When you think of crdoship terms, it becomes obvious that the biggest source of energy is the sun. Everything else is in the noise, So same conclusion. Tella is focused on playing the future of energy, transport, robotics in the ANI and this is a time when
others are just focused on managing around neartim trends. We think what we're doing is the right approach and if we execute on our objective, I think we will tell us that my predictions will become most valuable company in the world, probably long by long shot. I want to thank that tells the team once again for strong execution in a tough operating environment and looking forward to building an incredibly exciting future.
Thank you great, Thank you very much, Elan, and about personal bigger works as well.
Thanks our Q three results while we're on positive and once again demonstrate the skill which businesses involved with the generation of record operating cashlows of six month three billion. Automotive revenues grew both quarter and quarter a year early year while we had unit While in growth, we did experience reduction in asps, primarily due to the impact of
financing incentives. As a reminder, we're providing these incentives primarily using third party banks and financial institution and recognize the cost of these incentives as an affront reduction. Together, we released FSD for cyberchuk another features like actually Smart someone like PLA.
Talked about in North America, which.
Contributed three hundred and twenty six million of revenues in the quarter. We continue to see anybody levels of regular story credit sales with about two.
Billion of revenues so far.
This China continues to outperform US in Europe factor of three. This gives a signal of what is to come in other regions as customers acceptance of EV grows. A focus remains on growing unit volume while avoiding a build up of To support this strategy, we're continuing to offer extremely
compelling vehicle financing options in everyone. When you compare any vehicle in our lineup with other OEMs, believe our vehicles provide much better value, particularly when you consider the safety features, performance, and unparalleled software functionalities like David also talked about, include also what show had talked about around autonomy, music options,
parental controls, and much more. While every vehicle in our lineup comes up with these capabilities, there is an awareness cap not just with buyers, but at times even with existing owners. We plan on making these more visible in
our interactions with both existing and future customers. Automotive margins improved quarter or quarter as a result of equity feature release, discuss reform increase in our overall production and delivery varning, hoping it benefit from the marketing pricing and more localized deliveries in region which resulted in lower freight and duties.
Sustaining these margins in Q four, however, will be challenging given the current economic environment, so that we are focused on the cost per vehicle and there are numerous workstreams within the company to squeeze our costs without compromising on customer experience.
Yeah.
Something that's helpful, hopefully helpful macro trend is if there's a decline in interest rates, this has a massive effect on automotive vast majority of people is the demand is driven by the muscle payment. Connect with muscle payment, so most life well see continue to decline interest rates which helps with portability phables.
Yeah, I mean that is one trend which we observed in the industry that you know, because of the fudability being impacted because of interest rates, people are holding on to the cars longer, especially in the US, and that is actually having an impact on our industrytive. While we did see a decline in Q three, we expect to go deployment sequentially in Q four to end the year more than double the of last year thirtie margins in Q three for our record at more than thirty percent.
This is a function of mix up projects being deployed in the quarter. Note that there will be fluctuation in margins as we manage through deployments and our inventory. Our pipeline impact law continue to grow quarter work quarter as we fill our twenty twenty five production slots, and we're doing our best to keep with the demand. Just coming back on automotive margins, I talked about what is happening.
One other thing which I want to also share is that we will continue to keep whatever we can to see. Like I said before about squeezing out the costs, but this is something which we also are very capable of. I mean, just in Q three we really reach our lowest cost per wide and that is a trend which we want to focus on. Then going onto service another
we continue to show improvements in Q three. This was the result of better performance both in our service business which includes collision and parts, sales and merchandise, and continued growth and super charging his fleet based revenues will continue to grow as the lower of fleet size increases, our operating expenses decline quarter work quarter and an year on ear basin. This is partially due to the restructuring took in Q two. Cost saving from these initiatives were partially
offset by increasing costs related to our aifforts. We've started using the GPU cluster based out of our factory in US ahead of schedule, and I will track to get fifty k gbus deployed in Texas by the end of this month. One thing which I'd like to elaborate is that we're being very judicious on our AI compute span too, and saying how best we can utilize the existing infrastructure before making partner investments. On the capex front, we had
about three and a half billion in the quarter. This was a sequential increase margin because of investments in AI compute. We now expect the CAPEX for the year to be in excess of eleven billion. We shared our vision for the future at the view of what event at the beginning of the month. The TESTLA team is hyper focused on delivering on that version. All efforts are underway to
make it a reality. While we've achieved significant progress this year, it will take time to get this as we finally new and incredibly complex technologies and navigate a fragranted regulatory landscape. Future is incredibly bright and I want to tag the de'st looting once again called great.
Thank you very much for up. Now we'll go to investor questions. The first one is is Tesla still on track to deliver the more affordable model next year as mentioned by elan earlier? And how does it align with your AI product?
Right?
MAT permission has always been to lower the cost of the our vehicles to increase the adoption of sustainable energy and transport. Part of that is lowering the costs for current vehicles, which is where all of the personally.
Owned vehicles that we sell today come in.
But the next stage in that really is it fits into AI roommap is when we bring in romotax the specially as the initial cost of getting into an EV. And that's really where we see the marriage of EV roadmap and the ANIMI.
I'll be like with with incentives to thirty k, which is kind of a key threshold.
Right, thanks very much. A similar question next, when can we expect Tesla to give us the twenty five thousand dollars on regular taxi regular car model.
We're not making it on?
Yeah, all our vehicles today are rota. I think we've made very clear.
That future is autonomous. I mean it's going to be and I've actually said this many years ago, but that my strong believe and I believe that is panning out to be true and it will be very obvious in retrospect, is that the future is autonomous three vehicles and non autonomous gasoline vehicles in the future will be like riding a horse and using a fluff bone. It's not that there are no horses. Yeah, there are some horses, but they're unusual with their niche and you know, so, I
just everything's gonna be electric, autonomous. I think this is blind like it should be frankly, blindingly.
Obvious at this point that that is the future.
A lot of automotive companies, and most automotive companies have not not internalized this, which is surprising because we're shouting us from the rooftops for such a long time and it will accrue to their entrments in the future. But all of our vehicles in the future weed, yes, all the vehicles that we were really made. Obviously, millions of vehicles, the vast majority, are capable of autonomy. You know, we're currently making. It's on the order of thirty five thousand
autonomous vehicles a week. If you compare that to say, way most entire fleet if they have less than a thousand corners thirty five a week. Yeah, cars look normal and mostly normal. Cyber truck looks abnormal. In the cybercab SAT trug, we wanted to have something futuristic looking. It does look futuristic.
It's worth learnt.
Even with respect to the cybercab, it's not. It's especially not just a revolutionary vehicle design, but of revolution in vehicle manufacturing that is also coming with us with the cybercab. The cycle time, like the units per hour of the cybercab line is like, this is just really something special. I mean, listens will be half order of magas than other car manufacturing lines. Like not not even the same league, That's what I'm saying. Not in the same league. I said sart years ago.
And then maybe the.
Hottest part to comy will be the factory. Yeah, just like bio factory versus engineer or factory and it's empty mind. Yeah, so we're rapidly evolving acturing technology. So anyway, there's like basically I think having an irregular twenty five K models pointless, it would be silly, like it would be completely at us with what we believe.
In the time of this world. But matters as the lowest cost.
Per mile of the efficiency of that vehicle, and that's what we've done.
With Yeah, but the rope texting exactly autonomous, it's fully considered cost per mile is what matters. And if you try to make a car that is essentially a hybrid manual automatic cards, it's it's not going to be as good as a dedicated autonomous car. That's our cab is is just not going to have stereo wheels and bats
side otis or autonam. It'll cost on the order of cost roughly twenty five k. So it is a twenty five K car and you can you will be able to buy one on an exclusive exclusively if you want. So just what has steams there?
Neither?
Great, Thank you very much. The next question is what is Tesla doing to alleviate long way times at service centers?
So we aim on solving problems at the source, so at the factory before they can't even affect our customers. We believe the best services, no service don't even have them if the car doesn't break. Yeah, exactly, that's the best thing. Don't see anyone the test shirt. You either do it fixed the issue upstream where you would remotely do it through software, maybe be at.
Work or at home.
And you know, car we partner and we address fixed issue, and we've partnered the field with service to make.
Sure we're looking at the same issues.
And additionally, just in Q three and Q four this year long, we have opened and will open in total nearly seventy locations, and in North America we significantly expand the size of each location and have doubled the size last year compared to this year.
Yeah, I think it was like actually a lot of erics to have any large service set up because you can have specialization of labor, you can start, you can start your approach. Yeah, it should be more factory like where you can have dedicated lanes full particular types of service, and it's way easier for somebody to become expert in a few different types of repairs than in every repair. Exactly.
This has helped us with the base that these heavy repairs like cogging up the lane. They've dedicated lanes for different type of repairs and so it's through the matters and really treating it like a factory.
Yeah, this is this is where it tells.
The structure I think has a structural advantage relative to the rest of the order industry because we make the cars and reservice the cars, Whereas I think there's a bit of a conflict of interest with the dealer model and the traditional area and the dealer model where the dealerships make most of their money on service and so they don't have they obviously assist incented to reduce the servicing costume, whereas in our case we are incentive to
reduce the servicing costs because we we carry that servicede cost and we've got a good feedback without without cars.
Yeah, with the factory, with the service leaders together and send back people from the factors to the field and field to the factory to see it firsthand, I buy suggestions for you know, manufacturing as well as for engineering on design.
Yeah.
So I prove this as a structural fundamental structural advantage of tesla versa is the rest of the order indist to do a bunch.
Of work on the opperare side, only to automate diagnostics, so identifind what needs to be done to a car where it comes into service, but also automating all the preparation work and aligning all the resources that are necessary in order for the car and be very efficiently worked on once it arrived.
So the parts are there, like the.
Lift is schedule, the technician schedule, like everything like a card, this is what's wrong with me, and tell us fill the service center cart everything ready in advance.
Yeah, please tect me and this is what's wrong.
Yeah, instead of customer trying to translate the cars telling us directly and we're pulling that.
Yeah. Most of the time you don't need to diagnose the car when it arrives the car. Yeah, this is like again, have fun technology advantage, attructural advantage compared to the rest of the order industry.
It's I think it's under as to what all we are able to do. Yeah, and that's why because, like I said before, most of four cars except for cyber Drug look to say, right, so people don't realize that it has so much capability.
Yeah, that's better than other cars. Yeah, but they're not like obviously like super futuristic.
Great, thanks very much. The next question is please provide an update on the semi, what will the next stage of growth look like and when will f SD be ready?
Sure, we posted an earning that we're progressing softly on the build of the semi factory there in our data factory a reno. We've released all our meter cap co expenditures for that program and we're on track to start have the builds in the second half of next year, with production starting in the first half of twenty twenty six, in ramping really throughout the year to full production. Some you know, growth will largely depend on our customer's adoption of the product.
Well, I don't think we're going to be devided limited honestly.
What should I do say?
Which is like the brainer for the semi because it's really a commodity of total costs of.
Yes, exactly, it's we were kind of ridiculous demand for the semi in that world where it's about how much do I spend to exactines per mile? It's a new brainer. Fundamentally, if you've got a semi work with the fully considered cost pro mile or for a ton of transport is better than a diesel truck. Any company that doesn't adopt an electric semi, will will lose. It's not a subject of things we want the stat we want. We want
to have a beautiful semi drug. But frankly, if we made any semi drug would have matter.
And this is proving so in our fleets and in Pepsi's partner, in fact that PEPSI actually said last week their drivers don't want to go back.
As soon as we give anyone the electric semi. It's like that's the choice.
It's what they want to drive.
Yeah, yeah, yeah, that's like like so that like the most sceenor like if they're top drivers, will they get to drive. It's the thing they want to dry. It's super fun to drive. It's also very easy to drive.
It's easy to drive, and it holds fast and it's like fast maybe too fast, well, but i mean like you know, like you've seen like videos of world it has electric somehow can go uphill.
Yes, we can speeding past like the diesel trucking cars and cars so like it's responsive you floor it and the truck actually.
And that's a benefit not only for the driver and for the goods, but also for safety in terms of other drivers on the road. You don't get stuck behind the semi, You're not like, yeah, you know, in a slow down situation and on rad I mean how that plays into you know, FST which is the second part
of the question. All of the semis have been since a couple of hundred we've deployed already and the ones that will be building next year and throughout SUTURE have all of the hardware and the cameras necessary to deploy fifty And we're currently training with that small fan that we have and as soon as the fleet is trained and then they're on that's round, you know, we'll get FIST onto that platform.
Yeah, I mean, it'd be a massive improvement in dry or fatigue, you know, because and drivers safety. We've got sort of the anti Jack Knight thing software. You know, you don't have to worry about your breaks overheating if you go down there down the steep hill because that we use regenerating like that that energy goes back.
Into the pack.
Actually, when we leave ring on something, it's.
Like radically better than it is with someway. The drivers love it, guys, Thanks very much.
Our next question is when will Tesla incorporate X and ROCK in all the Tesla vehicles.
Well, I mean these are relatively small things, you know, but yeah, I think we'll keep expanding what is available in the car on the screen and also improving on the browsers, so you just generally you can access anything you want in the car. In fact, for that says
we'll just get to full autonomy. You actually want a system that you can do anything like if you want a browser Internet, if you want to ask AI questions, if you want to watch a movie, if you want to play a video game, if you want to do some productivity thing.
You can do anything you want. And you know, tom vehicle because you don't need to dry them.
So that's why the Cybercat got a nice picture and the great sound system, so you can what's a great movie with It's like being like an in personal movie theater, yeah, postal movie theater.
This is why we've been building this functionality and gaming to the car, adding MPs and other media applications of the car, because that's what you're gonna the cars will be built today.
Sometimes we find games by the way, picture people haven't tried it.
There's like Castle doing Bad and Platopia and a bunch of really fun games in the car.
We're constantly looking at what features to add next, and we're paying attention to what's most commonly rebuffed by our customers.
Yeah, playing Castle bombad Do you want to thank you guys very much?
The next question is Elon mentioned on supervised FSD in California and Texas next year. Does that mean regulators have agreed to it in the entire state for existing hardware three and four vehicles.
As I said earlier, California loves regulation, but they have a pathway. Yeah, I mean there's a pathway. Obviously Waymo operates in California. So there's a lot of forms and a lot of approvals that are required to be shocked if we don't get approved next year. But it's just not something we totally control. But I think we won't get approval next year in California and Texas, and for the end of the year it will branch out beyond California and Texas.
I think it's important to reiterate this mull union you're CERTI finding a vehicle out a federal level in the US is done by meeting FMVSS regulations. All our vehicles today that are produced that are autonomus capable meet all those regulations. The side of capital need build regulations and so the deployment of the vehicle to the road is no limitation, but is a limitation, is what you said at the state level where they control autonous vehicle deployment.
Some states are relatively easy, as you mentioned for Texas, and it's other ones have plays like California that may take a little longer. Other ones haven't set up anything yet, and so we will work for those state by state.
And I do think we should have a federal I agree that like autonomous vehicles should be approved, you should be it should be possible to condress.
If you're listening, let's get a federal av.
This should be a federal federal approval process for autonomous vehicles. I mean that's how the fmbsas it started and work Federal Motor Ahiric Club.
So I mean in twenty seventeen to eighteen, and you know, it's when regulators started looking at it, and it's really kind of stalled since then. But we would appreciate and would support helping out. It really needs to be that at like a national approval is important. You know, if there's an Apartment of Governor Efficiency try to help make that happen. And you took for everyone not just hasn't it.
But you know, something's in the US are state by state regulated, like for example, insurance. It' it's incredibly painful to do it state by state fifty states, and I think there should be a national approval process for autonomy.
Great guys. The next question is what is the plan for twenty twenty five.
I mean, basically we talked to this there's a lot going on. Multi mentioned that we're working on cheaper models to come out. I mean they're what Cosset the team is doing to get the factories ready today to try and make that happen done.
Yeah, to make a lower class car is insanely high. But like it is harder to get like twenty percent of the cost out of a car than it is to design the car and build the entire factory in the post place.
It's like excruciating.
And there's not a lot of movies made about the heroes who got twenty percent of the cost down of a car, but let me tell you there should be.
If we can got that is incredibly heroic. It's a little che isn't. It's not like a yeah, it's like there should be.
The heroes got twenty percent cost out of a car, is like, damn, I want respect to them. Like if you actually saw how hard, if people actually saw how hard was to do that, you'd be like, whoa, that's damn hard.
Just yesterday we were talking about potty.
I mean honestly, you're like literally yeah, I mean there's there's a lot of what I do call it sort of like getting cost out of things. It's kind of like it's like Game of Pennies. It's like Game of Thrones, but pennies. You know, first approximation, You've if you've got ten thousand nine ers in a car, very approximation, and each of them costs four dollars, then you have a
forty thousand dollars car. So then if you want to have a thirty five thousand dollars car, if you got to get fifty cents on average out of the ten thousand naders every time, every part, yeah, it's like yeah, and then obviously the best is you're delete some parts.
We're doun be able to lead a lot of parts.
I'm very I'm very excited about the cybercap design, how we're rethinking the design of a car for the cybercamp, designing it will alcohig volume production and then design a machine that builds the machine. That is I think revolutionary and that there's no other car company that's even trying to do what we're doing, Like I don't even heard of it. Actually, in fact I'm certain there's and when five five times better than additional factory like a.
Cycle time and like part deletion, material movement. I don't think any other car company has the same level of like integration of thought that we have when it comes to like when you design a part from the light shiet of paper, who's going to make it? Where is it going to be made, how's it going to be shipped, how's it going to be assembled into the vehicle, And like at any one point, if something is done in a silo, it becomes a bottleneck of either cost or time or efficiency.
But with the with the rod Taxi, the development, like we've done a good job on the combining all that and then like blowing up how it's made and saying it should be made this way and rethinking it.
Also it's the most efficient factory possible. That shows in our big wilship and our capex efficiency when we deploy it shows them in our parts, shows the simplicity and vehicle, but also have performance in terms of like end user state just on the energy front.
Also in twenty five, we will have Stardulent Fashion out the manufacturer Shanhai, will continue to increase our storage deployments. With all three, we plan to continue expanding our supercharging network, getting more wims on our network, some ramp as you want to talk about that will keep going. And then we also will have our lit refinery starting to produce.
Yeah, so many things.
I like crazy thing is like tell us winning basically out almost every single thing we're doing. If we're not winning, now we're one in an awareness with there our entire large companies that that's the only thing they do.
I mean it's a company.
There are multiple companies within the company.
Yeah, till those like many companies do one yeasic Thank you guys.
Just a few more.
What is going on the Tesla Roadster. Well, I'd just like to thank our long suffering deposit holders of the Tail Roadster. You know, the reason it hasn't come out yet is because it is that the roaster is not diet, not just the icing on the cake, it's the cherry
on the icing on the cake. And so you know, our large emission is to accelerate the progress towards a sustainable energy future, you know, try to do things that maximize the probably the future is good for humanity and for Earth, and so that necessarily means that like the things like that are kind of like dessert. We'd like, like we do all love to work on the tailor
next to in tesl roster. It is super fun and we are working on it, but it has to come behind them the more things, the things that have a more serious impact on the good of the world.
So just thank you to all lot long suffering tailor roaster deposit holders.
And we are actually finally making progress on that and we're close to finalizing the design on that. It's really going to be a big spectacular you know, frind Mine, Peter tail you know, and sometimes Peter til now right, we're really good friends. Peter, you know, was limiting how you know the future doesn't have flying cards. Well, we'll see before to come, I think very much.
The next one is quite similar to other questions we've had, so when I combine it with the final question, so briefly, could you just detail how ROBOTEXTI will roll out. We'll start with the tested deployed fleet and then allow customers to add THEIRS on the subscription model. And then will hardware three be capable of fix?
Because you're going the hardware three? What we solve it well not five?
Was it was easier to make progress with starting with Hardward four, then feeling on the solution, then backporting into
Hardbred three instead of directive working on Hardbred three. Given that Hardward four us more like fundamental hardware capabilities, I think that trend will continue into the next few quarters as well, by the first figure on the solution rapidly with four and then backport it trying the kind of it just takes longer to none of those things because it's not fundambly supported in the hardware and it's emulated. But yeah, it's initially working on Harvard for backboarding it to hundred three.
Yeah, it's actually answer we're not in harder sent sure, but I'm sure I mentioned by some measures hardware for it has several times the capability of Hardware three. It's it's easier to get things to work with Harvard for and then it takes a lot of effort to sort of squeeze that proximality into hardware three. There is some chance that hardware three does not achieve the safety level
that allows for untou advised FSD. If that turns out to be the case, we will upgrade those who have bought hardware three FSD for free, and we have designed the system to be upgradeable, you know, just to just sort of switch out the computer type thing. The cameras are okay, But we don't actually know the answer to that. But if it does turn out, we will make sure we take care of those who have board MSDY on Hardward three.
Right and in the last few minutes that we have left, we will try to get some analyst questions. The first question will be coming from here fairg and New Street.
Here.
Please feel free on me yourself.
I was wondering about the compute you're ramping up. You gave interesting statistics on how much you have, and you said you don't feel your compute constraint. I was wondering how you are putting to work this additional compute. Is that a game for you? Of creating larger and larger models, like next generation of models that are larger the way open air I go from GPT three to GPT four. Or is that more like you're set on your model and you need to throw more and more compute to
accelerate the pace of learning to improve reliability. And then I had a quick flow pretty quick on you rollout in Texas and in California next year. The plan as you see today, is it to roll out like a fleet or two with cars that will start with like a super visory, like some super a vision someone sitting at the wheel just in gaze and removing these super advisors posssively or are you aiming for going full fledged without even a human super super advisor when you get started.
Well, I guess regard to answer the question, the nature of real world AI is different from say ALAM in that you have a massive amount of context so that like case it tells the summary cameras are to night if you include internal camera that that that so you've got to tagabytes of context and and that that is then distilled down into a small number of control outputs. You know, whereas it's it's like you don't really it's very rare to have. In fact, I'm not sure any
ALM out there do gigabytes of context. And then you've got to you've got to the process that in the car with a very small amount of compute power. So you know it's it's it's all doable and it's happening. But it is a different problem than than what's a Gemini or open is doing.
Now.
Part of the way you can make up for the fact that the inferenced computer is quite small is by spending a lot of effort on training. Just like a human, the more you train on something, the less mental workload it takes when you do it. When the human starts driving, it absorbs your whole mind. But then as you train more and more on driving, get very good and the driving becomes a background task, it only absorbs a small amount of your mental capacity because of you have a
lot of training. So we can make up for the fact that the inferenced computer is tiny compared to a ten kilo or bank of gvus because we've got a few hundred watts of in furnst computer. We can make up that with heavy training. There's also vast amounts of data coming in and then's sorting out what training is important. With the vast amounts or video data coming in, what
is actually the most important for training. That's quite difficult, But as I said, we're not currently training compute construct.
You want to elaborate the.
Training has more train larger models also a train quicker, but in the end we've still got to pick which models are performing better. So the validation effort to picking the models customer sparent dimension is pretty large. Were to tive a lot of miles to go in close toop. We do have simulation and other ways to get those metrics, and those two help, but in the end that's a big bottleneck. That's why we're not training complete constraint alone.
And there's other access of scating as well, which is a data figuring offic straight as more useful than that is an important as we're focusing on that.
Yeah, so as it really is.
The second part of your question here about safety drivers and rolling it out. Each state has different requirements that you know in terms of how many miles and how much time you need to have a safety driver. And now I have a safety driver, we're going to follow all those were not if regulations you're out there with safety as a priority. But well was obviously already and safety is there will over.
From the right jet.
We think that we'll be able to have drive all those test those doing paid rides next year. Sometimes makes sure all thank you and our next question comes from Adam Jonas at Morgan Stanley. Adam, please feel free to unnet yourself.
Okay, thanks, everybody, just had a question about the relationship between Tesla and Xai. Many investors are still not clear how the work at Xai is truly beneficial to Tesla. Some even take the view that the two companies may be in competition with each other in terms of talent and tech and even your time, elon. So what's your message to investors on that relationship between Tesla and Xai and where do you see it going over time?
Thanks?
Well, I should say that you know Xai has been helpful to Tesla AI quite a few times in terms of things like scaling up training just even recently in the last week or so. It's in training or where if you're doing a big training one note fails be.
Able to continue training. But there are different problems.
X is working on artificial eral intelligence or official superintelligence it tells us trying to make autonomous cars and autonomous robots.
There are different problems.
Also, like all not all e is equal, right, I mean there's a is a broad spectrum.
Yeah, and we have our own swim lanes.
Yeah, there are certain things which we can collaborate on if needed, But for the most part we're sloving different issues.
Ye tells us focused on real well data and that is saying it's it is quite a bit different from an element that you have. You have massive context in the more video and some amount of video that's going to be distilled. Like with extremely efficient in furns compute, I do think tells us the most efficient in the world in terms of inference compute. Like because out of necessity we have to we have to be very good at in efficient in ferns. We can't pretend to kill
ons GPUs in a car. We've got a couple of hundred ones. You know, it's pretty well designed, tells the AI chow, but it's still a cup hundred one. But there are different problems. I mean, it's you know, like the s not that XAI is like we're just running en Burns. I mean it's running inference like answering questions, answering Tesla questions on a ten.
Killer what rank? It's like, ye't put that in a car. It's a different problem. No, exactly, XAI is.
Because because I felt there wasn't a truth seeking digital super intelligence company out there. That's what it came down to that they needed to be a truth seecret like an AI company that is very rugorous about a truthful And so I'm not saying XI is perfect, but that is that that is at least the franspiration, even if something is politically corrected, should still be truthful. I think this is very important for a safety. I think XI has been helpful to Tesla and will continue to be
helped to Tesla. But there are very different problems. Great, you know what other co company has a world class trip design team?
Zero? What are the cock companies a world class that a team like Tesla does? Zero? Those all startups? The creator for scratch.
Great, thank you a lot, and I think that's unfortunately all the time that we have for today, and we appreciate all of your questions and we look forward to eating squarter. Thank you very much, and goodbye.