Latest / Elon Musk Thinking / Latest Interview of Elon Musk, Talking About Tesla!!!

Transcript

Latest Interview of Elon Musk, Talking About Tesla!!! · Elon Musk Thinking

0:00
Speaker 1: We're currently making it on the order of thirty five
0:02
Speaker 1: thousand autonomous vehicles a week, competed to say way most
0:06
Speaker 1: entire fleet if they have less than a thousand corners
0:08
Speaker 1: thirty five k a week. The fsty is actually getting
0:11
Speaker 1: so good that it takes us a while to actually
0:14
Speaker 1: find mistakes. I feel confidence saying that we have most
0:18
Speaker 1: advanced human right robot by longshot, and we're moreover the
0:22
Speaker 1: only company that really has all of the ingredients necessary
0:24
Speaker 1: to scale human right robots. We have a kind of
0:27
Speaker 1: ridiculous demand for the semi roaster is not diet, not
0:31
Speaker 1: just the icing on the cake.
0:32
Speaker 2: It's the cherry on the icing on the cake. We
0:35
Speaker 2: think that we'll be.
0:36
Speaker 1: Able to have drival as Tesla's during paid riots next year.
0:41
Speaker 1: One of the clock companies a world class that a
0:43
Speaker 1: team like Tesla does zero. If you look at eed
0:48
Speaker 1: companies worldwide, it's the best my knowledge. No eb company
0:53
Speaker 1: is even profitable, and it's the best my knowledge. There's
0:55
Speaker 1: no easy division of any company of any existing order
0:59
Speaker 1: company that profitable. So it is notable that Tela is
1:03
Speaker 1: profitable despite very challenging one about the environment. And this
1:07
Speaker 1: quarter actually is a rec world Q three for US.
1:10
Speaker 1: So we produced to our seven million vehicle actually just yesterday,
1:14
Speaker 1: So congratulations to the teams that made it happen in Tesla.
1:17
Speaker 1: That's the straggling, the events, amount of work to make
1:20
Speaker 1: seven million cars.
1:21
Speaker 3: So to see.
1:23
Speaker 1: We also have enterry storage business is growing like wildfire
1:28
Speaker 1: with strong demand.
1:29
Speaker 2: Will both make it.
1:29
Speaker 1: Back and powerbol and as people know, on October tenth,
1:33
Speaker 1: we laid out a vision for an autonomous and looked
1:36
Speaker 1: for future that I think is very compelling something you
1:40
Speaker 1: know that the Tesla team did a phenomenal job there
1:42
Speaker 1: with actually giving people an opportunity to experience the future
1:45
Speaker 1: where you have humanoid robots walking among the crowd, not
1:49
Speaker 1: can video presentation or anything, but picially walking among the crowd,
1:53
Speaker 1: so many drinks and whatnot. And we had fifty autonomous vehicles.
1:56
Speaker 1: There were no twenty cyber caps, but there were an
1:58
Speaker 1: additional thirty model wise operating fully autonomously the entire night
2:03
Speaker 1: and carrying thousands who was no incident the entire night
2:06
Speaker 1: for those who went there that it were empsizing that
2:08
Speaker 1: the Cybercap had no steering wheel or break or accelerated pedals.
2:14
Speaker 1: I mean there was no way for anyone to intervene
2:17
Speaker 1: manually even if they wanted to, and the whole night
2:19
Speaker 1: went very smoothly. So regarding the vehicle business, we are
2:22
Speaker 1: still contracted to deliver affordable models starting in the first
2:26
Speaker 1: half of twenty twenty five.
2:28
Speaker 2: But this is I think probably people are wanting to
2:30
Speaker 2: who what should they assume for vehicle sales growth next year,
2:34
Speaker 2: And at the risk of to take a bit of
2:37
Speaker 2: risk here, I do want to give some rough estimate,
2:40
Speaker 2: which is I think twenty to thirty percent vehicle growth
2:43
Speaker 2: next year, notwithstanding negative external events like if there's some
2:47
Speaker 2: force masure events like some big war breaks out or
2:52
Speaker 2: interest rates to go sky high or something like that,
2:54
Speaker 2: then we can't overcome massive force masure events. But I
2:58
Speaker 2: think with other cost hicles with the advent of autonomy,
3:04
Speaker 2: something like a twenty to thirty percent growth next year
3:06
Speaker 2: is my best guess.
3:08
Speaker 1: And then Cybercamp reaching buying production in twenty six. Do
3:12
Speaker 1: you feel confident of cyber reaching buying production twenty six
3:14
Speaker 1: A starting production reaching buling production in twenty six, that
3:17
Speaker 1: should be substantial. Both we're aiming for at least two
3:20
Speaker 1: million units a year of Cybercamp. That will be in
3:25
Speaker 1: more than one factory, but I think it's at least
3:26
Speaker 1: two million units a year, maybe four million ultimately. So yeah,
3:30
Speaker 1: these are just my best guesses. But if you ask
3:32
Speaker 1: me my best guesses, those are my best guesses. The
3:34
Speaker 1: twenty six to sell forty six to eighty lines, the
3:37
Speaker 1: team is actually doing great work there. For forty six
3:39
Speaker 1: to eighty is rapidly approaching the point where it is
3:43
Speaker 1: the most competitive sell. I think that the fully landed
3:47
Speaker 1: the cost of a battery pack fully landed in the
3:51
Speaker 1: US net of incentives and duties, the forty six eighty
3:57
Speaker 1: is tracking to be the most competitive lower costic than
4:00
Speaker 1: what are fully considered than any other alternative is We're
4:04
Speaker 1: not quite there yet, but we're close to being there,
4:07
Speaker 1: which I think is extremely exciting, and we've got several
4:11
Speaker 1: a lot of ideas to go well beyond that. So
4:15
Speaker 1: if I think this, if we execute well, the forty
4:19
Speaker 1: six eight will have the the internally produced cell will
4:22
Speaker 1: be the most cost competitive sell in to the North America,
4:26
Speaker 1: the testament to a tremendous amount of hard work there
4:28
Speaker 1: by the team. So that's say, we will continue to
4:31
Speaker 1: buy a lot of seals from our competitors, so ours
4:33
Speaker 1: not to make to provide to make cells just eternally,
4:39
Speaker 1: so I don't want to set off any law mails here.
4:41
Speaker 1: We're alsously increasing substantially our big output and our stationary
4:48
Speaker 1: storage output, so we need a lot of cells, and
4:51
Speaker 1: most of them will still come from suppliers. But I
4:54
Speaker 1: think it is some good news that the Tesla internal
4:58
Speaker 1: cell is likely tray needs to be the.
5:00
Speaker 2: Most competitive in the users.
5:02
Speaker 1: So with respect to autonomy, as people are experiencing in
5:06
Speaker 1: the cars, really from week to week, there are significant
5:10
Speaker 1: improvements in the miles between interventions. So with the new
5:16
Speaker 1: version twelve poind five, the release of fullstof driving and
5:19
Speaker 1: cyber truck combining the code into a single stack. The
5:24
Speaker 1: city driving and the engine and highway driving are one stack,
5:27
Speaker 1: which is a big improvement for the highway driving. So
5:30
Speaker 1: it's just all neural nets and the release of actually
5:33
Speaker 1: smart summon. We try to have a sense of humor
5:35
Speaker 1: here at test. That's twelve point five. Version thirteen of
5:39
Speaker 1: FSD is going out soon show we will elaborate more
5:43
Speaker 1: on that setator in the coal we expect to see
5:46
Speaker 1: SUT roughly a five or six fold improvement in miles
5:49
Speaker 1: between interventions compared to twelve point five, and actually looking
5:52
Speaker 1: at the year's whole, the improvement in miles between the interventions,
5:58
Speaker 1: we think we'll bring it at least creors bang to
6:01
Speaker 1: so that's a very dramatic improvement in the course of
6:05
Speaker 1: a year. We expect that to continue next year. The
6:08
Speaker 1: characters internal expectation or the chessea FSD having longer miles
6:14
Speaker 1: between inventional than human is the second quarter of next year,
6:19
Speaker 1: which means it made up being the third quarter, but
6:22
Speaker 1: it's it's next It seems extremely likely to be next year.
6:25
Speaker 2: Do you want to well.
6:27
Speaker 4: Yeah, interestioning miles between critical interventions like you mentioned it
6:32
Speaker 4: on we already made that one hundred next improvement twelve
6:36
Speaker 4: or five from running of this year, and then with
6:38
Speaker 4: the thirteen release, we expect to be one thousand next
6:41
Speaker 4: from the beginning from January of this year on my
6:43
Speaker 4: production release software. And this came in because of uge
6:47
Speaker 4: improvements going to end to end having higher frame rate,
6:51
Speaker 4: partly also helped by Harvard Force more capabilities, and we
6:55
Speaker 4: hope that like Q two next year we should cross
7:02
Speaker 4: over the average human might spe critical intervention corision that case,
7:07
Speaker 4: I mean that is just our internal estimate. Yes, aternal
7:10
Speaker 4: inestimate is Q two next year to be safe than
7:13
Speaker 4: human and then to continue with rapid improvements thereafter.
7:18
Speaker 2: The vast majority of humanity has no idea that doesn't
7:21
Speaker 2: ye themselves. So, especially for something like a Model three
7:24
Speaker 2: or Model ye.
7:25
Speaker 1: It looks like a normal car, so you don't expect
7:27
Speaker 1: normal car to be able to be intelligent enough to
7:30
Speaker 1: drive itself. If cybocab looks to fix different, cyber truck
7:33
Speaker 1: looks different, but bottle wine at all? Three are They're
7:37
Speaker 1: good looking cards but look look fairly normal. You don't
7:40
Speaker 1: expect a very normal looking car to have the intelligence
7:43
Speaker 1: enough AI to be able to drive itself, but it does.
7:47
Speaker 2: So we do want to expose that to more people.
7:50
Speaker 1: Every time we have a significant improvement in the software,
7:53
Speaker 1: we'll roll out another sort of third day trial to
7:56
Speaker 1: encourage people to try it again. And we are seeing
8:00
Speaker 1: significant improvement and adoption. It will take great for FST
8:04
Speaker 1: is improved substantially, especially after the ten ten evment. Yeah,
8:08
Speaker 1: so if there's no need to wait for robot taxi
8:09
Speaker 1: or stab acab for to experience the full autonomy, we
8:12
Speaker 1: expect to achieve that next year with our existing big line.
8:15
Speaker 4: The car ab dedrive itself to the user within paired
8:18
Speaker 4: parking lots. Currently it is speed limited, but then it's
8:21
Speaker 4: kind of quickly we increase. We already had more than
8:25
Speaker 4: a million US which we have Fortesla employees in the
8:29
Speaker 4: Bay Area. You're right healing your body, so you can
8:33
Speaker 4: actually use with the development app. You can request a
8:36
Speaker 4: ride and it'll take you anywhere in the Bay Area.
8:38
Speaker 4: We do have a safety driver for now, but a
8:41
Speaker 4: software required to do that we've developed. David, do you
8:44
Speaker 4: want to collaborate on the Yeah, sure, Stavid.
8:47
Speaker 5: We shared different screenshots of this in the Key one
8:50
Speaker 5: shareholder deck, but this is real. We've been testing it
8:52
Speaker 5: for the part of the year. The building blocks that
8:54
Speaker 5: we needed to build this functionality in delivery and production
8:57
Speaker 5: we've been thinking about working on for years. It just
8:59
Speaker 5: so happened is that we've used those building blocks to
9:02
Speaker 5: deliver great features for our customers in the meantime, such
9:05
Speaker 5: as sharing your profile, synchronizing and across cars so that
9:08
Speaker 5: every single car that you jump into, whether it's another
9:11
Speaker 5: car that you own or some a car that somebody's
9:13
Speaker 5: loaned to you, or a rental car that you jump into,
9:15
Speaker 5: it looks exactly like yours. Everything synchronized, seat mirror positions,
9:19
Speaker 5: media navigation, everything is the same, just what you would
9:23
Speaker 5: expect from one of our robotaxis. We gave that functionality
9:27
Speaker 5: to our customers right now because we've built it intending
9:30
Speaker 5: for it to be used.
9:30
Speaker 2: In the future.
9:31
Speaker 5: Releasing that functionality now all the cybersecurity that we knew
9:36
Speaker 5: we were going to need to deliver that functionality. Sending
9:38
Speaker 5: a navigation to destination from your phone to the vehicle.
9:42
Speaker 5: You're doing that now with the right Hailing app, but
9:44
Speaker 5: it's something that we've made available customers for years. Seeing
9:48
Speaker 5: the progress on a route in the mobile app. That's
9:51
Speaker 5: something you'll need for the right Haling app. But again,
9:53
Speaker 5: we released it in the meantime, so it's not like
9:55
Speaker 5: we're just starting to think about this enough right now
9:57
Speaker 5: while we're building out the early stages of our right
9:59
Speaker 5: Haling network. We've been thinking about this for quaitely a
10:01
Speaker 5: long time, and we're excited to get the function out.
10:03
Speaker 1: There, and we do expect to roll out right Heiling
10:06
Speaker 1: in California and Texas next year to the public. The
10:10
Speaker 1: California is somewhere that's quite a long regulatory approval process,
10:15
Speaker 1: it shouldn't. I think we should get approval next year,
10:17
Speaker 1: but it's continued upon a regulatory approval. Texas is a
10:21
Speaker 1: lot faster, so I'd say we'll definitely have available in
10:23
Speaker 1: Texas and probably haven't available in California subject torctigulatory approval,
10:28
Speaker 1: and maybe some other states actually next year as well,
10:30
Speaker 1: but at least California and Texas, so I think that'd
10:32
Speaker 1: be very exciting. But it's really a profound change. The
10:35
Speaker 1: Tesli becomes more than a sort of vehicle and a
10:39
Speaker 1: free main battering company. At that point, we published Q
10:43
Speaker 1: three Bibicle Safety Report which shows one faster for every
10:47
Speaker 1: seven miles of water pilot that compressed the US average
10:50
Speaker 1: water crash roughly every seven hundred thousand miles. So it's
10:53
Speaker 1: currently showing a tanx safety improvement relative to the US average.
10:58
Speaker 1: And we continue to expand our a training capacity to
11:01
Speaker 1: accommodate the needs of both the FSTY and optimists. We're
11:04
Speaker 1: currently not a training compute constraint. It's probably the big
11:07
Speaker 1: slimity factors that the EPSTY is actually getting so good
11:10
Speaker 1: that it takes us a while to actually find mistakes
11:14
Speaker 1: when you start getting to where it can take ten
11:17
Speaker 1: thousand miles to find a it's a mistake.
11:21
Speaker 2: It takes a while to actually figure out which it is.
11:24
Speaker 1: Is this well, it's a software bolt better than the
11:27
Speaker 1: software bolt, a bet than software bolet B. It actually
11:29
Speaker 1: takes a while to figure it out because neither one
11:31
Speaker 1: of them will making mistakes. It takes a long time
11:34
Speaker 1: to make mistakes. How long does it take us to
11:36
Speaker 1: figure out which version is better? Sort of high class problem.
11:39
Speaker 1: Obviously having a giant fleet is very helpful for breaking
11:42
Speaker 1: this out. I think with Optimists we start of a
11:45
Speaker 1: mass of improvement in Optimist to exterity movement on or
11:48
Speaker 1: two a tenth and our next gen handed form, which
11:52
Speaker 1: is twenty two degrees of freedom double the prior handed
11:56
Speaker 1: worong it's extremely human like, and will say it's much
11:59
Speaker 1: better attacked ale sensing. I feel confiden saying that we
12:03
Speaker 1: have the most advanced human right robot by long shot,
12:07
Speaker 1: and we're moreover the only company that really has all
12:10
Speaker 1: of the ingredients necessary to scale human right robots, because
12:14
Speaker 1: the things that that what other companies missing is that
12:16
Speaker 1: they're missing the AI brain, they're missing the blot really
12:21
Speaker 1: scale to very high volume production, so you sometimes see
12:24
Speaker 1: some impressive video demos. But what they like the they
12:28
Speaker 1: like the localized AI and the polite to scale volume
12:31
Speaker 1: to very high nverse. As I've said on a few
12:34
Speaker 1: occasions before, I think optimists will ultimately be most it's
12:39
Speaker 1: a good chance of being the most valuable product ever
12:41
Speaker 1: made for the energy business that's doing extremely well. The
12:44
Speaker 1: lad Through megapack factory reached two hundred megapacks a week,
12:49
Speaker 1: which is now a forty gigle what hour year, and
12:52
Speaker 1: we have a second factory in Shanghai that will begin
12:56
Speaker 1: with the twenty gig what our year run rates Q
12:59
Speaker 1: one next year so just the next quarter, and that'll
13:01
Speaker 1: pop also scale up. It won't be long before we're
13:04
Speaker 1: shipping one hundred gig white hours a year stationary storage,
13:07
Speaker 1: that's all, and that will ultimately grow, I think to
13:11
Speaker 1: multiple turor whatte hours per years. It has to actually
13:13
Speaker 1: in order to have a sustainable energy future. If you're
13:16
Speaker 1: not at the terraboid scale, you're you're not really moving
13:18
Speaker 1: the needle.
13:19
Speaker 2: So if if you look.
13:20
Speaker 1: At our very complicated last master plan, which I think
13:24
Speaker 1: actually is too much, email, maybe least to analyze it
13:27
Speaker 1: and shorten enough yet give us the tail VR on
13:32
Speaker 1: the last master plan. But we showed that mass plan
13:35
Speaker 1: that it is possible to take all of OOSE to
13:39
Speaker 1: a fully sustainable energy situation using sustainable energy power generation
13:44
Speaker 1: in batteries and electric transport. And there are no fundamental
13:48
Speaker 1: material limitations, like there's not some very rare material that
13:53
Speaker 1: we don't have enough blog on it. We actually have
13:55
Speaker 1: enough with the raw materials to take all of human
13:58
Speaker 1: civilization make it fully sustainable. And even if civilization dramatically
14:03
Speaker 1: increased its trust, the usage could still be fully sustainable.
14:06
Speaker 1: You know, one way to think of the progress of
14:09
Speaker 1: a civilization is percentage completion of cutdoship scale Cardashi of
14:14
Speaker 1: scale one would be you're using all the power of
14:19
Speaker 1: a planet. We're currently less than one percent on cudship
14:23
Speaker 1: level one. Level two would be you're using all the
14:26
Speaker 1: power of the sun. Now level three all the power
14:29
Speaker 1: of the galaxy long ago. When you think of crdoship terms,
14:35
Speaker 1: it becomes obvious that the biggest source of energy is
14:38
Speaker 1: the sun. Everything else is in the noise, So same conclusion.
14:44
Speaker 1: Tella is focused on playing the future of energy, transport,
14:47
Speaker 1: robotics in the ANI and this is a time when
14:50
Speaker 1: others are just focused on managing around neartim trends. We
14:54
Speaker 1: think what we're doing is the right approach and if
14:58
Speaker 1: we execute on our objective, I think we will tell
15:02
Speaker 1: us that my predictions will become most valuable company in
15:04
Speaker 1: the world, probably long by long shot. I want to
15:08
Speaker 1: thank that tells the team once again for strong execution
15:12
Speaker 1: in a tough operating environment and looking forward to building
15:16
Speaker 1: an incredibly exciting future.
15:17
Speaker 6: Thank you great, Thank you very much, Elan, and about
15:20
Speaker 6: personal bigger works as well.
15:22
Speaker 7: Thanks our Q three results while we're on positive and
15:26
Speaker 7: once again demonstrate the skill which businesses involved with the
15:30
Speaker 7: generation of record operating cashlows of six month three billion.
15:33
Speaker 7: Automotive revenues grew both quarter and quarter a year early
15:38
Speaker 7: year while we had unit While in growth, we did
15:40
Speaker 7: experience reduction in asps, primarily due to the impact of
15:44
Speaker 7: financing incentives. As a reminder, we're providing these incentives primarily
15:49
Speaker 7: using third party banks and financial institution and recognize the
15:53
Speaker 7: cost of these incentives as an affront reduction. Together, we
15:57
Speaker 7: released FSD for cyberchuk another features like actually Smart someone
16:01
Speaker 7: like PLA.
16:01
Speaker 3: Talked about in North America, which.
16:04
Speaker 7: Contributed three hundred and twenty six million of revenues in
16:06
Speaker 7: the quarter. We continue to see anybody levels of regular
16:09
Speaker 7: story credit sales with about two.
16:11
Speaker 3: Billion of revenues so far.
16:13
Speaker 7: This China continues to outperform US in Europe factor of three.
16:19
Speaker 7: This gives a signal of what is to come in
16:21
Speaker 7: other regions as customers acceptance of EV grows. A focus
16:26
Speaker 7: remains on growing unit volume while avoiding a build up
16:29
Speaker 7: of To support this strategy, we're continuing to offer extremely
16:34
Speaker 7: compelling vehicle financing options in everyone. When you compare any
16:38
Speaker 7: vehicle in our lineup with other OEMs, believe our vehicles
16:43
Speaker 7: provide much better value, particularly when you consider the safety features, performance,
16:49
Speaker 7: and unparalleled software functionalities like David also talked about, include
16:55
Speaker 7: also what show had talked about around autonomy, music options,
17:00
Speaker 7: parental controls, and much more. While every vehicle in our
17:04
Speaker 7: lineup comes up with these capabilities, there is an awareness
17:08
Speaker 7: cap not just with buyers, but at times even with
17:11
Speaker 7: existing owners. We plan on making these more visible in
17:15
Speaker 7: our interactions with both existing and future customers. Automotive margins
17:20
Speaker 7: improved quarter or quarter as a result of equity feature release,
17:24
Speaker 7: discuss reform increase in our overall production and delivery varning,
17:29
Speaker 7: hoping it benefit from the marketing pricing and more localized
17:33
Speaker 7: deliveries in region which resulted in lower freight and duties.
17:37
Speaker 7: Sustaining these margins in Q four, however, will be challenging
17:41
Speaker 7: given the current economic environment, so that we are focused
17:46
Speaker 7: on the cost per vehicle and there are numerous workstreams
17:49
Speaker 7: within the company to squeeze our costs without compromising on
17:53
Speaker 7: customer experience.
17:54
Speaker 3: Yeah.
17:55
Speaker 1: Something that's helpful, hopefully helpful macro trend is if there's
17:59
Speaker 1: a decline in interest rates, this has a massive effect
18:04
Speaker 1: on automotive vast majority of people is the demand is
18:08
Speaker 1: driven by the muscle payment. Connect with muscle payment, so
18:12
Speaker 1: most life well see continue to decline interest rates which
18:15
Speaker 1: helps with portability phables.
18:17
Speaker 7: Yeah, I mean that is one trend which we observed
18:20
Speaker 7: in the industry that you know, because of the fudability
18:23
Speaker 7: being impacted because of interest rates, people are holding on
18:27
Speaker 7: to the cars longer, especially in the US, and that
18:31
Speaker 7: is actually having an impact on our industrytive. While we
18:34
Speaker 7: did see a decline in Q three, we expect to
18:37
Speaker 7: go deployment sequentially in Q four to end the year
18:41
Speaker 7: more than double the of last year thirtie margins in
18:44
Speaker 7: Q three for our record at more than thirty percent.
18:48
Speaker 7: This is a function of mix up projects being deployed
18:50
Speaker 7: in the quarter. Note that there will be fluctuation in
18:54
Speaker 7: margins as we manage through deployments and our inventory. Our
18:59
Speaker 7: pipeline impact law continue to grow quarter work quarter as
19:02
Speaker 7: we fill our twenty twenty five production slots, and we're
19:06
Speaker 7: doing our best to keep with the demand. Just coming
19:09
Speaker 7: back on automotive margins, I talked about what is happening.
19:13
Speaker 7: One other thing which I want to also share is
19:16
Speaker 7: that we will continue to keep whatever we can to see.
19:21
Speaker 7: Like I said before about squeezing out the costs, but
19:24
Speaker 7: this is something which we also are very capable of.
19:27
Speaker 7: I mean, just in Q three we really reach our
19:31
Speaker 7: lowest cost per wide and that is a trend which
19:33
Speaker 7: we want to focus on. Then going onto service another
19:36
Speaker 7: we continue to show improvements in Q three. This was
19:39
Speaker 7: the result of better performance both in our service business
19:42
Speaker 7: which includes collision and parts, sales and merchandise, and continued
19:46
Speaker 7: growth and super charging his fleet based revenues will continue
19:50
Speaker 7: to grow as the lower of fleet size increases, our
19:53
Speaker 7: operating expenses decline quarter work quarter and an year on
19:57
Speaker 7: ear basin. This is partially due to the restructuring took
20:00
Speaker 7: in Q two. Cost saving from these initiatives were partially
20:03
Speaker 7: offset by increasing costs related to our aifforts. We've started
20:07
Speaker 7: using the GPU cluster based out of our factory in
20:10
Speaker 7: US ahead of schedule, and I will track to get
20:14
Speaker 7: fifty k gbus deployed in Texas by the end of
20:16
Speaker 7: this month. One thing which I'd like to elaborate is
20:19
Speaker 7: that we're being very judicious on our AI compute span too,
20:24
Speaker 7: and saying how best we can utilize the existing infrastructure
20:27
Speaker 7: before making partner investments. On the capex front, we had
20:31
Speaker 7: about three and a half billion in the quarter. This
20:33
Speaker 7: was a sequential increase margin because of investments in AI compute.
20:38
Speaker 7: We now expect the CAPEX for the year to be
20:40
Speaker 7: in excess of eleven billion. We shared our vision for
20:43
Speaker 7: the future at the view of what event at the
20:45
Speaker 7: beginning of the month. The TESTLA team is hyper focused
20:48
Speaker 7: on delivering on that version. All efforts are underway to
20:51
Speaker 7: make it a reality. While we've achieved significant progress this year,
20:56
Speaker 7: it will take time to get this as we finally
20:58
Speaker 7: new and incredibly complex technologies and navigate a fragranted regulatory landscape.
21:05
Speaker 7: Future is incredibly bright and I want to tag the
21:08
Speaker 7: de'st looting once again called great.
21:10
Speaker 6: Thank you very much for up. Now we'll go to
21:13
Speaker 6: investor questions. The first one is is Tesla still on
21:17
Speaker 6: track to deliver the more affordable model next year as
21:19
Speaker 6: mentioned by elan earlier? And how does it align with
21:22
Speaker 6: your AI product?
21:23
Speaker 3: Right?
21:23
Speaker 8: MAT permission has always been to lower the cost of
21:26
Speaker 8: the our vehicles to increase the adoption of sustainable energy
21:29
Speaker 8: and transport. Part of that is lowering the costs for
21:32
Speaker 8: current vehicles, which is where all of the personally.
21:35
Speaker 3: Owned vehicles that we sell today come in.
21:37
Speaker 8: But the next stage in that really is it fits
21:39
Speaker 8: into AI roommap is when we bring in romotax the
21:41
Speaker 8: specially as the initial cost of getting into an EV.
21:44
Speaker 8: And that's really where we see the marriage of EV
21:47
Speaker 8: roadmap and the ANIMI.
21:49
Speaker 1: I'll be like with with incentives to thirty k, which
21:52
Speaker 1: is kind of a key threshold.
21:54
Speaker 6: Right, thanks very much. A similar question next, when can
21:59
Speaker 6: we expect Tesla to give us the twenty five thousand
22:01
Speaker 6: dollars on regular taxi regular car model.
22:04
Speaker 2: We're not making it on?
22:05
Speaker 3: Yeah, all our vehicles today are rota. I think we've
22:08
Speaker 3: made very clear.
22:08
Speaker 1: That future is autonomous. I mean it's going to be
22:12
Speaker 1: and I've actually said this many years ago, but that
22:16
Speaker 1: my strong believe and I believe that is panning out
22:18
Speaker 1: to be true and it will be very obvious in retrospect,
22:21
Speaker 1: is that the future is autonomous three vehicles and non
22:27
Speaker 1: autonomous gasoline vehicles in the future will be like riding
22:30
Speaker 1: a horse and using a fluff bone. It's not that
22:34
Speaker 1: there are no horses. Yeah, there are some horses, but
22:36
Speaker 1: they're unusual with their niche and you know, so, I
22:41
Speaker 1: just everything's gonna be electric, autonomous. I think this is
22:43
Speaker 1: blind like it should be frankly, blindingly.
22:46
Speaker 2: Obvious at this point that that is the future.
22:48
Speaker 1: A lot of automotive companies, and most automotive companies have
22:51
Speaker 1: not not internalized this, which is surprising because we're shouting
22:56
Speaker 1: us from the rooftops for such a long time and
22:58
Speaker 1: it will accrue to their entrments in the future. But
23:01
Speaker 1: all of our vehicles in the future weed, yes, all
23:06
Speaker 1: the vehicles that we were really made. Obviously, millions of vehicles,
23:10
Speaker 1: the vast majority, are capable of autonomy. You know, we're
23:14
Speaker 1: currently making. It's on the order of thirty five thousand
23:17
Speaker 1: autonomous vehicles a week. If you compare that to say,
23:19
Speaker 1: way most entire fleet if they have less than a
23:21
Speaker 1: thousand corners thirty five a week. Yeah, cars look normal
23:26
Speaker 1: and mostly normal. Cyber truck looks abnormal. In the cybercab
23:31
Speaker 1: SAT trug, we wanted to have something futuristic looking. It
23:33
Speaker 1: does look futuristic.
23:35
Speaker 2: It's worth learnt.
23:36
Speaker 1: Even with respect to the cybercab, it's not. It's especially
23:39
Speaker 1: not just a revolutionary vehicle design, but of revolution in
23:43
Speaker 1: vehicle manufacturing that is also coming with us with the cybercab.
23:48
Speaker 1: The cycle time, like the units per hour of the
23:51
Speaker 1: cybercab line is like, this is just really something special.
23:55
Speaker 1: I mean, listens will be half order of magas than
24:01
Speaker 1: other car manufacturing lines. Like not not even the same league,
24:04
Speaker 1: That's what I'm saying. Not in the same league. I
24:06
Speaker 1: said sart years ago.
24:07
Speaker 2: And then maybe the.
24:08
Speaker 1: Hottest part to comy will be the factory. Yeah, just
24:11
Speaker 1: like bio factory versus engineer or factory and it's empty mind. Yeah,
24:15
Speaker 1: so we're rapidly evolving acturing technology. So anyway, there's like
24:18
Speaker 1: basically I think having an irregular twenty five K models pointless,
24:22
Speaker 1: it would be silly, like it would be completely at
24:23
Speaker 1: us with what we believe.
24:25
Speaker 3: In the time of this world. But matters as the
24:27
Speaker 3: lowest cost.
24:28
Speaker 8: Per mile of the efficiency of that vehicle, and that's
24:31
Speaker 8: what we've done.
24:31
Speaker 1: With Yeah, but the rope texting exactly autonomous, it's fully
24:34
Speaker 1: considered cost per mile is what matters. And if you
24:37
Speaker 1: try to make a car that is essentially a hybrid
24:41
Speaker 1: manual automatic cards, it's it's not going to be as
24:44
Speaker 1: good as a dedicated autonomous car. That's our cab is
24:48
Speaker 1: is just not going to have stereo wheels and bats
24:51
Speaker 1: side otis or autonam. It'll cost on the order of
24:55
Speaker 1: cost roughly twenty five k. So it is a twenty
24:57
Speaker 1: five K car and you can you will be able
25:00
Speaker 1: to buy one on an exclusive exclusively if you want.
25:03
Speaker 1: So just what has steams there?
25:06
Speaker 2: Neither?
25:06
Speaker 6: Great, Thank you very much. The next question is what
25:09
Speaker 6: is Tesla doing to alleviate long way times at service centers?
25:13
Speaker 9: So we aim on solving problems at the source, so
25:16
Speaker 9: at the factory before they can't even affect our customers.
25:19
Speaker 9: We believe the best services, no service don't even have
25:22
Speaker 9: them if the car doesn't break. Yeah, exactly, that's the
25:24
Speaker 9: best thing. Don't see anyone the test shirt. You either
25:27
Speaker 9: do it fixed the issue upstream where you would remotely
25:30
Speaker 9: do it through software, maybe be at.
25:32
Speaker 2: Work or at home.
25:33
Speaker 9: And you know, car we partner and we address fixed issue,
25:37
Speaker 9: and we've partnered the field with service to make.
25:40
Speaker 3: Sure we're looking at the same issues.
25:42
Speaker 9: And additionally, just in Q three and Q four this
25:44
Speaker 9: year long, we have opened and will open in total
25:47
Speaker 9: nearly seventy locations, and in North America we significantly expand
25:51
Speaker 9: the size of each location and have doubled the size
25:54
Speaker 9: last year compared to this year.
25:56
Speaker 2: Yeah, I think it was like actually a lot of
25:57
Speaker 2: erics to have any large service set up because you
26:01
Speaker 2: can have specialization of labor, you can start, you can
26:04
Speaker 2: start your approach. Yeah, it should be more factory like
26:07
Speaker 2: where you can have dedicated lanes full particular types of service,
26:10
Speaker 2: and it's way easier for somebody to become expert in
26:12
Speaker 2: a few different types of repairs than in every repair. Exactly.
26:16
Speaker 9: This has helped us with the base that these heavy
26:19
Speaker 9: repairs like cogging up the lane. They've dedicated lanes for
26:22
Speaker 9: different type of repairs and so it's through the matters
26:25
Speaker 9: and really treating it like a factory.
26:27
Speaker 2: Yeah, this is this is where it tells.
26:29
Speaker 1: The structure I think has a structural advantage relative to
26:32
Speaker 1: the rest of the order industry because we make the
26:36
Speaker 1: cars and reservice the cars, Whereas I think there's a
26:39
Speaker 1: bit of a conflict of interest with the dealer model
26:43
Speaker 1: and the traditional area and the dealer model where the
26:46
Speaker 1: dealerships make most of their money on service and so
26:50
Speaker 1: they don't have they obviously assist incented to reduce the
26:53
Speaker 1: servicing costume, whereas in our case we are incentive to
26:57
Speaker 1: reduce the servicing costs because we we carry that servicede
27:00
Speaker 1: cost and we've got a good feedback without without cars.
27:05
Speaker 9: Yeah, with the factory, with the service leaders together and
27:08
Speaker 9: send back people from the factors to the field and
27:11
Speaker 9: field to the factory to see it firsthand, I buy
27:14
Speaker 9: suggestions for you know, manufacturing as well as for engineering
27:17
Speaker 9: on design.
27:18
Speaker 3: Yeah.
27:18
Speaker 1: So I prove this as a structural fundamental structural advantage
27:22
Speaker 1: of tesla versa is the rest of the order indist
27:24
Speaker 1: to do a bunch.
27:25
Speaker 5: Of work on the opperare side, only to automate diagnostics,
27:29
Speaker 5: so identifind what needs to be done to a car
27:31
Speaker 5: where it comes into service, but also automating all the
27:34
Speaker 5: preparation work and aligning all the resources that are necessary
27:37
Speaker 5: in order for the car and be very efficiently worked
27:40
Speaker 5: on once it arrived.
27:41
Speaker 2: So the parts are there, like the.
27:43
Speaker 5: Lift is schedule, the technician schedule, like everything like a card,
27:46
Speaker 5: this is what's wrong with me, and tell us fill
27:48
Speaker 5: the service center cart everything ready in advance.
27:51
Speaker 2: Yeah, please tect me and this is what's wrong.
27:56
Speaker 9: Yeah, instead of customer trying to translate the cars telling
27:59
Speaker 9: us directly and we're pulling that.
28:00
Speaker 2: Yeah. Most of the time you don't need to diagnose
28:03
Speaker 2: the car when it arrives the car. Yeah, this is
28:06
Speaker 2: like again, have fun technology advantage, attructural advantage compared to
28:11
Speaker 2: the rest of the order industry.
28:13
Speaker 7: It's I think it's under as to what all we
28:15
Speaker 7: are able to do. Yeah, and that's why because, like
28:18
Speaker 7: I said before, most of four cars except for cyber
28:21
Speaker 7: Drug look to say, right, so people don't realize that
28:26
Speaker 7: it has so much capability.
28:27
Speaker 1: Yeah, that's better than other cars. Yeah, but they're not
28:31
Speaker 1: like obviously like super futuristic.
28:33
Speaker 6: Great, thanks very much. The next question is please provide
28:37
Speaker 6: an update on the semi, what will the next stage
28:39
Speaker 6: of growth look like and when will f SD be ready?
28:42
Speaker 8: Sure, we posted an earning that we're progressing softly on
28:45
Speaker 8: the build of the semi factory there in our data
28:48
Speaker 8: factory a reno. We've released all our meter cap co
28:50
Speaker 8: expenditures for that program and we're on track to start
28:54
Speaker 8: have the builds in the second half of next year,
28:56
Speaker 8: with production starting in the first half of twenty twenty six,
28:58
Speaker 8: in ramping really throughout the year to full production. Some
29:01
Speaker 8: you know, growth will largely depend on our customer's adoption
29:03
Speaker 8: of the product.
29:04
Speaker 2: Well, I don't think we're going to be devided limited honestly.
29:06
Speaker 3: What should I do say?
29:07
Speaker 8: Which is like the brainer for the semi because it's
29:09
Speaker 8: really a commodity of total costs of.
29:12
Speaker 1: Yes, exactly, it's we were kind of ridiculous demand for
29:16
Speaker 1: the semi in that world where it's about how much
29:18
Speaker 1: do I spend to exactines per mile? It's a new brainer. Fundamentally,
29:23
Speaker 1: if you've got a semi work with the fully considered
29:26
Speaker 1: cost pro mile or for a ton of transport is
29:29
Speaker 1: better than a diesel truck. Any company that doesn't adopt
29:34
Speaker 1: an electric semi, will will lose. It's not a subject
29:36
Speaker 1: of things we want the stat we want. We want
29:39
Speaker 1: to have a beautiful semi drug. But frankly, if we
29:41
Speaker 1: made any semi drug would have matter.
29:42
Speaker 8: And this is proving so in our fleets and in
29:45
Speaker 8: Pepsi's partner, in fact that PEPSI actually said last week
29:48
Speaker 8: their drivers don't want to go back.
29:50
Speaker 2: As soon as we give anyone the electric semi. It's
29:53
Speaker 2: like that's the choice.
29:54
Speaker 3: It's what they want to drive.
29:54
Speaker 1: Yeah, yeah, yeah, that's like like so that like the
29:57
Speaker 1: most sceenor like if they're top drivers, will they get
30:00
Speaker 1: to drive. It's the thing they want to dry. It's
30:02
Speaker 1: super fun to drive. It's also very easy to drive.
30:05
Speaker 2: It's easy to drive, and it holds fast and it's
30:07
Speaker 2: like fast maybe too fast, well, but i mean like
30:12
Speaker 2: you know, like you've seen like videos of world it
30:14
Speaker 2: has electric somehow can go uphill.
30:17
Speaker 1: Yes, we can speeding past like the diesel trucking cars
30:20
Speaker 1: and cars so like it's responsive you floor it and
30:23
Speaker 1: the truck actually.
30:24
Speaker 8: And that's a benefit not only for the driver and
30:26
Speaker 8: for the goods, but also for safety in terms of
30:28
Speaker 8: other drivers on the road. You don't get stuck behind
30:30
Speaker 8: the semi, You're not like, yeah, you know, in a
30:32
Speaker 8: slow down situation and on rad I mean how that
30:35
Speaker 8: plays into you know, FST which is the second part
30:37
Speaker 8: of the question. All of the semis have been since
30:40
Speaker 8: a couple of hundred we've deployed already and the ones
30:42
Speaker 8: that will be building next year and throughout SUTURE have
30:45
Speaker 8: all of the hardware and the cameras necessary to deploy
30:48
Speaker 8: fifty And we're currently training with that small fan that
30:51
Speaker 8: we have and as soon as the fleet is trained
30:53
Speaker 8: and then they're on that's round, you know, we'll get
30:56
Speaker 8: FIST onto that platform.
30:57
Speaker 1: Yeah, I mean, it'd be a massive improvement in dry
31:00
Speaker 1: or fatigue, you know, because and drivers safety. We've got
31:04
Speaker 1: sort of the anti Jack Knight thing software. You know,
31:07
Speaker 1: you don't have to worry about your breaks overheating if
31:09
Speaker 1: you go down there down the steep hill because that
31:11
Speaker 1: we use regenerating like that that energy goes back.
31:14
Speaker 2: Into the pack.
31:16
Speaker 3: Actually, when we leave ring on something, it's.
31:18
Speaker 2: Like radically better than it is with someway. The drivers
31:20
Speaker 2: love it, guys, Thanks very much.
31:22
Speaker 6: Our next question is when will Tesla incorporate X and
31:25
Speaker 6: ROCK in all the Tesla vehicles.
31:27
Speaker 1: Well, I mean these are relatively small things, you know,
31:31
Speaker 1: but yeah, I think we'll keep expanding what is available
31:36
Speaker 1: in the car on the screen and also improving on
31:40
Speaker 1: the browsers, so you just generally you can access anything
31:42
Speaker 1: you want in the car. In fact, for that says
31:44
Speaker 1: we'll just get to full autonomy. You actually want a
31:47
Speaker 1: system that you can do anything like if you want
31:50
Speaker 1: a browser Internet, if you want to ask AI questions,
31:52
Speaker 1: if you want to watch a movie, if you want
31:54
Speaker 1: to play a video game, if you want to do
31:56
Speaker 1: some productivity thing.
31:57
Speaker 2: You can do anything you want. And you know, tom
32:00
Speaker 2: vehicle because you don't need to dry them.
32:01
Speaker 1: So that's why the Cybercat got a nice picture and
32:04
Speaker 1: the great sound system, so you can what's a great
32:08
Speaker 1: movie with It's like being like an in personal movie theater, yeah,
32:11
Speaker 1: postal movie theater.
32:12
Speaker 5: This is why we've been building this functionality and gaming
32:15
Speaker 5: to the car, adding MPs and other media applications of
32:17
Speaker 5: the car, because that's what you're gonna the cars will
32:20
Speaker 5: be built today.
32:20
Speaker 2: Sometimes we find games by the way, picture people haven't
32:22
Speaker 2: tried it.
32:23
Speaker 1: There's like Castle doing Bad and Platopia and a bunch
32:26
Speaker 1: of really fun games in the car.
32:28
Speaker 5: We're constantly looking at what features to add next, and
32:31
Speaker 5: we're paying attention to what's most commonly rebuffed by our customers.
32:35
Speaker 2: Yeah, playing Castle bombad Do you want to thank you
32:37
Speaker 2: guys very much?
32:38
Speaker 6: The next question is Elon mentioned on supervised FSD in
32:41
Speaker 6: California and Texas next year. Does that mean regulators have
32:44
Speaker 6: agreed to it in the entire state for existing hardware
32:47
Speaker 6: three and four vehicles.
32:49
Speaker 1: As I said earlier, California loves regulation, but they have
32:53
Speaker 1: a pathway. Yeah, I mean there's a pathway. Obviously Waymo
32:56
Speaker 1: operates in California. So there's a lot of forms and
32:58
Speaker 1: a lot of approvals that are required to be shocked
33:00
Speaker 1: if we don't get approved next year. But it's just
33:02
Speaker 1: not something we totally control. But I think we won't
33:04
Speaker 1: get approval next year in California and Texas, and for
33:07
Speaker 1: the end of the year it will branch out beyond
33:09
Speaker 1: California and Texas.
33:11
Speaker 8: I think it's important to reiterate this mull union you're
33:13
Speaker 8: CERTI finding a vehicle out a federal level in the
33:15
Speaker 8: US is done by meeting FMVSS regulations. All our vehicles
33:19
Speaker 8: today that are produced that are autonomus capable meet all
33:21
Speaker 8: those regulations. The side of capital need build regulations and
33:24
Speaker 8: so the deployment of the vehicle to the road is
33:26
Speaker 8: no limitation, but is a limitation, is what you said
33:29
Speaker 8: at the state level where they control autonous vehicle deployment.
33:32
Speaker 8: Some states are relatively easy, as you mentioned for Texas,
33:34
Speaker 8: and it's other ones have plays like California that may
33:36
Speaker 8: take a little longer. Other ones haven't set up anything yet,
33:39
Speaker 8: and so we will work for those state by state.
33:41
Speaker 2: And I do think we should have a federal I
33:44
Speaker 2: agree that like autonomous vehicles should be approved, you should
33:47
Speaker 2: be it should be possible to condress.
33:48
Speaker 3: If you're listening, let's get a federal av.
33:50
Speaker 1: This should be a federal federal approval process for autonomous vehicles.
33:54
Speaker 1: I mean that's how the fmbsas it started and work
33:57
Speaker 1: Federal Motor Ahiric Club.
33:58
Speaker 8: So I mean in twenty seventeen to eighteen, and you know,
34:00
Speaker 8: it's when regulators started looking at it, and it's really
34:02
Speaker 8: kind of stalled since then. But we would appreciate and
34:05
Speaker 8: would support helping out. It really needs to be that
34:08
Speaker 8: at like a national approval is important. You know, if
34:10
Speaker 8: there's an Apartment of Governor Efficiency try to help make
34:12
Speaker 8: that happen. And you took for everyone not just hasn't it.
34:15
Speaker 1: But you know, something's in the US are state by
34:17
Speaker 1: state regulated, like for example, insurance. It' it's incredibly painful
34:21
Speaker 1: to do it state by state fifty states, and I
34:23
Speaker 1: think there should be a national approval process for autonomy.
34:27
Speaker 6: Great guys. The next question is what is the plan
34:30
Speaker 6: for twenty twenty five.
34:32
Speaker 7: I mean, basically we talked to this there's a lot
34:34
Speaker 7: going on. Multi mentioned that we're working on cheaper models
34:39
Speaker 7: to come out. I mean they're what Cosset the team
34:42
Speaker 7: is doing to get the factories ready today to try
34:44
Speaker 7: and make that happen done.
34:45
Speaker 1: Yeah, to make a lower class car is insanely high.
34:50
Speaker 1: But like it is harder to get like twenty percent
34:52
Speaker 1: of the cost out of a car than it is
34:54
Speaker 1: to design the car and build the entire factory in
34:56
Speaker 1: the post place.
34:57
Speaker 2: It's like excruciating.
34:59
Speaker 1: And there's not a lot of movies made about the
35:02
Speaker 1: heroes who got twenty percent of the cost down of
35:04
Speaker 1: a car, but let me tell you there should be.
35:07
Speaker 2: If we can got that is incredibly heroic. It's a
35:10
Speaker 2: little che isn't. It's not like a yeah, it's like
35:12
Speaker 2: there should be.
35:13
Speaker 1: The heroes got twenty percent cost out of a car,
35:15
Speaker 1: is like, damn, I want respect to them. Like if
35:19
Speaker 1: you actually saw how hard, if people actually saw how
35:21
Speaker 1: hard was to do that, you'd be like, whoa, that's
35:24
Speaker 1: damn hard.
35:25
Speaker 3: Just yesterday we were talking about potty.
35:27
Speaker 1: I mean honestly, you're like literally yeah, I mean there's
35:31
Speaker 1: there's a lot of what I do call it sort
35:33
Speaker 1: of like getting cost out of things. It's kind of
35:34
Speaker 1: like it's like Game of Pennies. It's like Game of Thrones,
35:37
Speaker 1: but pennies. You know, first approximation, You've if you've got
35:40
Speaker 1: ten thousand nine ers in a car, very approximation, and
35:44
Speaker 1: each of them costs four dollars, then you have a
35:46
Speaker 1: forty thousand dollars car. So then if you want to
35:47
Speaker 1: have a thirty five thousand dollars car, if you got
35:49
Speaker 1: to get fifty cents on average out of the ten
35:51
Speaker 1: thousand naders every time, every part, yeah, it's like yeah,
35:55
Speaker 1: and then obviously the best is you're delete some parts.
35:58
Speaker 2: We're doun be able to lead a lot of parts.
36:00
Speaker 1: I'm very I'm very excited about the cybercap design, how
36:05
Speaker 1: we're rethinking the design of a car for the cybercamp,
36:08
Speaker 1: designing it will alcohig volume production and then design a
36:13
Speaker 1: machine that builds the machine. That is I think revolutionary
36:16
Speaker 1: and that there's no other car company that's even trying
36:20
Speaker 1: to do what we're doing, Like I don't even heard
36:22
Speaker 1: of it. Actually, in fact I'm certain there's and when
36:25
Speaker 1: five five times better than additional factory like a.
36:29
Speaker 8: Cycle time and like part deletion, material movement. I don't
36:32
Speaker 8: think any other car company has the same level of
36:35
Speaker 8: like integration of thought that we have when it comes
36:38
Speaker 8: to like when you design a part from the light
36:40
Speaker 8: shiet of paper, who's going to make it? Where is
36:42
Speaker 8: it going to be made, how's it going to be shipped,
36:44
Speaker 8: how's it going to be assembled into the vehicle, And
36:46
Speaker 8: like at any one point, if something is done in
36:49
Speaker 8: a silo, it becomes a bottleneck of either cost or
36:53
Speaker 8: time or efficiency.
36:54
Speaker 10: But with the with the rod Taxi, the development, like
36:57
Speaker 10: we've done a good job on the combining all that
37:01
Speaker 10: and then like blowing up how it's made and saying
37:03
Speaker 10: it should be made this way and rethinking it.
37:05
Speaker 8: Also it's the most efficient factory possible. That shows in
37:07
Speaker 8: our big wilship and our capex efficiency when we deploy
37:10
Speaker 8: it shows them in our parts, shows the simplicity and vehicle,
37:13
Speaker 8: but also have performance in terms of like end user
37:16
Speaker 8: state just on the energy front.
37:18
Speaker 7: Also in twenty five, we will have Stardulent Fashion out
37:21
Speaker 7: the manufacturer Shanhai, will continue to increase our storage deployments.
37:25
Speaker 7: With all three, we plan to continue expanding our supercharging network,
37:30
Speaker 7: getting more wims on our network, some ramp as you
37:33
Speaker 7: want to talk about that will keep going. And then
37:37
Speaker 7: we also will have our lit refinery starting to produce.
37:41
Speaker 2: Yeah, so many things.
37:43
Speaker 1: I like crazy thing is like tell us winning basically
37:46
Speaker 1: out almost every single thing we're doing. If we're not winning,
37:49
Speaker 1: now we're one in an awareness with there our entire
37:52
Speaker 1: large companies that that's the only thing they do.
37:54
Speaker 3: I mean it's a company.
37:56
Speaker 7: There are multiple companies within the company.
37:58
Speaker 1: Yeah, till those like many companies do one yeasic Thank
38:01
Speaker 1: you guys.
38:02
Speaker 2: Just a few more.
38:03
Speaker 1: What is going on the Tesla Roadster. Well, I'd just
38:06
Speaker 1: like to thank our long suffering deposit holders of the
38:08
Speaker 1: Tail Roadster. You know, the reason it hasn't come out
38:11
Speaker 1: yet is because it is that the roaster is not diet,
38:14
Speaker 1: not just the icing on the cake, it's the cherry
38:16
Speaker 1: on the icing on the cake. And so you know,
38:19
Speaker 1: our large emission is to accelerate the progress towards a
38:22
Speaker 1: sustainable energy future, you know, try to do things that
38:25
Speaker 1: maximize the probably the future is good for humanity and
38:29
Speaker 1: for Earth, and so that necessarily means that like the
38:33
Speaker 1: things like that are kind of like dessert. We'd like,
38:35
Speaker 1: like we do all love to work on the tailor
38:38
Speaker 1: next to in tesl roster. It is super fun and
38:41
Speaker 1: we are working on it, but it has to come
38:43
Speaker 1: behind them the more things, the things that have a
38:46
Speaker 1: more serious impact on the good of the world.
38:48
Speaker 2: So just thank you to all lot long suffering tailor
38:50
Speaker 2: roaster deposit holders.
38:52
Speaker 1: And we are actually finally making progress on that and
38:56
Speaker 1: we're close to finalizing the design on that. It's really
38:59
Speaker 1: going to be a big spectacular you know, frind Mine,
39:03
Speaker 1: Peter tail you know, and sometimes Peter til now right,
39:07
Speaker 1: we're really good friends. Peter, you know, was limiting how
39:12
Speaker 1: you know the future doesn't have flying cards. Well, we'll
39:15
Speaker 1: see before to come, I think very much.
39:18
Speaker 6: The next one is quite similar to other questions we've had,
39:21
Speaker 6: so when I combine it with the final question, so briefly,
39:23
Speaker 6: could you just detail how ROBOTEXTI will roll out. We'll
39:27
Speaker 6: start with the tested deployed fleet and then allow customers
39:30
Speaker 6: to add THEIRS on the subscription model. And then will
39:33
Speaker 6: hardware three be capable of fix?
39:35
Speaker 2: Because you're going the hardware three? What we solve it
39:38
Speaker 2: well not five?
39:38
Speaker 4: Was it was easier to make progress with starting with
39:41
Speaker 4: Hardward four, then feeling on the solution, then backporting into
39:44
Speaker 4: Hardbred three instead of directive working on Hardbred three. Given
39:47
Speaker 4: that Hardward four us more like fundamental hardware capabilities, I
39:51
Speaker 4: think that trend will continue into the next few quarters
39:55
Speaker 4: as well, by the first figure on the solution rapidly
39:57
Speaker 4: with four and then backport it trying the kind of
40:00
Speaker 4: it just takes longer to none of those things because
40:02
Speaker 4: it's not fundambly supported in the hardware and it's emulated.
40:06
Speaker 4: But yeah, it's initially working on Harvard for backboarding it
40:10
Speaker 4: to hundred three.
40:11
Speaker 1: Yeah, it's actually answer we're not in harder sent sure,
40:15
Speaker 1: but I'm sure I mentioned by some measures hardware for
40:18
Speaker 1: it has several times the capability of Hardware three. It's
40:21
Speaker 1: it's easier to get things to work with Harvard for
40:24
Speaker 1: and then it takes a lot of effort to sort
40:26
Speaker 1: of squeeze that proximality into hardware three. There is some
40:29
Speaker 1: chance that hardware three does not achieve the safety level
40:33
Speaker 1: that allows for untou advised FSD. If that turns out
40:38
Speaker 1: to be the case, we will upgrade those who have
40:41
Speaker 1: bought hardware three FSD for free, and we have designed
40:44
Speaker 1: the system to be upgradeable, you know, just to just
40:48
Speaker 1: sort of switch out the computer type thing. The cameras
40:50
Speaker 1: are okay, But we don't actually know the answer to that.
40:52
Speaker 1: But if it does turn out, we will make sure
40:54
Speaker 1: we take care of those who have board MSDY on
40:56
Speaker 1: Hardward three.
40:57
Speaker 6: Right and in the last few minutes that we have left,
41:00
Speaker 6: we will try to get some analyst questions. The first
41:03
Speaker 6: question will be coming from here fairg and New Street.
41:06
Speaker 3: Here.
41:07
Speaker 2: Please feel free on me yourself.
41:09
Speaker 11: I was wondering about the compute you're ramping up. You
41:11
Speaker 11: gave interesting statistics on how much you have, and you
41:14
Speaker 11: said you don't feel your compute constraint. I was wondering
41:17
Speaker 11: how you are putting to work this additional compute. Is
41:22
Speaker 11: that a game for you? Of creating larger and larger models,
41:26
Speaker 11: like next generation of models that are larger the way
41:29
Speaker 11: open air I go from GPT three to GPT four.
41:33
Speaker 11: Or is that more like you're set on your model
41:36
Speaker 11: and you need to throw more and more compute to
41:39
Speaker 11: accelerate the pace of learning to improve reliability. And then
41:44
Speaker 11: I had a quick flow pretty quick on you rollout
41:47
Speaker 11: in Texas and in California next year. The plan as
41:50
Speaker 11: you see today, is it to roll out like a
41:54
Speaker 11: fleet or two with cars that will start with like
41:57
Speaker 11: a super visory, like some super a vision someone sitting
42:01
Speaker 11: at the wheel just in gaze and removing these super
42:04
Speaker 11: advisors posssively or are you aiming for going full fledged
42:09
Speaker 11: without even a human super super advisor when you get started.
42:12
Speaker 1: Well, I guess regard to answer the question, the nature
42:15
Speaker 1: of real world AI is different from say ALAM in
42:19
Speaker 1: that you have a massive amount of context so that
42:22
Speaker 1: like case it tells the summary cameras are to night
42:25
Speaker 1: if you include internal camera that that that so you've
42:29
Speaker 1: got to tagabytes of context and and that that is
42:32
Speaker 1: then distilled down into a small number of control outputs.
42:36
Speaker 1: You know, whereas it's it's like you don't really it's
42:38
Speaker 1: very rare to have. In fact, I'm not sure any
42:41
Speaker 1: ALM out there do gigabytes of context. And then you've
42:44
Speaker 1: got to you've got to the process that in the
42:47
Speaker 1: car with a very small amount of compute power. So
42:52
Speaker 1: you know it's it's it's all doable and it's happening.
42:54
Speaker 1: But it is a different problem than than what's a
42:58
Speaker 1: Gemini or open is doing.
43:00
Speaker 2: Now.
43:01
Speaker 1: Part of the way you can make up for the
43:02
Speaker 1: fact that the inferenced computer is quite small is by
43:07
Speaker 1: spending a lot of effort on training. Just like a human,
43:10
Speaker 1: the more you train on something, the less mental workload
43:13
Speaker 1: it takes when you do it. When the human starts driving,
43:16
Speaker 1: it absorbs your whole mind. But then as you train
43:21
Speaker 1: more and more on driving, get very good and the
43:23
Speaker 1: driving becomes a background task, it only absorbs a small
43:26
Speaker 1: amount of your mental capacity because of you have a
43:28
Speaker 1: lot of training. So we can make up for the
43:32
Speaker 1: fact that the inferenced computer is tiny compared to a
43:35
Speaker 1: ten kilo or bank of gvus because we've got a
43:37
Speaker 1: few hundred watts of in furnst computer. We can make
43:41
Speaker 1: up that with heavy training. There's also vast amounts of
43:44
Speaker 1: data coming in and then's sorting out what training is important.
43:47
Speaker 1: With the vast amounts or video data coming in, what
43:51
Speaker 1: is actually the most important for training. That's quite difficult,
43:54
Speaker 1: But as I said, we're not currently training compute construct.
43:58
Speaker 2: You want to elaborate the.
43:59
Speaker 4: Training has more train larger models also a train quicker,
44:03
Speaker 4: but in the end we've still got to pick which
44:05
Speaker 4: models are performing better. So the validation effort to picking
44:08
Speaker 4: the models customer sparent dimension is pretty large. Were to
44:12
Speaker 4: tive a lot of miles to go in close toop.
44:14
Speaker 4: We do have simulation and other ways to get those metrics,
44:17
Speaker 4: and those two help, but in the end that's a
44:19
Speaker 4: big bottleneck. That's why we're not training complete constraint alone.
44:22
Speaker 4: And there's other access of scating as well, which is
44:24
Speaker 4: a data figuring offic straight as more useful than that
44:27
Speaker 4: is an important as we're focusing on that.
44:29
Speaker 3: Yeah, so as it really is.
44:31
Speaker 8: The second part of your question here about safety drivers
44:34
Speaker 8: and rolling it out. Each state has different requirements that
44:37
Speaker 8: you know in terms of how many miles and how
44:39
Speaker 8: much time you need to have a safety driver. And
44:41
Speaker 8: now I have a safety driver, we're going to follow
44:42
Speaker 8: all those were not if regulations you're out there with
44:45
Speaker 8: safety as a priority. But well was obviously already and
44:48
Speaker 8: safety is there will over.
44:52
Speaker 3: From the right jet.
44:53
Speaker 1: We think that we'll be able to have drive all
44:55
Speaker 1: those test those doing paid rides next year. Sometimes makes
45:00
Speaker 1: sure all thank you and our next question comes from
45:04
Speaker 1: Adam Jonas at Morgan Stanley. Adam, please feel free to
45:07
Speaker 1: unnet yourself.
45:08
Speaker 12: Okay, thanks, everybody, just had a question about the relationship
45:13
Speaker 12: between Tesla and Xai. Many investors are still not clear
45:17
Speaker 12: how the work at Xai is truly beneficial to Tesla.
45:20
Speaker 12: Some even take the view that the two companies may
45:23
Speaker 12: be in competition with each other in terms of talent
45:25
Speaker 12: and tech and even your time, elon. So what's your
45:28
Speaker 12: message to investors on that relationship between Tesla and Xai
45:32
Speaker 12: and where do you see it going over time?
45:35
Speaker 3: Thanks?
45:35
Speaker 1: Well, I should say that you know Xai has been
45:38
Speaker 1: helpful to Tesla AI quite a few times in terms
45:42
Speaker 1: of things like scaling up training just even recently in
45:46
Speaker 1: the last week or so. It's in training or where
45:50
Speaker 1: if you're doing a big training one note fails be.
45:53
Speaker 2: Able to continue training. But there are different problems.
45:57
Speaker 1: X is working on artificial eral intelligence or official superintelligence
46:02
Speaker 1: it tells us trying to make autonomous cars and autonomous robots.
46:06
Speaker 2: There are different problems.
46:08
Speaker 7: Also, like all not all e is equal, right, I
46:11
Speaker 7: mean there's a is a broad spectrum.
46:13
Speaker 3: Yeah, and we have our own swim lanes.
46:15
Speaker 7: Yeah, there are certain things which we can collaborate on
46:18
Speaker 7: if needed, But for the most part we're sloving different issues.
46:21
Speaker 1: Ye tells us focused on real well data and that
46:25
Speaker 1: is saying it's it is quite a bit different from
46:29
Speaker 1: an element that you have. You have massive context in
46:31
Speaker 1: the more video and some amount of video that's going
46:35
Speaker 1: to be distilled. Like with extremely efficient in furns compute,
46:39
Speaker 1: I do think tells us the most efficient in the
46:42
Speaker 1: world in terms of inference compute. Like because out of
46:45
Speaker 1: necessity we have to we have to be very good
46:49
Speaker 1: at in efficient in ferns. We can't pretend to kill
46:52
Speaker 1: ons GPUs in a car. We've got a couple of
46:55
Speaker 1: hundred ones. You know, it's pretty well designed, tells the
46:58
Speaker 1: AI chow, but it's still a cup hundred one. But
47:00
Speaker 1: there are different problems. I mean, it's you know, like
47:02
Speaker 1: the s not that XAI is like we're just running
47:05
Speaker 1: en Burns. I mean it's running inference like answering questions,
47:08
Speaker 1: answering Tesla questions on a ten.
47:10
Speaker 2: Killer what rank? It's like, ye't put that in a car.
47:14
Speaker 2: It's a different problem. No, exactly, XAI is.
47:18
Speaker 1: Because because I felt there wasn't a truth seeking digital
47:22
Speaker 1: super intelligence company out there. That's what it came down
47:25
Speaker 1: to that they needed to be a truth seecret like
47:29
Speaker 1: an AI company that is very rugorous about a truthful
47:32
Speaker 1: And so I'm not saying XI is perfect, but that
47:34
Speaker 1: is that that is at least the franspiration, even if
47:38
Speaker 1: something is politically corrected, should still be truthful. I think
47:42
Speaker 1: this is very important for a safety. I think XI
47:46
Speaker 1: has been helpful to Tesla and will continue to be
47:48
Speaker 1: helped to Tesla. But there are very different problems. Great,
47:52
Speaker 1: you know what other co company has a world class
47:55
Speaker 1: trip design team?
47:56
Speaker 2: Zero? What are the cock companies a world class that
47:58
Speaker 2: a team like Tesla does? Zero? Those all startups? The
48:02
Speaker 2: creator for scratch.
48:03
Speaker 6: Great, thank you a lot, and I think that's unfortunately
48:06
Speaker 6: all the time that we have for today, and we
48:07
Speaker 6: appreciate all of your questions and we look forward to
48:09
Speaker 6: eating squarter. Thank you very much, and goodbye.