Tesla (TSLA) Q3 2026 earnings call transcript

I am Travis Axelrod, the head of investment relations, and I am joined by Elon Musk, the executive president, and a number of other executive managers. Our results for the third quarter were announced at around 3 p. m. in the middle of America, and that was on the presentation that we published on the same link of this broadcast. During this conversation, we will discuss our expectations for our work and discuss some Thank you.

The third quarter was strong to the end, and I would like to congratulate Tesla's team for its complete efforts to achieve this success. The demand for our cars is increasing at a speed that is much higher than our production.

It is clear that we will increase production in a noticeable way, and we are indeed increasing it in a large way. It is the entire self-driving driving system that is the largest for the demand, and there is no real competition for it in any car available for purchase. and to clarify, we produce and deliver the largest number of cars and we strive to strive to obtain the full control of the self-driving system in additional markets. We have indeed obtained approval in many European countries and we hope to obtain approval at the level of the European Union.

It is difficult to predict that, but perhaps in the course of a few months, but I will be surprised if it is more than a quarter of the first. But it is clear that this is not a timeline we are controlling, but we expect to get full compliance in all parts of Europe and in many other countries.

We also started producing Tesla's Semi in the third quarter, and we did an experimental presentation for the new Roadster, which was, in my opinion, one of the most exciting events. This is a proof of the amazing work of the Tesla and SpaceX team. Therefore, we also put the foundations for a great increase in production energy. We aim to produce 5 million cars annually by the year 2033 approximately.

And I think we have a good opportunity to achieve this goal. Yes, and the demand for energy is very strong. We have agreed to sell all of Megaback's stocks by the year 2027. and we work to increase production as quickly as possible to meet the demand. We also expect a big increase in the overall profit.

In short, after years of recession and relative recession, I think we are now heading towards a great and exciting growth. In what concerns self-driving cars, we witness rapid growth in the number of cars cut or, in fact, the number of cars cut entirely.

And from The difficulty of exaggerating the importance of this point The growth in the number of self-driving cars is fast until a week after a week, and I don't think there is any company in the world capable of competing with the size of our production of self-driving vehicles as we continue to expand. For example, the fleet of CyberCab cars in Austin has grown ten times since its launch about a month ago. And the distance cut off is really amazing, because when I look at the weekly data drawing, I find it really amazing. It looks like a rocket running in its orbit, and it grows at a high speed. And if you want to experience it, I advise you to visit Austin and you will see the CyberCab cars used in a wide area there, as many have seen in the Austin City Limits event recently.

So This project will be a great success. What about For Optimus, we have started what I call early production on the high-size production line in Fremont. Now, I want to expand on this point because I I mentioned this before, Optimus's growth is similar to the development of the S-Mobile, but the first part of this S-Mobile is a small and irregular object because it is a new product in general. Therefore, as you know, we must think about the fact that there are about 10,000 workers who must work on what is needed to produce the robot, and the robot will move at a slower speed than these 10,000 robots.

But we do have it. The vast majority of production line is 1 million units of each production line in Fremont now. But I expect it to be exhausted in a year at least, and maybe a little more, to reach this annual rate of 20,000 units a week. Well, there is an essential matter because there are parts of the design that are still in the process of change.

Many people ask me about Optimus, and for this reason I think I should clarify this point. Because people, if they do not understand the new technology or the huge production of complex devices, may think that it is not simply possible to produce large quantities of it and send it immediately, but in reality, it is necessary to build a huge global supply chain and manufacturer.

Local Very large And to do everything efficiently And there are still some adjustments that we need in the design Because we want to provide a safe, reliable and long-term robot So that it does not take a month Well, there is still a lot of work to be done in the field of AI in the robot to become useful in the open tasks. But as you said, I expect the problems of AI, mechanical electricity design, reliability and life of use, i. e. the life of the robot, to be solved by the end of the next year.

So, this is the expected time To hand over Optimus to companies outside Tesla And I'm still convinced that Optimus will be the biggest product on earth, and so Robots are similar to humans in general So, if Optimus is a very important thing, but we must not lower the ceiling of expectations, we still have a lot to accomplish, and it will take about a year to complete all of that.

To a certain extent, there is no one who is even close to the level of Tesla in terms of a human robot with a variety of features. As for Lantarafab, the research plant in Austin is advancing rapidly.

We will start operating the machines in the first half of the next year. We hope things will go as planned, and we will make the first test sheet to see if we can make any sheet at all in May or June. This is exciting. to the extent.

Well, then if you are our intention, it is a kind of useful products in large quantities in the third quarter of the year two thousand and twenty-eight.

And as I mentioned before, I think that the Terafab factory will be the most expensive building in the world, and perhaps its economic value will be among the best 20 countries.

It is a huge project and will be a very important milestone if it is not necessary for the expansion, especially the Optimus project, where we can deal with the gaps in the transportation networks, whether in the closure of the memory or the logic or anything else.

As for the possible integration with SpaceX, I think I said everything I should have said yesterday. It was very complicated. Well, that's all. In the end, we are excited to go on our way in the field of self-driving and robotics. There are many great things to come, and I would like to thank Tesla again for their special performance and all our supporters for joining us on this journey.

Great.

Thank you very much, Elon. And FybHaf has an inspirational word as well. FybHaf.

Thank you, Rafis. We have reached our destination on many fronts in the third quarter. Elon has actually gone to many of them. In addition to that, I would like to add that we also started delivering MIGABAC batteries supplied with internal LFP batteries. For the first time, we registered 40 consecutive deliveries of more than 480,000 vehicles.

As for the car sector, we have completed the quarter with the highest number of requests we have in four years, where delivery operations were mainly controlled by production. Therefore, we focus on increasing production in all our factories to meet this increasing demand. While we are working to increase production, it takes some time to adjust each side of the supply chain to manufacture and logistics services for delivery. Model Y's car reached its peak during the third quarter, verifying the numbers of delivery in many major markets. Based on the initial data that we collected, we believe that it was the most sold electric car in Europe and the most sold passenger car in South Korea. We also achieved new standards in other major markets, including Australia, Japan and Taiwan.

As Elon Musk said, the FSD, the main engine for the demand, is still the full-fledged leadership technology. In the third quarter, more than 75% of the delivery operations in North America were subscribed to FSD when delivered, which is nearly 20% of the second quarter. In general, the FSD usage rate continued to improve to more than 1.7 million customers worldwide, of which 50%, and for the first time, they depend on subscriptions. In view of the FSD power in the United States, we take additional steps to sell specific cars in Europe with some advantages in an attempt to increase its sales before obtaining a wider agreement on FSD.

As for the profits in the car sector, with the exception of the dependencies, it has risen in a consistent way from 16.3% to 18.1%. This improves the profits to increase the ratio of deliveries in the Asian region and the Pacific. While the prices of essential goods and the changes in profit prices continue to increase our costs and create additional challenges, we have adjusted the car financing programs to partially compensate for these costs.

As for the energy sector, we achieved the second best quarter of the year in terms of distributing 13.7 gigawatts an hour of energy storage. The overall profit of energy has improved from 20.4% to 33.1%. This improvement is mainly due to the adoption of the national drawings that were recorded during the quarter and the guarantee of the agreement with one of the suppliers about some of the costs related to the cells that will be compensated by them. As Elon mentioned, the sector witnesses a huge demand from data centers, and we are making a small effort to increase production. Thus, we will witness an improvement in the average selling price and more improvement in profit margins over time.

Service margins and other margins have witnessed a continuous improvement of 14.1% to 15.9%, which is the highest level at all. This improves the performance of services in many business sectors, including high-charge and security services. The car industry has reached the leadership level, which is growing among other services and areas, with the launch of the CyberCab service in Austin. The total amount of cut-off goods reached eight times during the quarter, and we achieved a weekly growth of two times the quarter. We benefit from all the useful lessons from our expansion in Austin to help us prepare to launch CyberCab on a wider range in other markets. Our priorities are to increase the density and expand the network through vehicles that provide the best possible experience, such as CyberCab.

We will also add other types of vehicles to our fleet as time passes to meet the needs of land transport for all residents. I would like to confirm that we are ready to control all aspects of land transportation. Our goal is not to shorten the transportation of individuals, but rather to ensure that the entire transportation system is available worldwide We are in a huge investment stage to our future Therefore as the announcement previously stated the operating profits continue to rise continuously Most of this rise is due to research and development activities. and the cost of developing information systems, such as the consumption of additional computing assets, and the development of technologies, and the cost of pre-production of new products such as Optimus and Megapack 3 and the Semi charger.

This will continue to increase in operating profits until we introduce these new products in the markets. The increase in the cost of income affected the decrease in the value of our stock in SpaceX of 142 million dollars, which was compensated with a profit of the price of The market price of Bitcoin is close to $290 million.

We have made our efforts to manage the market with efficiency and responsibility. We have completed a quarter of the free cash flow, which adds up to $1.5 billion, and the cash and investment revenue, which adds up to $45 billion. We review our expectations for the capital inflows to reach about $20 billion this year. This is a result of our effective management of the capital distribution, and we plan to fund it mainly from our operating cash flows.

In addition to that, we achieved during the quarter the possibility of borrowing up to $30 billion to support the expected growth in capital inflows. We have increased our confidence in our business ability to generate cash and investments that we are keen to secure short-term funding. We do not plan to withdraw any excess from any of the new easements during this year.

And the reinstatement of the money we received from the collection of the national drawings to expand the production energy in the United States.

I would like to confirm that we are making great and direct efforts to operate the production facilities. So, the capital inflows will continue to expand over the next few years with the addition of more self-driving car lines, expanding our production energy, building a solar power plant, and building a computer-based infrastructure and a factory for the transportation of the Thank you, Faibhaf.

Now we move to questions from the site say. com.

Ashouk, would you like to answer this question?

Okay, I'll start from here. First, we launched Cybertruck this quarter, which is an exciting matter to the extent. We have achieved this through learning mainly from the data acquired from other vehicle platforms. For example, Model Y, Model 3, and so on. We will be able to expand the program in Dallas, and then expand the program itself to include from other cities side by side with Model Y robot car.

Great. Thank you very much.

We have talked about this matter, but But using external resources is not the problem for Optimus. There are many things that we must adjust to in terms of building artificial intelligence, and we are still working on it. We are making good progress, but this needs a solution. In order for Optimus to get the necessary training data, we need to actually create a large number of Optimus robots. So we need to create about 10 or 20,000 robots, then make them play independently. And we are working on creating Optimus Academy, where Optimus can train on the performance of a very wide group of tasks. Then we can use these data in a continuous improvement process to improve its performance in these different tasks.

The big difference between Optimus and cars is that we do not have a huge fleet of cars to collect training data from. So with Optimus, we have to start by making robots. Play with it, which is the way humans learn to interact with the world. As you know, you can notice that when you have a small child, and when he grows up, he is at first hesitant and does not know how to do things, and in the end, as you know, he learns how to clean himself by himself, and it is a great achievement.

As any person has children, then they learn to read and do different kinds of things. and they learn all this through playing and we actually need to play Optimus on a wide range so that he can be a good right in performing his duties in the real world and Shukin, do you want to add anything to that?

Yes, as I mentioned Elon, for the car we have data from its internal design but for Optimus there is no external data that can be reached by humans and turned into a doll or something similar يجب أن يتعلم أوبتيموس من خلال مشاهدة الآخرين وهم يؤدون المهام في البداية ثم يبدأ بممارسة المهمة بنفسه. على سبيل المثال تخيل أنك تشاهد مباراة كرة قدم فمجرد مشاهدتك لها لا يعني أنك ستصبح لعبا ماهرا. عليك أن تبدأ بالتدرب وتتقن اللعبة وتتعلم من تجربة الروبوت. ونحن بصدد إنشاء دورة تعلم معزز ضخمة تبدأ بتدريب أساسي من البشر ثم نطبقه على أوبتيموس ونجعله يتدرب بنفسه كثيرا. whether in the real world or in the real world, to learn from its internal design. I think we are very close to achieving excellent results in this field.

Yes, I think we may be ready to release a video for Optimus in about a month. I want to make sure that this video, when it is released, is not just an experimental show ready, as you know. Some similar robot offers seem exciting to the eye, but they do not use any kind of advanced or integrated AI. I mean, they are just predefined tasks. If you try to use this robot in the real world, you will face many problems. It will not work easily. So I want to release a video showing Optimus doing other things. I think this will take a month or two approximately to present to you an experimental presentation that you can believe what you see.

Then there is another important achievement we want to achieve, which is the integration of the Tesla AI5 artificial intelligence chip in Optimus, which, if things go according will be in April of next year, approximately. This is a huge improvement in capabilities, as it allows for a larger number of operations in seconds. In other words, AI represents five more computing powers, ten times more than AI-4. This will help Optimus a lot, because the degrees of freedom that human The human robot is more likely to deal with it than the car. Therefore, integrating everything required to work with all levels of freedom is a difficult thing. And with this wide range of tasks, you can imagine that the human robot will do a thousand times the things that the car does.

This is a huge number. Therefore, I think that the additional computing power of AI5 will make a big difference. and the expected to start producing it in large quantities in April of the next year. For this reason, I believe that all elements, such as improvements in the electromechanical power supply and the assurance of the ability of Optimus to operate for long periods without eating, and the perfection of the AI program and the integration of a 5A chip, are all factors that complement each other. For this reason, I think that after about a year, the product will be ready to be shipped to customers and will also reach the production stage in a wide range. So, maybe this is the right time to think about it.

As for the early customers of Optimus, I think they will use it at the beginning in industrial applications, such as The dangerous and repetitive tasks Well, things people don't want to do So, this is what you want to accomplish over the hour I mean, things you do over the hour This will be the highest value for Optimus at the beginning With the increase in production, of course, it will become something people can buy As you know, for their homes But the primary use of the high value, I think, will be in heavy industries and dangerous tasks or things that may cause repeated injuries, and this is what you need to accomplish over time. So this is the way to think about it.

Great. Thank you.

Ashouk, do you want to start this?

Yes, as you mentioned earlier, the Cybertruck and Model Y program is the same, and we will be able to soon release the same program as part of the V15 version for customers. So, the only additional condition is to make sure that the Model Y car, especially for the client, for example, can be part of the system that determines the geographical locations and the management of the fleet that Tesla provides to be fully self-driving.

We expect to be able to operate some of the first customer cars to be part of the fleet of mobile car prices in the second quarter of the next year.

Great. Thank you, Ashu.

Larsi, do you want to start this?

Certainly. Thank you, Rafis. I think this completes a little of Ashok's article for the TOWI. But in short, we are ready to provide Cybertruck cars to customers by the end of this year, 2026. It is clear that we produce them in our production line in Texas, but they are all currently used in the fleet of our private car dealerships. and with the growth of Intisala, we will focus on delivering self-driving vehicles designed entirely for our customers. And CyberCab will be the first of these vehicles. Therefore, with the start of our project on a small area this year, and our acquisition of more knowledge about everything related to delivery, ownership and management of self-driving vehicles for customers, we will continue to expand that area.

Elon, I do not know if you have anything to add to that.

No, I mean that CyberCab will basically serve as a transportation service. While our other cars will be more, as you know, the cars that people want to own like the 3 and what not, and the Roadster car of course, but as you know, the majority of the services of CyperCab are owned by Tesla and will be provided as a transportation service.

Excellent. Thank you.

So, Fiti, do you want to dive into the issue of the economy of the unit?

Yes, as far as premium cars are concerned, we are actually witnessing an improvement in the use in the main geographical areas that we work in. As Elon said, we look a week after a week at the amount and number of paid trips. Have we reached our goal of performing in 30 years today? No, but do we see a way to get there? Yes, certainly. And it's all about expanding the scope of operations. And what you will see in the coming few months is that we will launch the CyberCab service, which is the main product of self-driving luxury cars in Austin only.

But we plan to launch it in other cities before the end of this year. In addition to more CyberCab cars, we also learn from the experience of operating luxury car fleet, transportation, cleaning, and others. So we work to meet all these costs. And for this reason, as you mentioned earlier, we see a clear path to our goal of 30 years.

Then, do you want to cover the cost of Optimus?

Yes, I think we follow the cost of Optimus internally in a good way in terms of design development. But from the past, for the sake of saying that we have reached our goal, because we, as Elon said, still make many adjustments to design. The good thing is, as you know, when you look at the matter from the perspective of cost, we are approaching the ultimate cost expected to serve electric car prices.

Great. Thank you very much, Fibhaf.

Yes, yes, you are completely right. We should not rely on external maps. Therefore, we build our own maps, but not in the same way as building other maps. We build a new building that is based on artificial intelligence in drawing maps and mapping. It will be great to the extent that it will be combined with voice control. You can talk to it and ask for different routes. It will learn your personal preferences in addition to the preferences of the car fleet. will be a great help to all cars in making maps and traffic in a great way. We will start and you will see that the beginning and end of the trips use this technology very soon as part of the V15.

And in the coming few months, all the tasks of traffic and traffic jams and dealing with the directions will be taken care of, and it will introduce you in a better way than you know yourself in some areas related to the mission of driving.

I will take full responsibility. Of course, we cannot produce unless we have at least one operation or component. Five Half has mentioned that we run all production areas effectively. As you know, and as we talked about in our previous conversations, we have made global strategic investments to increase our current factory's ratio over the past year. This quarter we started to see that profit, and we expect this to continue in the next quarter with the start of the growth of these investments. In addition, we are also working to increase the production of Cybertruck trucks in the We have submitted requests for a required license to expand the factories in Berlin. As Elon also mentioned, we have started to build a new factory for the shipping of the similar factory to Renault.

In short, yes, we are building new factories.

Good. And we aim to achieve the goal of 5 million by 2030, as Elon mentioned. Great. Thank you very much.

This is all about the questions of Say. com. Now we will move on to the questions of the analysts. To remind you, as Elon said, we will not answer any questions today about a possible integration with SpaceX.

I will leave Ashoka to explain in detail, but it is essentially a very complex problem. Yes, it is easy to reach the first 90% and create a clear and attractive display, but But when you delve into the details of this problem, you will find it very long to the extent that people do not realize that any sane person does not need an attempt to solve it. It is complex to the extent and long term. Especially to reach a level of security that allows safe leadership of millions of people without injuries. It will be very difficult. And with our launch of the FSD, the FSD, which is not subject to the supervision of our clients, I think that no one will be able to compete for many years.

And there is another thing I would like to say, if it is not clear, and that is when the clients come to our exhibition hall here in the United States, they ask us, can we get a FSD-powered car? They don't ask for a 3-way or a Cybertruck car, but they ask us, can we get a FSD-powered car? And this has become the main reason, because it allows greater flexibility for different types of people who used to face difficulty in driving.

Now they are ready to use the FSD-powered car. And many people see it as a kind of transfer Meaning if you haven't tried it And I always say it If you tried it You won't be able to drive any other car I understand you, my friend I drive everywhere What?

It's hard, but as you know, for some cars, it was able to operate the FSD self-driving system. Everyone who tried it there realized how much it surpassed the competitors. In reality, perhaps you did not know that. Some executive leaders of competing companies visited the United States and tried to drive cars supplied with an FSD system and were amazed at their capabilities. Therefore, the answer is similar. And again, we do not deserve this. We work actively with the organizational parties in China because we are We previously issued a safety report for the FSD system, as well as a safety report for self-driving car.

We believe that this is the best way to provide a safe environment, because the number of accidents that occur due to the lack of attention of drivers is large.

I mean, I think Elon talked about this a little at the beginning, as you know By using the US growth accounts, this will become an important part of our profits and earnings in 2027 Yes I would say that this quarter will not make a big difference, but it may start to affect the appearance in the second quarter, then it becomes of great importance by the third or fourth quarter of the next year.

Good. If you think about it from the size of the ship, I think the area of the production line of the CyberCab cars has more than 125,000 units a year. Production began recently.

How do you expect to be the increase in production on the CyberCab line?

I'll start with this.

And I will leave Lars Yudli with his opinion as well But I would like to invite the engineering team They designed this product because I don't only talk to engineers But I also talk to the manufacturing team They believe that this product is one of the easiest products that can expand the production area So it will not be a big challenge I believe that the main thing that we are working hard to ensure Thank you very much.

Well, I think time is still early in this matter. We don't know the full extent of how his development I think we will understand that better in the first quarter or the second quarter but it has great capabilities but it is difficult to determine the size of these capabilities now Good, understood and we continue to talk about self-driving cars It seems that you have reached a turning point in Austin last month, almost since the launch of the CyperCab car.

I think there are some programming challenges that we are still working on. As Ashouk mentioned, maybe his words are correct. For example, we started the service of the car from 6 a. m. to 9 p. m. Then we widened it to include 10 p. m. and 11 p. m. And now the service will be available soon until 2 a. m. So we gradually expand the service of car cars.

And this depends on the capabilities of artificial intelligence software. We just want to make sure that no security accident happens. This is the principle of our work. and even if there is a minimum risk, we do not want to endure it. So we are very careful to avoid dangers in this field, O Hassan. We want no injuries or injuries to happen. This is our main principle. And we will see during the rest of this year that we will do the service throughout the hour and move to high-speed roads. And we believe that all this is necessary to expand and provide a great product to a large number of users. And for this reason, we wanted to start with the comprehensive expansion through work in multiple cities.

We have started with six cities so far. We also work on the deep expansion by operating a large number of vehicles in one city, which is Austin. And we will expand to include different weather conditions, night and day conditions, fast roads and high speeds. And we hope that all this will be done by the end of this year. And then we will witness a larger expansion and the coverage of more cities with self-driving cars. So, it's not an organizational search or something like that. It's a technology push, but we see that it's quickly turning into an organizational, maybe in the first quarter or the second.

Well, we have to determine. The relative distribution between SpaceX and Tesla is related to who pays what in Tiravap. But spending on the main equipment will not be less than tens of billions even. Early year 2028, and after that, the amount will rise significantly, and perhaps exceed 100 billion dollars in the long term.

Yes, I would like to add that we have already submitted most of the orders for the purchase of the first set of equipment for the factory that Tesla is building.

For this reason, when Elon spoke about the middle of next year, it was from the first factory, which is the research factory. Yes, we are currently determining how to operate the equipment, but the actual production will start in the late years of 2028, which is a record time for the factories of the same type.

Yes, it represents the biggest part of the difficulty. I have Constantine, so why don't you ask this question?

In reality, the hand is one of the most challenging areas of the device. Most The electromechanical system, as I say, has high relative confidence, meaning that it has turned into something that works well.

The hand, from the electromechanical point of view, is more complex than the engines of the movement system, for example. Also, there are some designed lessons that are a big disadvantage in terms of consumption And the reliability that we are still working on developing Yes, this is the most important thing about the winner by hand Specifically By adjusting its design accurately, especially the fingers, it is a difficult problem because humans have the ability to fix themselves, our hands fix themselves and our bodies fix themselves, but robots do not do that. So you have to have our fingers soft, kind of, and contain a lot of sensations in our skin. It is difficult to make a hand that can withstand a lot of use and wear, but it is at the time itself in a way that is enough to capture the things accurately and contains all the senses and does not require the senses to perform continuous movements.

There are many small details that make you realize in the end how great the human hand is. Well, it is mostly related to the electromechanical side, i. e. the hands. There are other aspects that improve the robot as well, but this is the most important.

As you mentioned, there is a program for the embedded robot intelligence using artificial intelligence 5, which is a type transfer in intelligence that will greatly enhance the capabilities of Optimus. As you said before, robots similar to humans are the largest product on earth. and its benefit will be almost double the value of the electromagnetic and artificial intelligence software and artificial intelligence chip.

All these three aspects are witnessing rapid improvement. I know that we use this word a lot, but the value, as you know, is almost the result of hitting these three things together, some of them, then Optimus robots will be built, a few of them by building Optimus robots in the future, so you have a repetitive evidence. Therefore, when you have a repetitive, double-edged evidence, things will grow at a tremendous speed. It may start from a small base, but it will grow to imaginary levels, and I think that at some point there will be a number of robots similar to humans on Earth above the number of humans.

Yes, I mean, remember when you compare us, you compare us with the service of riding that is available to drivers. As for us, no. In addition to that, we make the car ourselves. Therefore, the basis of the cost is different for us, and I think I mentioned that in the word of the introduction, and I will repeat it. I don't want to repeat myself too much, but our main goal is to control all types of transportation.

I mean, I think we will do what governments have not stopped, and it is a possible possibility. For example, if they were trying to protect a local industry, I think Tesla will eventually take over most of the transportation of passengers around the world in any country that allows them to work there freely. A time. And you can come back and look at this expectation. But I feel confident about it. It is really useful.

Yes, we will continue to expand in the future. We do not see any obstacles in the way. It is clear that there is a required work, but we see a clear path towards expanding to millions of vehicles during the coming few months. And this is mainly due to our technical investments over the past years. and also to V15 group that we referred to in a previous time of this year. Well, all this is hard to see, it is like a mountain. You do not see except its top. There is a huge effort in it, and now we are starting to see the fruit of this effort from the development of integrated neural networks and learning from humans and so on.

All this gives its fruit. This is what you see as a result of many years of work.

Yes. I think the numbers I shared are correct. And with the expansion of the service area, we will start sharing more of the standards related to the entire self-government system, not only the wills, but also the other main standards.

The number of drivers who stop subscribing Yes, as you mentioned, we will share more measurements about the stoppage rates from subscribing and what is similar But the reason for not doing so at this stage is our busyness with many things, especially with the addition of new countries. Then we provide free experiences, and I would like to emphasize an important matter when I say in North America, it is more than 75% of the cars that we give are free of charge for a month in the FSD full self-driving system. And yet, people still participate, despite the availability of this free subscription.

It seems that there is a lot of what he can do in relation to the role of training and developing some I believe there is a lot of things that can be done about the role of training and developing some of the and we work to improve all aspects of the supply chain and production system.

So we do all these things in parallel, like the huge production line in our factory in Fremont, and the huge supply chain, in addition to the adjustment. Design and engineering simultaneously With artificial intelligence software All these things happen at one time And we work at the fastest possible speed I don't think anyone can do it Optimus competition I would be surprised if one robot found it competing even if in a light way When we start to hand it over to the clients which is what is likely to happen at the end of next year.

Yes, and we must learn that. Not only the Fremont factory, but we also started building a factory here in Austin, which will be able to produce a larger number of robots compared to the Fremont factory. So, back to what Elon said, we are making our efforts short. We will not stop. We are just moving at the fastest possible speed.

Yes, we are certainly, as they say in Las Vegas, looking forward to the future.

It has no great impact at the moment. I mean, in reality, our capital spending was a little less than we expected.

But it will become more important next year and will reach tens of billions in 2028. We are moving according to the plan.