Latest / Elon Musk Thinking / Elon Musk Latest Announced at Tesla's 2025 Shareholders Event!!!

Transcript

Elon Musk Latest Announced at Tesla's 2025 Shareholders Event!!! · Elon Musk Thinking

0:00
Speaker 1: I'm going to say a bunch of things that probably
0:01
Speaker 1: I shouldn't say, but but that's what keeps it interesting.
0:07
Speaker 1: Have you watched any other annual shareholder meeting? Honestly, it's
0:11
Speaker 1: like if you need to go to sleep. Sure, I
0:14
Speaker 1: mean that other shareholder meetings are like snoozefest, but ours
0:18
Speaker 1: are bangers. Look, this is sick. Yeah, We're got like
0:26
Speaker 1: the cyberpunk nightclub here with real robots just standing there
0:31
Speaker 1: and milling around and dancing and round out our engineering
0:36
Speaker 1: headquarters in palab Alto. The robots just walking around the
0:39
Speaker 1: office twenty four to seven with no one minding them.
0:42
Speaker 1: They just and then they go charge themselves. And Yeah,
0:45
Speaker 1: the scale of Optimists, I like, I said, that's really
0:48
Speaker 1: gonna be something else. I think it's gonna be the
0:50
Speaker 1: biggest product of all time by far. Yeah, so bigger
1:00
Speaker 1: than cell phones, bigger than anything.
1:03
Speaker 2: I guess the way to think about it is that.
1:05
Speaker 1: Every human on earth is gonna want to have their
1:07
Speaker 1: own personal R two D two C three field, So
1:13
Speaker 1: who wouldn't. But actually, optimists will be even better than
1:17
Speaker 1: the like O R two D two would beep at
1:20
Speaker 1: you and hard to figure out what is good talking
1:21
Speaker 1: about three field to translate, but Optimist is gonna be.
1:27
Speaker 2: Like, everyone's gonna want one.
1:29
Speaker 1: I think in terms of industry providing products and services,
1:33
Speaker 1: I think it's probably I don't know, three to five
1:37
Speaker 1: robots in industry for everyone that's a personal robot. I
1:40
Speaker 1: think there could be tens of billions of Optimist robots
1:44
Speaker 1: out there now. Obviously it's very important we pay close
1:47
Speaker 1: attention to safety here because we do want the Star
1:50
Speaker 1: Wars movie, not.
1:51
Speaker 2: The Gim Cameron movie. I love Jim Cameron's movies.
1:54
Speaker 1: But yeah, yeah, So we're gonna launch on the fastest
2:00
Speaker 1: production ramp of any product, of any large complex manufacturer
2:04
Speaker 1: product ever and starting with building a million unit production
2:09
Speaker 1: line in Fremont, and that's Line one, and then a
2:15
Speaker 1: ten million unit per year production line here on the
2:26
Speaker 1: I don't know where we're gonna put one hundred million
2:27
Speaker 1: unit production line, maybe on Mars, I don't know, but
2:34
Speaker 1: I think it's gonna literally get to one hundred million
2:36
Speaker 1: a year, maybe even a billion a year. And if
2:39
Speaker 1: you can get to the five second cycle time. So
2:50
Speaker 1: it's a lot of cars. So these will be everywhere
2:56
Speaker 1: in the future. Oh, and wanted, We wanted to look futuristic,
3:06
Speaker 1: so like it changes the look of the roads.
3:10
Speaker 2: Now. The Yeah, the ingredients.
3:13
Speaker 1: When you look at what Optimists is, what's required to
3:17
Speaker 1: make Optimists and the various ingredients. What do you need
3:20
Speaker 1: to do to make to do high volume humanoid robot production.
3:25
Speaker 2: I think it's worth considering that.
3:26
Speaker 1: Really the cars we make are already robots, but there
3:29
Speaker 1: are four wheeled robots. So Tesla is already the biggest
3:33
Speaker 1: robot manufacturer in the world because every car we make
3:35
Speaker 1: is a robot.
3:37
Speaker 2: And when you break it down to the fundamental.
3:39
Speaker 1: Elements, you've got batteries, power electronics, motors, gearboxes, you've got connectivity,
3:47
Speaker 1: you've got a vision based AI high optimists and all
3:52
Speaker 1: the various pieces that you need for a humanoid robot.
3:56
Speaker 1: You need the AI chip, you need the AI software,
3:59
Speaker 1: you need to build manage a large fleet. And so
4:03
Speaker 1: really Optimus is a robot with arms and legs as
4:06
Speaker 1: opposed to a robot with wheels. Telsa's ideally suited I
4:11
Speaker 1: think to make to succeed in this arena, you will
4:14
Speaker 1: see certainly many companies showing demonstration robots. There's really three
4:20
Speaker 1: things that are super difficult about robots. One is the
4:23
Speaker 1: engineering of the forum in hand, because the human hand
4:27
Speaker 1: is an incredible thing. Actually, it's super dexterous engineering the
4:32
Speaker 1: hand really well, the real world AI, and then volume manufacturing.
4:39
Speaker 1: Those are generally the things that are missing one or
4:43
Speaker 1: more of those things missing from other companies. So Telsa
4:45
Speaker 1: is the only one that has all three of those.
4:48
Speaker 2: Yeah.
4:49
Speaker 1: Yes, So this is the Optimist kind of initial it's
5:01
Speaker 1: kind of a prototype production line.
5:04
Speaker 2: The high volume production line.
5:05
Speaker 1: Will be very automated, obviously, but this is really the
5:08
Speaker 1: production line that we use to make the prototypes. So
5:11
Speaker 1: you can get a sort of rough sense for what
5:14
Speaker 1: it takes to build the robot.
5:17
Speaker 2: You got to pull the finger.
5:32
Speaker 1: And then, as I've said before, I think once we
5:34
Speaker 1: reach about a million units per year of sustained production,
5:37
Speaker 1: or in excess of that, I think probably the cost
5:40
Speaker 1: of production is around twenty thousand dollars in curry your dollars,
5:46
Speaker 1: so this will be something very affordable. And yeah, like
5:52
Speaker 1: I said, I think Optimists will ultimately increase the size
5:57
Speaker 1: of the economy, probably by a factor of ten or
6:00
Speaker 1: or next year we start production with Optimist version three.
6:04
Speaker 1: This is what you're seeing here is Optimists version two
6:06
Speaker 1: point five. Optimist three is an incredibly good design. The
6:13
Speaker 1: tales the engineering team is amazing. When you see Optimists three, Yeah,
6:25
Speaker 1: it will seem as though that there's someone like a
6:28
Speaker 1: person in a robot outfit, which is how we started
6:31
Speaker 1: with Optimists.
6:34
Speaker 2: It really is going to be something special.
6:36
Speaker 1: And then Optimist Optimus four that hopefully starts production in
6:41
Speaker 1: twenty seven, and then Optimus five and twenty eight, so
6:45
Speaker 1: it's like an annual release cycle with significant improvements with
6:48
Speaker 1: each one and gigantic increases in the scale of production. Yeah,
6:56
Speaker 1: sustainable abundance via AI in robotics, that's the future. Credible
7:10
Speaker 1: and as I think most people here know, the safety
7:14
Speaker 1: statistics show that miles driven on FSD are much safer
7:18
Speaker 1: than miles driven without it. What this will translate to
7:22
Speaker 1: ultimately is saving the lives of millions of people and
7:25
Speaker 1: preventing hundreds of millions of accidents.
7:28
Speaker 2: So a massive.
7:30
Speaker 1: Increase in lives solved and tragedy is avoided.
7:34
Speaker 2: It's going to be amazing.
7:44
Speaker 1: How many people here have tried fourteen point one? Okay,
7:49
Speaker 1: all right, cool, Yeah, you can see the even with
7:52
Speaker 1: the point releases, it's getting quite a bit better. Should
7:55
Speaker 1: be pretty smooth at this point, but really fourteen point
7:58
Speaker 1: two is there are major changes to fourteen point two
8:02
Speaker 1: and then fourteen point three, and I think by fourteen
8:05
Speaker 1: point three is when we'll really be at the point
8:08
Speaker 1: where you can just pretty much pull asleep and wake
8:10
Speaker 1: up at your destination. And then I've been putting a
8:14
Speaker 1: lot of time into the new Tesla chip design because
8:18
Speaker 1: in order to have a functional robot, you have to
8:21
Speaker 1: have a great AI chip, and it needs to be
8:24
Speaker 1: an inexpensive chip, and it needs to be very power efficient.
8:28
Speaker 1: So we believe the AI five chip will be probably
8:33
Speaker 1: about a third of the power of say something like
8:36
Speaker 1: a Blackwell and a video Blackwell, which is a great
8:38
Speaker 1: chip for roughly a comparable performance and much less than
8:45
Speaker 1: ten pescent of the cost. So this is a chip
8:55
Speaker 1: that is very much optimized for the Tesla AI software stack.
8:59
Speaker 1: So it's not meant to be a general purpose chip.
9:01
Speaker 1: It's meant to be an amazing chip for the Tesla
9:04
Speaker 1: AI software And a couple of things that I think
9:08
Speaker 1: make like how E Tailless are able to achieve such
9:11
Speaker 1: an improvement. I think it is because because we are specialized,
9:14
Speaker 1: we're not trying to Nvidia has to serve the superset
9:18
Speaker 1: of all past and future customers, so all of their requirements,
9:22
Speaker 1: all of the software that they're written, has to work,
9:24
Speaker 1: which is a very difficult problem, whereas we just need
9:27
Speaker 1: to make it work for our software, and so we're
9:30
Speaker 1: able to simplify the chip dramatically. And then we also think,
9:34
Speaker 1: I think we're unique in this, but we have an
9:37
Speaker 1: integer based system, and integer operations are fundamentally more efficient
9:42
Speaker 1: than floating point operations, so we can do floating point
9:46
Speaker 1: but we but the vast majority of our inference is
9:49
Speaker 1: done in integer, which is, if you're familiar with sort
9:54
Speaker 1: of logic gates like the simplicity of integer, and it's
9:59
Speaker 1: integer is much more power eficient and much more silicon efficient.
10:02
Speaker 1: But you have to you actually have to train for
10:05
Speaker 1: integer inference, which everyone else is training for flooding point.
10:09
Speaker 1: It's like a niche technical detail, but it's actually very important.
10:14
Speaker 2: This is going to be a great chip.
10:16
Speaker 1: So this chip will be made in basically in four places, TISMC, Taiwan,
10:23
Speaker 1: Samsung Career, TSMC Arizona, and TSOMC Texas.
10:34
Speaker 2: And we already know what improvements to make for AI six.
10:38
Speaker 1: So I'm hopeful that we can within less than the
10:42
Speaker 1: year of AI five starting production, we can actually transition
10:46
Speaker 1: in the same fab to AI six and double all
10:49
Speaker 1: of the performance metrics. I'm super hardcore on chips right now,
10:58
Speaker 1: as you may be able to tell chips on the brain,
11:02
Speaker 1: I dream about chips literally. I can drow the at
11:09
Speaker 1: least the broad rush stroke physical design of the A
11:11
Speaker 1: I five chip by well by heart.
11:13
Speaker 2: At this point, it's a good chip. It's good chips, sir,
11:20
Speaker 2: So this is really key. Now.
11:22
Speaker 1: One of the things I'm trying to figure out is
11:24
Speaker 1: how do we make enough chips. I have a lot
11:27
Speaker 1: of respect for our the Tesla partners, TSMC and Samsung. Yeah,
11:32
Speaker 1: maybe we'll do something with Intel. We haven't signed any deal,
11:36
Speaker 1: but it's probably worth having discussions with Intel. But even
11:40
Speaker 1: when we extrapolate the best case scenario for chip production
11:44
Speaker 1: from our suppliers, it's still not enough. So I think
11:51
Speaker 1: we may have to do a Tesla Terra fab. It's
12:01
Speaker 1: like Gigabit way bigger. I can't see any other way
12:05
Speaker 1: to get to the volume of chips that that we're
12:07
Speaker 1: looking for, So I think we're probably gonna have to
12:10
Speaker 1: build a gigantic chip fab Yeah, gotta be done. So anyway,
12:29
Speaker 1: some of this stuff I've already talked about. Yep, we've
12:32
Speaker 1: done a tremendous amount for sustainable energy and that is
12:35
Speaker 1: only going to grow over time. The world is moving
12:38
Speaker 1: towards a solar battery economy, which is ultimately the that's
12:44
Speaker 1: it's where, that's where it's going to go anyway. But
12:46
Speaker 1: what tells us do is accelerate, accelerate that outcome. Sometimes
12:51
Speaker 1: people don't understand quite how much energy comes from the Sun.
12:55
Speaker 1: So the Sun is ninety nine point eight percent of
12:58
Speaker 1: all mass in the solar system being zero point one percent,
13:01
Speaker 1: and then zero point one percent is miscellaneous, Earth being
13:04
Speaker 1: in the miscellaneous category. So the total amount of energy.
13:08
Speaker 1: So some nice people, what say, will build the fusion
13:11
Speaker 1: reactors on Earth, it's actually the giant fusion reactor in
13:14
Speaker 1: the sky is basically impossible to beat to such agree
13:18
Speaker 1: that even if you could burn Jupiter in a thermon
13:20
Speaker 1: nuclear reactor, the amount of energy produced by the Sun
13:23
Speaker 1: would still round up to one hundred percent. That's how
13:26
Speaker 1: much energy the Sun produces. So solar power is necessarily
13:33
Speaker 1: the future, and I think there's going to be a
13:36
Speaker 1: lot of solar powered AI satellites, and I think Tells
13:41
Speaker 1: is going to play a role in that. We've obviously
13:51
Speaker 1: refreshed the product line so S three x Y. If
13:55
Speaker 1: if people haven't tried the well S three x or
13:58
Speaker 1: Y or the Cybertry, I recommend at least getting a
14:01
Speaker 1: test drive or a test ride as the case may be,
14:04
Speaker 1: try out the full self driving and I think you'll
14:06
Speaker 1: be blown away to those who do not. If you
14:09
Speaker 1: might be listening and don't have a Tesla, you should
14:10
Speaker 1: try one. And of course we've got the cyber Truck,
14:15
Speaker 1: which is the toughest truck of all time. It's literally bulletproof,
14:19
Speaker 1: faster than a Portion nine eleven and can out toe
14:22
Speaker 1: at Ford F three fifty.
14:25
Speaker 2: So it's a great car, great truck.
14:35
Speaker 1: And then starting next year, we manufacture the Tesla Semi.
14:41
Speaker 1: So this is already we already have a lot of
14:49
Speaker 1: prototype Tesla Semis in operation. PepsiCo and other companies have
14:53
Speaker 1: been using the a Tesla Semi for quite some time.
14:56
Speaker 1: But we will start volume production at our Northern Nevada
15:00
Speaker 1: factory in twenty twenty six. So we've got two big
15:03
Speaker 1: products or three three massive products starting production next year,
15:06
Speaker 1: we've got optimists, we've got Tesla Semi, and we've got
15:09
Speaker 1: the cybercap.
15:19
Speaker 2: And then battery packs. So the if you look at total.
15:26
Speaker 1: US power generation capability, it's roughly a tero WAT, but
15:31
Speaker 1: the average power usage is less than half a tarrawatt,
15:35
Speaker 1: and that's because there are big differences in power usage
15:38
Speaker 1: between day and night. So daily and seasonal variations in
15:42
Speaker 1: power consumption mean that the United States and really every
15:46
Speaker 1: country is only using about half It is only producing
15:49
Speaker 1: about half as much electrical energy as it could because
15:53
Speaker 1: without batteries, there's no effective way to buffer the energy.
15:57
Speaker 1: So what batteries actually enable is even if you don't
16:00
Speaker 1: build any incremental power plants, you could double the energy
16:04
Speaker 1: output of the United States just with batteries.
16:08
Speaker 2: This is a super.
16:09
Speaker 1: Big deal, and in fact, I think that's really where
16:12
Speaker 1: most of the incremental energy production in the United States
16:15
Speaker 1: is going to come from.
16:16
Speaker 2: It's literally batteries, so a.
16:19
Speaker 1: Bigger deal than it may seem. And then we've got
16:29
Speaker 1: we keep improving the battery design, so the megablock, which
16:32
Speaker 1: makes it really easy to deploy battery utility scale batteries.
16:37
Speaker 2: So we've just simplified and.
16:38
Speaker 1: Brought more of the components to be internal to the batteries,
16:42
Speaker 1: so you can just show up and drop off a
16:43
Speaker 1: battery and it works, and then hopefully without hopefully over
16:49
Speaker 1: time we will actually add more and more of the
16:51
Speaker 1: power electronics so that megapac will actually be able to
16:54
Speaker 1: output up to thirty five killer volts directly, so you
16:59
Speaker 1: won't need to up station.
17:00
Speaker 2: That's what I'm saying. You can just.
17:01
Speaker 1: Literally drop it off, like the way that a power wall.
17:04
Speaker 1: You just connect it to the house. The utility wires
17:06
Speaker 1: go on one side and the other side goes to
17:08
Speaker 1: the house mains and that's it. So we want to
17:11
Speaker 1: get megapacked to the point where you just literally take
17:14
Speaker 1: the utility wires and you plug them in and it
17:16
Speaker 1: just works. Then we've got We've also built the world's
17:19
Speaker 1: largest supercharger network, so we do a lot of things
17:22
Speaker 1: here at Tesla. That's the biggest supercharger network in the
17:27
Speaker 1: world by far, and ultimately you'll be able to go
17:30
Speaker 1: anywhere on Earth using a Tesla supercharger. And it's pretty
17:33
Speaker 1: close to anywhere on Earth, but it's it's going to
17:36
Speaker 1: be ultimately just anywhere.
17:37
Speaker 2: It'll just work anywhere.
17:39
Speaker 1: The supercharger team has done great work expanding that and
17:41
Speaker 1: improving the efficiency of the supercharger network, and in North
17:44
Speaker 1: America they did such a good job that the other
17:48
Speaker 1: car companies basically said that they'll just use the Tesla
17:50
Speaker 1: supercharge network.
17:51
Speaker 2: Okay, sounds good to us.
18:01
Speaker 1: It's always important to have it a slant on safety
18:03
Speaker 1: in the factory, so we continue to improve safety.
18:05
Speaker 2: For our factory workers.
18:07
Speaker 1: We care a lot about their well being, and like
18:10
Speaker 1: one way you can just tell if a company is
18:12
Speaker 1: a good company or not if you just walk through the.
18:14
Speaker 2: Factory or walk through the office and.
18:15
Speaker 1: Catch the vibe, and the vibe in the tails of
18:18
Speaker 1: factory is good. People are happy, and that's how you
18:21
Speaker 1: know it's a good company. We've also put a lot
18:30
Speaker 1: of investment into raw materials. We've built in South Texas
18:35
Speaker 1: and Corpus Christy the biggest lithium refinery outside of China,
18:40
Speaker 1: I believe. So it's going to it's starting off up
18:42
Speaker 1: with fifty gig what hours of lithium and we'll expand
18:46
Speaker 1: from there. So this is a very important to have
18:49
Speaker 1: in a worst case scenario, that we have the ingredients
18:52
Speaker 1: necessary to make a battery very important. And then we've
18:56
Speaker 1: got here on this site, the Cathode factory, which is
18:59
Speaker 1: just the sort of giant building about a half mile
19:03
Speaker 1: that way, and we're just making sure that we promised
19:06
Speaker 1: supply chain standpoint are resilient against any potential geopltical challenges.
19:19
Speaker 1: And then also at this factory we also make the
19:21
Speaker 1: forty six eighty cell, which is getting better and better,
19:24
Speaker 1: and that forty six eighty cell will be used in
19:28
Speaker 1: the cybercab cyber It is being used in the cyber truck,
19:31
Speaker 1: and we'll be used in the cybercab and also an optimus.
19:35
Speaker 1: So that's going well. But we continue obviously to get
19:39
Speaker 1: cells from many suppliers. It's kind of like the chip
19:42
Speaker 1: bad thing. Will take as many chips as our suppliers
19:44
Speaker 1: can provide us, but then beyond that, if they can
19:48
Speaker 1: provide us with any incremental cells or chips we have
19:51
Speaker 1: making ourselves or we get stuck. That's the Tesla semi factory.
20:00
Speaker 1: It's going to be pretty cool to see these going
20:01
Speaker 1: down the road and scale.
20:13
Speaker 2: Yeah.
20:14
Speaker 1: So that's the basic plan, sustainable abundance for all.
20:29
Speaker 2: All right, So.
20:32
Speaker 1: Yeah, I guess we can go into Queen now, maybe
20:39
Speaker 1: at the next annual shoulder meeting. We'll have optimists takes
20:42
Speaker 1: some of the questions. That'd be cool. So let's see Alejohn,
20:57
Speaker 1: thank you for your help.
20:58
Speaker 3: By the way, Yeah, please go ahead, is it?
21:09
Speaker 4: No, let me hold it.
21:10
Speaker 1: It's okay.
21:12
Speaker 5: What a year has been?
21:13
Speaker 1: Right?
21:13
Speaker 4: Just a year ago President Trump got elected on the
21:15
Speaker 4: same day. These twelve months have been a heck of
21:19
Speaker 4: twelve months. Thank you to the board because I think
21:21
Speaker 4: they had to weather quite some storms. Institutional investors, and
21:26
Speaker 4: obviously thank you to you, and we stand with you.
21:29
Speaker 4: I think we showed it and we will show.
21:32
Speaker 2: It at.
21:41
Speaker 4: So we would like to express a wish please. This
21:44
Speaker 4: is a nice venue. We love coming to the factory
21:46
Speaker 4: and I think it's actually great we're doing this at
21:48
Speaker 4: the factory. But there are thousands of retail investors who
21:52
Speaker 4: are crying because they cannot they come to me. I
21:55
Speaker 4: love taking the Mama, but I now give it to you.
21:58
Speaker 4: Please organize a bigger venue, a bigger than Berkshire, and
22:03
Speaker 4: we will do better than Berkshire.
22:04
Speaker 1: Okay, sure we can pay. Yeah, see, Tessa will be
22:11
Speaker 1: very bigger than Berkshire. Long term, it's gonna be naughty,
22:15
Speaker 1: all right. We'll do the next one were able in
22:17
Speaker 1: the yeah, like the yeah, maybe they're downtown arena or
22:22
Speaker 1: the soccer stadium or something like that, and get your
22:24
Speaker 1: security check.
22:25
Speaker 4: We want to make sure that you're safe.
22:27
Speaker 2: What if I just we.
22:31
Speaker 4: But we want to party. And if you can bring
22:34
Speaker 4: the other Elon companies there as well, let's just.
22:36
Speaker 1: Would really appreciate.
22:37
Speaker 2: All right, So that's good.
22:38
Speaker 4: Thank you very much for listening to.
22:49
Speaker 2: Elon as a father.
22:50
Speaker 6: I just want to first thank you for the everything
22:53
Speaker 6: you're doing in the world, especially freedom of speech.
22:55
Speaker 2: Thank you.
22:56
Speaker 6: And one question. Will you a Tesla ever consider FSD
23:03
Speaker 6: tied to an owner's account rather than a vehicle to
23:06
Speaker 6: encourage more frequent upgrades, provided a transfer is to only
23:09
Speaker 6: a brain new Tesla vehicle, while flustering brand loyalty. If
23:13
Speaker 6: the vehicle has sold the trade of about upgrade to
23:15
Speaker 6: another Tesla f S, the ownership would end.
23:21
Speaker 1: We have done that a few times. I guess we
23:23
Speaker 1: could extend it again. All right, we'll extend it. We'll
23:28
Speaker 1: extend it for at least another quarter and then play
23:30
Speaker 1: it by a year after that.
23:32
Speaker 7: All right, howdy Elon, congrats on not having to show
23:38
Speaker 7: up to work for free anymore.
23:40
Speaker 8: Yeah, we now have over a forty billion dollar war chest.
23:51
Speaker 7: We're cash flow positive and remain that. We know how
23:56
Speaker 7: you feel about FIAT already.
23:57
Speaker 2: Is it time to take a look at bitcoin?
23:59
Speaker 7: What's your belief on that? Also, you've hinted towards that
24:03
Speaker 7: there's a wheelchair accessible model in the works where you're
24:06
Speaker 7: referring to the Reboven, and if so, can we please
24:09
Speaker 7: speed up production to help the least fortunate.
24:13
Speaker 2: Sure?
24:13
Speaker 1: Yeah, obviously we need a vehicle that's big enough to
24:16
Speaker 1: fit a wheelchair wheelchair accessible.
24:19
Speaker 2: So I think that.
24:21
Speaker 1: That is the Reboven or Robust or whatever we call it.
24:24
Speaker 1: It's not like we're slowing down because we want to
24:26
Speaker 1: slow down. It's like we're spending in like gazillion plates here.
24:30
Speaker 1: But I do think it would be very cool because
24:32
Speaker 1: I think esthetically as well, it's just it's just would
24:34
Speaker 1: change the look of the roads and make it feel
24:35
Speaker 1: like the future.
24:37
Speaker 2: So it won't be long before we.
24:38
Speaker 1: Make that, but it's but since we do have optimists,
24:42
Speaker 1: the we've got cybercaf optimists and semi all next year
24:47
Speaker 1: it will probably have to be maybe the year that,
24:51
Speaker 1: you know, a couple of years from now or something
24:52
Speaker 1: like that. But we certainly will make a wheelchair accessible vehicle.
24:57
Speaker 9: Thanks Hi, so she I was privileged enough to attend
25:06
Speaker 9: the investor day, And you've also talked about the factory
25:09
Speaker 9: being a product, and I saw the dry cathode method
25:12
Speaker 9: and so forth, and I wanted to know the progress
25:15
Speaker 9: in that and also would often be working on that
25:17
Speaker 9: in the future production line in the future.
25:20
Speaker 1: Yeah, I guess the dry cathode man that's turned out
25:23
Speaker 1: to be a lot harder than we thought.
25:25
Speaker 2: It does. Looked like it's going to be successful and it.
25:27
Speaker 1: Will have some cost advantage relative to wet cathode. But
25:30
Speaker 1: if had to wind the clock back, I would probably
25:32
Speaker 1: have gone with wet cathode instead of dry cathode because
25:35
Speaker 1: it just turned out to be a lot harder to
25:36
Speaker 1: make it capable of high buluing production with super higher liability. Yeah,
25:41
Speaker 1: but we will be we will be scaling up battery
25:45
Speaker 1: cell production at Tesla and looking for cell production from
25:48
Speaker 1: ou suppliers as well, because we're going to be ramping
25:51
Speaker 1: up production very dramatically at Tesla now that we believe
25:55
Speaker 1: we have full self dropping, that we have autonomy solved,
25:58
Speaker 1: or at least or within a few months of having
26:00
Speaker 1: it unsupvised autonomy solved at a reliability level significantly better
26:05
Speaker 1: than human It's not that means it's time to ramp
26:08
Speaker 1: up production because the value proposition is.
26:11
Speaker 2: Now much greater than a regular car.
26:13
Speaker 1: The killer app really is for people, can you text
26:16
Speaker 1: and drive or can you sleep and drive? Can you
26:19
Speaker 1: can the car take you to your destination? Or do
26:21
Speaker 1: you need to pay attention and be and.
26:23
Speaker 2: Have to drive it?
26:24
Speaker 1: And before we allow the car to be driven without
26:29
Speaker 1: paying attention, we need to make sure it's very safe.
26:32
Speaker 2: But like I said, we're on the cusp of that.
26:34
Speaker 1: I know said that a few times, but we really
26:36
Speaker 1: are at this point and you can feel it for
26:38
Speaker 1: yourselves with the fourteen point one release. So the what
26:44
Speaker 1: we're going to try to we're going to push to
26:47
Speaker 1: expand vehicle production as fast as we possibly can. So
26:50
Speaker 1: aspirationally would aim to increase vehicle production by about fifty
26:54
Speaker 1: percent by the end of next year. So yeah, so
26:59
Speaker 1: that's it's very hard to increased production, but that's roughly
27:01
Speaker 1: I don't know, maybe we get I'm just guessing at
27:05
Speaker 1: an exit rate by the Internet year of around two
27:08
Speaker 1: point six two point seven million vehicles annualized production and
27:13
Speaker 1: then aim to get to maybe four million by the
27:18
Speaker 1: annualized rate by the end of twenty seven, and then
27:22
Speaker 1: maybe five million by the end of twenty eight. Those
27:24
Speaker 1: are rough those are our aspirational goals. So these are
27:28
Speaker 1: This is a gigantic increase in output, which means that
27:32
Speaker 1: the entire supply chain has to move in unison with
27:36
Speaker 1: that with that increase in volume, and the nature of
27:39
Speaker 1: producing a large, complex product is that it moves as
27:42
Speaker 1: fast as the least lucky, dumbest element in the entire system.
27:46
Speaker 1: And there's ten thousand plus items. But this, like I said,
27:50
Speaker 1: this really is a new It's not just a new
27:52
Speaker 1: chapter for Teesla. It's a new book, and that new
27:54
Speaker 1: book is massively increasing vehicle production and ramping up OPTIMS
28:00
Speaker 1: production faster than any things ever been wrapped up before
28:03
Speaker 1: in history.
28:06
Speaker 2: Awesome, Elon, thank you, Thank you from all of us.
28:09
Speaker 10: The retail shareholders really care and so echo the sentiment
28:13
Speaker 10: of a larger venue work can come as a testament
28:17
Speaker 10: to that. I was here, I was fortunate enough to
28:19
Speaker 10: be her last year and since then touring the factory
28:23
Speaker 10: talking to people that work here. I'm just a retail shareholder,
28:27
Speaker 10: but increase my holdings twelve times, So I know it's
28:30
Speaker 10: really engaging and gives people confidence, and so thank you
28:34
Speaker 10: for having us, and I do echo that I hope
28:35
Speaker 10: it grows. Last year, I'm the one that asked you
28:38
Speaker 10: about your well being and safety.
28:40
Speaker 2: Oh yeah, full life.
28:48
Speaker 10: It was a broad general question and it was before
28:52
Speaker 10: some of the uncertainty that's unfolded since then. And yes,
28:55
Speaker 10: I hope that had a positive impact and we all care.
28:59
Speaker 2: And so my question this year.
29:03
Speaker 10: You've talked recently about the most mind blowing product demo
29:07
Speaker 10: of all time and a shot at being the most
29:10
Speaker 10: memorable product.
29:11
Speaker 2: Oh yeah ever roadster. Yeah yeah.
29:15
Speaker 10: And I have patiently been waiting on my Founder's series
29:21
Speaker 10: Roaster Reservation on behalf of the Founder Series. Guys that
29:24
Speaker 10: have stuck in there, can we be unvited to this unveil?
29:30
Speaker 2: Oh yeah, sure, absolutely definitely.
29:32
Speaker 10: Okay, yes, all Founders series can be in.
29:36
Speaker 1: Yes, it's the least we could do, frankly for people
29:39
Speaker 1: that have along suffering Roaster Reservation holders. The I I
29:46
Speaker 1: feel confident in saying it will be the most exciting
29:50
Speaker 1: product on bail ever, and I hope that whether it
29:56
Speaker 1: goes well or it doesn't go well.
30:00
Speaker 10: Some them up here. Can I get the first one?
30:03
Speaker 1: I guess the it's according to who whoever put down
30:07
Speaker 1: the deposits in that sequence, So that's the but you'll
30:11
Speaker 1: get a very early one.
30:12
Speaker 2: And the new.
30:13
Speaker 1: Roaster is very much like it's not even the icing
30:15
Speaker 1: on the cake, it's the cherry on the icing on
30:17
Speaker 1: the cake. It's really it's not essential for sustainable abundance.
30:21
Speaker 2: Put it that way.
30:23
Speaker 1: But I do think there should be very cool technology
30:26
Speaker 1: in the world that is, let's wave beyond anything that's
30:29
Speaker 1: ever existed. And I think I'd like those even if
30:35
Speaker 1: I could never have access to those things, I'd like
30:36
Speaker 1: to know that they exist and see the future happening.
30:40
Speaker 1: So I think it will be inspiring to a lot
30:41
Speaker 1: of people. And just it's extreme.
30:44
Speaker 2: It's like.
30:46
Speaker 1: The coolest car if it even is a car that
30:49
Speaker 1: is ever that will probably ever exist. Yeah, Eilon, congrats
30:56
Speaker 1: on the proposal plan.
30:57
Speaker 2: That was amazing. The compensation question.
30:59
Speaker 1: Back to Chick, the chips will be the limiting factor
31:02
Speaker 1: to the future, so on chips and electricity are the
31:06
Speaker 1: two limiting factors.
31:07
Speaker 2: Yeah, and we got the.
31:07
Speaker 11: Energy and power and energy packs ready to roll. So
31:10
Speaker 11: I think on the chips. I think you said TSMC, Samsung,
31:13
Speaker 11: perhaps Intel fourteen A and several different sites are a
31:16
Speaker 11: couple of US factory sources, which.
31:18
Speaker 2: Is we're like everyone every day. Yeah yeah, And.
31:21
Speaker 11: So I guess the is there an open door in
31:24
Speaker 11: the future of investing directly into some of those boundaries.
31:27
Speaker 11: And second question, how big is a terraff factory? Can
31:30
Speaker 11: you put that in scale? And oh yeah, yeah, that's
31:34
Speaker 11: TERROR factory.
31:37
Speaker 2: Yeah.
31:37
Speaker 1: The thing is that we actually have agreed to buy
31:40
Speaker 1: all the chips that are made from the FAB, so
31:42
Speaker 1: it's basically a money printing machine for TSMC and Samsung.
31:46
Speaker 1: It's like literally the fast you make the trips are faster,
31:48
Speaker 1: we sent them money, but it's still not going at
31:52
Speaker 1: just fast enough. So that's why I think, as far
31:59
Speaker 1: as I can see, the only option is to go
32:01
Speaker 1: boild some very big chip bab and then you've got
32:05
Speaker 1: to solve memory and packaging too. But otherwise you just
32:08
Speaker 1: tap out at whatever the chip production rate is. And
32:11
Speaker 1: so I guess TERROR would be you'd want to say,
32:13
Speaker 1: it's got to be at least one hundred thousand way
32:15
Speaker 1: for starts per month size fab and maybe that would
32:20
Speaker 1: be one of ten in a complex. So will ultimately
32:23
Speaker 1: be a million wafers starts per month? Yeah, exactly. You
32:28
Speaker 1: can tell when it gets the giggle factor. That's probably
32:31
Speaker 1: a good sign that we're onto something special. But I
32:37
Speaker 1: wouldn't be surprised for long term, it's like a million
32:39
Speaker 1: wafers a month.
32:41
Speaker 2: Yeah.
32:45
Speaker 12: Hi, I'm a long term investor, so I've been holding
32:48
Speaker 12: the shares for twelve years. I worked at the Autopilot
32:51
Speaker 12: team development for ten years, and I brought my dad
32:54
Speaker 12: here from Brazil. We actually won the testlevision contest with
32:58
Speaker 12: the cyber truck b VideA. I hope you got to
33:00
Speaker 12: see that at some point. So thanks for the free mode.
33:02
Speaker 12: The why, and then my question is about obviously market expansion,
33:06
Speaker 12: but not only South America, Brazil, but also into Mars.
33:09
Speaker 12: Like with the upcoming rendezvous of Mars and Earth, what's
33:12
Speaker 12: going to be in a pillow where we're going to send.
33:14
Speaker 1: There, Optimist is going to play a big role Optimists,
33:18
Speaker 1: and I think tells the vehicles will play a big.
33:20
Speaker 2: Role on the Moon and Mars.
33:23
Speaker 1: So for a Moon base and a Mars city, tells
33:26
Speaker 1: the vehicles and Optimus robots are natural fit for building
33:30
Speaker 1: and operating a Moon base and a Mars city.
33:35
Speaker 2: So sorry tyber truck. Also yeah, yeah, it's cyber truck.
33:39
Speaker 1: You will need to drive around and pressurized vehicle if
33:41
Speaker 1: there's a person inside. But yeah, it'll be something cool
33:45
Speaker 1: the next level Moon buggy or my Mars buggy.
33:48
Speaker 13: All right, Hi, mister musk GOREV been a fan of
33:53
Speaker 13: yours since two thousand and six. I think I saw
33:56
Speaker 13: a piece in a popular science magazine. It was really cool,
33:59
Speaker 13: and I think I was fifteen at the time. Shareholders.
34:02
Speaker 13: Since ten years now, it's been really rewarding. It's changed
34:05
Speaker 13: my life, So thank you very much to you and Tesla.
34:08
Speaker 13: My question is regarding Tesla's mission for a sustainable future
34:13
Speaker 13: as it pertains to autonomy. I personally fully expect a
34:17
Speaker 13: deflationary period after unlocked by autonomy, basically both in moving
34:23
Speaker 13: humans and goods and whatnot, excluding Optimus even yeah, what
34:28
Speaker 13: efforts are is Tesla going to focus on to reduce
34:32
Speaker 13: the dollars per mile rough math? I think like thirty
34:36
Speaker 13: cents a mile would be really nice. Fifty cents is
34:40
Speaker 13: pretty good too, looking at inflation later on, and then
34:43
Speaker 13: I guess the second part of that, how low does
34:45
Speaker 13: it have to be for people to just stop buying
34:48
Speaker 13: cars like where it doesn't make economic sense to do that.
34:51
Speaker 13: I expect that to affect economies of scale and then
34:54
Speaker 13: further increase the pricing of vehicles they're on out. So like,
34:57
Speaker 13: how what is that node at which inflection happens.
35:02
Speaker 2: In terms of cost per mile?
35:03
Speaker 1: We do see a path with a lot of work
35:06
Speaker 1: to get below twenty cents a mile in current ear dollars,
35:09
Speaker 1: and I someone agree that there's things will probably be
35:12
Speaker 1: deflationary as productivity increases. Because you can think of money
35:16
Speaker 1: supply as being the ratio of goods and services to
35:19
Speaker 1: what's the growth in goods and services versus the growth in.
35:22
Speaker 2: The money supply.
35:23
Speaker 1: That ratio is basically inflation, and so if the output
35:27
Speaker 1: of goods and services grows faster than the money supply,
35:29
Speaker 1: then you necessarily have deflation or vice versa. They try
35:32
Speaker 1: to make it sound complicated, but it's not. I'm not
35:34
Speaker 1: sure that even government overspending can actually I think government
35:39
Speaker 1: overspending will not be will be lower than the increase
35:42
Speaker 1: in productivity of goods and services, which would apply to
35:45
Speaker 1: deflation as you expect.
35:50
Speaker 13: Sorry, what about economies of scale? And be impacted. Has
35:53
Speaker 13: to how you respond as an automotive company when people
35:56
Speaker 13: stop potentially buying cars.
35:59
Speaker 2: Economies, sorry, economies of scale in.
36:01
Speaker 13: Like auto manufacturering for instance, it only makes sense if
36:04
Speaker 13: lots of people buy your vehicles.
36:06
Speaker 2: Oh, the total number of vehicles will decrease.
36:09
Speaker 1: This sort of vehicle fleet out there is about two
36:11
Speaker 1: billion of cars, but to two billion if you add
36:13
Speaker 1: up old corson trucks that are not at a scrap yard.
36:16
Speaker 1: I think it's around two billion, but that number would
36:19
Speaker 1: decrease with autonomous vehicles, so the total fleet size would drop.
36:24
Speaker 1: I actually think miles driven will increase because it's now
36:27
Speaker 1: much less painful to travel somewhere. So if somebody's thinking
36:31
Speaker 1: about traveling across the busy city, then they'll take into
36:35
Speaker 1: account how much pain do I have to deal with
36:36
Speaker 1: If I have to do it through two hours of traffic,
36:38
Speaker 1: I probably won't won't do it. But if you're just
36:41
Speaker 1: say sitting in a cybercab watching a movie or doing
36:44
Speaker 1: some work, then it's just like sitting in a little lounge,
36:47
Speaker 1: And so I think you'll see probably a significant increase
36:50
Speaker 1: in total miles driven. But at the same time, a
36:52
Speaker 1: decrease in the total active fleet of vehicles.
36:58
Speaker 14: I'd just like to reiterate how grateful we all are
37:01
Speaker 14: to have you on board, ready to lead us to
37:03
Speaker 14: another seven trillion in market cap at.
37:05
Speaker 2: Least at least obviously thank you.
37:10
Speaker 14: My question is how much of a concern is it
37:12
Speaker 14: that when Cybercabs starts production in Q two next year,
37:16
Speaker 14: that regulation won't be there yet to where you can
37:18
Speaker 14: deploy cybercabs being produced. Or are you guys confident that
37:22
Speaker 14: every cybercab you guys make you'll be able to deploy.
37:26
Speaker 2: Yeah.
37:26
Speaker 1: I think the rate at which we'll we receive regulatory
37:29
Speaker 1: approval will roughly match the rate of cybercab production. It'll
37:34
Speaker 1: be maybe a little tight, but that's it's about right.
37:38
Speaker 1: And i'd like to thank Way more for paving the
37:40
Speaker 1: path here. It's very helpful. Yeah, but I think we'll
37:44
Speaker 1: be able to deploy all the cyberclbs that we produce.
37:49
Speaker 2: And the other thing is, once it becomes like extremely
37:53
Speaker 2: normal in cities, it's.
37:55
Speaker 1: Just going to become like the regulators will have just
37:58
Speaker 1: fewer and fewer reasons to say no.
38:00
Speaker 2: And then you've got this, You've got the.
38:02
Speaker 1: Accident statistics at scale, and you can show that what
38:05
Speaker 1: autonomous miles save lives then, and you've got unequivocal billions
38:12
Speaker 1: of miles to prove it, then I think it's hard
38:14
Speaker 1: for regulators to say no. Yeah, true. All right.
38:22
Speaker 15: First of all, Leland, thank you so much for everything
38:24
Speaker 15: you do for the Tesla shareholders. Do you see a
38:27
Speaker 15: path for optimists to have consciousness downloaded to it?
38:34
Speaker 1: You mean human consciousness or yeah, it's not immediate, but
38:42
Speaker 1: if you say it down the road, would you be
38:44
Speaker 1: able to say, with a neural link, have a snapshot
38:51
Speaker 1: of what is an approximate snapshot of somebody's mind and
38:56
Speaker 1: then upload that approximate snapshot to optimus body. I think
39:01
Speaker 1: that at some point that technology becomes possible, and it's
39:05
Speaker 1: probably less than twenty years. Yeah, but of course you
39:09
Speaker 1: won't quite be the same, be a little different because
39:12
Speaker 1: you'll be in a robot body and the mental snapshot
39:15
Speaker 1: will not be precise. It'll be a probably pretty close,
39:19
Speaker 1: but not exactly the same. On the other hand, are
39:22
Speaker 1: you the same person that you were five years ago? Nope,
39:24
Speaker 1: A lot of things have changed. Yeah, but I guess
39:27
Speaker 1: at some point, if you want to be uploaded to
39:30
Speaker 1: a robot body. My guess is that becomes possible.
39:35
Speaker 16: Yeah, hi, elon space based data center is a great idea,
39:41
Speaker 16: and I agree with that.
39:42
Speaker 2: Yeah.
39:43
Speaker 16: I was curious on your views on space based solar power,
39:47
Speaker 16: the idea of beaming down microwave energy down to Earth
39:50
Speaker 16: so that you can just point it to where it
39:52
Speaker 16: needs to go without transmission, without distribution lines, and it's
39:55
Speaker 16: more energy dense, so you know, less band use. Just
40:00
Speaker 16: curious on your views.
40:03
Speaker 1: Space solar power until the advent of something like Starship,
40:07
Speaker 1: where the cost per ton to orbit drops by orders
40:10
Speaker 1: of magnitude. The cost of getting Pale to orbit is
40:13
Speaker 1: so high that there was no possibility of space solar
40:16
Speaker 1: power solving anything.
40:17
Speaker 2: Really.
40:18
Speaker 1: Now with Starship, like I see a path to Starship
40:22
Speaker 1: cost per ton to orbit being.
40:24
Speaker 2: Lower than air flight.
40:26
Speaker 1: If you were to fly to a long distance airflight
40:30
Speaker 1: with cargirls, they took seven forty seven from here to Sydney,
40:34
Speaker 1: Australia or something like that.
40:37
Speaker 2: I think I think ultimately Starship.
40:38
Speaker 1: Will be able to do that trip for less cost
40:42
Speaker 1: per ton than an aircraft. So then you so now
40:46
Speaker 1: that now you've got that opens up a what very
40:51
Speaker 1: wide range of possibilities, like the most obvious one I
40:54
Speaker 1: think is actually solar powered AI satellites, So to move
40:59
Speaker 1: the AI to orbit and essentially deep space over time,
41:05
Speaker 1: because the you can actually access over a billion times
41:09
Speaker 1: more energy from the Sun in deep space than you
41:12
Speaker 1: can on Earth. The scaling to scale to crterships any
41:18
Speaker 1: to make any progress on a crdoship two scale, which
41:21
Speaker 1: is using some non trivial amount of energy from the Sun.
41:27
Speaker 2: You have to do space solar power.
41:29
Speaker 1: Now you could only beam a tiny amount of that
41:32
Speaker 1: back to Earth or you would melt Earth.
41:34
Speaker 2: So Earth actually receives a very.
41:37
Speaker 1: Tiny amount of the Sun's energy, where Earth is a tiny.
41:40
Speaker 2: Dust mote you see, like Earth to.
41:43
Speaker 1: Scale with the Sun, we look like a little crumb.
41:47
Speaker 2: So the to scale civilization.
41:52
Speaker 1: Like said to be at all relevant on a cordship
41:54
Speaker 1: of two scale, like to even use a millionth of
41:57
Speaker 1: a percent of the Sun's energy, you really have to
42:00
Speaker 1: be have your soul power in deep space and you
42:05
Speaker 1: could be in some of that back to Earth too,
42:07
Speaker 1: but you can only be a tiny bit of it
42:09
Speaker 1: back or you'd melt Earth.
42:13
Speaker 17: And this will be our last question.
42:16
Speaker 2: Well do okay, we'll do a few more we'll do
42:17
Speaker 2: a few more. Sorry, I'm sorry, all right, Elan. Look
42:22
Speaker 2: around you.
42:22
Speaker 18: This is a very small sample of Tesla soldiers. Maybe
42:28
Speaker 18: actually I should say, elin Musk soldiers.
42:31
Speaker 1: Thank you for your sport. So thank you everyone. I'd
42:34
Speaker 1: like to say thank you from the bottom of my
42:37
Speaker 1: heart for your support.
42:38
Speaker 2: Thank you, and thank you every one out there.
42:40
Speaker 18: So I know for sure a lot of them sold
42:46
Speaker 18: the farm to buy Tesla, and they stuck with Tesla
42:50
Speaker 18: through thick and thin, good days and bad days.
42:54
Speaker 2: Yes, So my.
42:56
Speaker 18: Question to you is, how do you feel about the thing,
43:00
Speaker 18: the success, but in a safer way where you get
43:04
Speaker 18: qualified Tesla retail investor only to invest in SpaceX, avoiding
43:10
Speaker 18: all these market manipulators, those crooks, and these short sellers.
43:15
Speaker 1: It is a tough problem that you highlight the basically
43:19
Speaker 1: the unfortunately, over time, the parasitic load of being a
43:24
Speaker 1: public company has just grown over time, and so you
43:27
Speaker 1: get all these spurious lawsuits, obviously, and they just make
43:31
Speaker 1: it very difficult to operate effectively as a public company.
43:35
Speaker 1: But I do want to try to figure out some
43:36
Speaker 1: way for Tesla shareholders to participate in SpaceX.
43:40
Speaker 19: That would be very cool.
44:04
Speaker 1: I've been giving a lot of thought to how to
44:07
Speaker 1: give people access to SpaceX stock, because I do want
44:10
Speaker 1: supporters to have SpaceX stock. But there's a sort of
44:14
Speaker 1: twenty five hundred shareholder threshold for before you become impacted.
44:18
Speaker 2: A public company.
44:19
Speaker 1: But I don't know, maybe at some point tells it,
44:22
Speaker 1: maybe made some point SpaceX should become a public company
44:24
Speaker 1: despite all the downsides of being public.
44:28
Speaker 20: All right, Okay, first of all, thank you very much, Elon.
44:34
Speaker 20: Just an amazing job you're doing, not only for Tesla
44:37
Speaker 20: shareholders but humanity in general. So we got a couple
44:41
Speaker 20: of questions. Many of us are asking, like, we want
44:42
Speaker 20: to know. Many of us think you are one of
44:44
Speaker 20: the most consequential business people honor of ever and we
44:49
Speaker 20: want to understand how do you think, For example, like
44:52
Speaker 20: a couple of one, how did you know in the
44:53
Speaker 20: master plant part two, this is like twenty eighteen or
44:56
Speaker 20: something like that, you calculated that point it takes six
45:00
Speaker 20: billion miles of FSD miles driven before maybe the world
45:04
Speaker 20: the regatory would prove it. And then the second is
45:08
Speaker 20: like when you were deciding if Tessa's going to have
45:11
Speaker 20: the hardware chips AI chips, did you know at that
45:14
Speaker 20: time that at some point when you get to AA five,
45:17
Speaker 20: it's going to be used not only for cars but
45:20
Speaker 20: for bots and for AI data centers.
45:23
Speaker 2: Was that luck a little bit or was it.
45:25
Speaker 21: Like you that was going to happen?
45:28
Speaker 1: I generally, I mean in terms of the estimates, like
45:31
Speaker 1: the for the self driving stuff, I generally try to
45:33
Speaker 1: get things to within an order of magnitude. So if so,
45:37
Speaker 1: it seemed to me like probably and this was technically
45:40
Speaker 1: in kilometers, but would be more closer to ten billion kilometers,
45:46
Speaker 1: which is roughly six billion miles. Then it would be
45:48
Speaker 1: to one billion, but I think I thought it would
45:50
Speaker 1: not take one hundred billion kilometers.
45:53
Speaker 2: So it's really just.
45:54
Speaker 1: Trying to when you're trying to guess something where there's
45:57
Speaker 1: a lot of uncertainty, just try to get it get
46:00
Speaker 1: the estimate to the nearest order mangitude closest factor of ten.
46:04
Speaker 1: And that's why I said probably around ten billion kilometers
46:07
Speaker 1: or six billion miles. And then yeah, the chip the
46:10
Speaker 1: reason Tesla created a chip team, And it's important though,
46:14
Speaker 1: like the vast majority of Tesla is like a dozen
46:16
Speaker 1: startups in one and the only we've only really done
46:21
Speaker 1: one major acquisition, which is solar City and then had
46:24
Speaker 1: some very small ones. So all of us is almost
46:27
Speaker 1: all of this is organic. I built the CHIP team
46:30
Speaker 1: from scratch and the AI team from scratch. The it
46:34
Speaker 1: was just because it became a limiting factor. Hardware too,
46:39
Speaker 1: we used in video, but in video was at that
46:42
Speaker 1: time focused on making really AI server hardware, which obviously
46:48
Speaker 1: was a smart bet. Though the most val currently the
46:50
Speaker 1: most valuable company in the world, and Jensen Long his
46:54
Speaker 1: team have done an incredible job at in video.
46:56
Speaker 2: My headsaft to them.
46:57
Speaker 1: I'm a huge admirer of Jensen and in video done
47:00
Speaker 1: amazing work, but they were they didn't want to do
47:03
Speaker 1: a cut a low cost, power efficient car computer at
47:07
Speaker 1: the time AI car computer. So it's like, okay, I
47:10
Speaker 1: guess we need to start a chip team to solve that.
47:13
Speaker 1: So then that's when I hired Jim Keller and we
47:16
Speaker 1: built the chip team and did AI three, then AI four,
47:21
Speaker 1: and then we to be honest, we made a few
47:23
Speaker 1: we made some mistakes there with AI five, but now
47:26
Speaker 1: AI five is back on track and we'll have a
47:29
Speaker 1: very rapid cadence to AI six and so forth, and
47:32
Speaker 1: this will really this is really necessary for for the
47:36
Speaker 1: car to like like With AI four, I think we
47:39
Speaker 1: can get to two or three hundred percent better than
47:43
Speaker 1: human safety, maybe four hundred percent better. But with AI five,
47:47
Speaker 1: I think we can do one thousand percent better or
47:50
Speaker 1: maybe even better than that, and then human safety at
47:53
Speaker 1: a certain point. Actually, it might be too much intelligence
47:55
Speaker 1: for a car because like I was thinking, like, what
47:57
Speaker 1: if you get stuck in a car and you have
47:59
Speaker 1: too much intelligence? But then one of the things we
48:02
Speaker 1: could do is when the car is idle, is use
48:05
Speaker 1: the car as a massive distributed AI infrince fleet with
48:10
Speaker 1: the content of customers. We're like, do you want your
48:11
Speaker 1: car to earn money for you while it's sitting in your.
48:14
Speaker 2: Garage of night. I don't know.
48:17
Speaker 1: We'll pay one hundred dollars a month or two hundred
48:20
Speaker 1: dollars a month or whatever the right number is if
48:23
Speaker 1: you allow Tesla to do AI inference workloads when you're
48:28
Speaker 1: not using your car. So that will also helped the
48:32
Speaker 1: AI in the car not get bored, because I imagine
48:35
Speaker 1: what if I got stuck in a car and then well,
48:39
Speaker 1: and the highlight of your day was driving. But they
48:41
Speaker 1: don't always want to drive, so then what do you
48:43
Speaker 1: do with the rest of the time. So I think
48:45
Speaker 1: Tesla could actually end up having the largest might end
48:49
Speaker 1: up having the most amount of AI inference compute.
48:53
Speaker 2: In the world. Like I think, like maybe if we.
48:55
Speaker 1: Had one hundred million car fleet, and at some point
48:58
Speaker 1: we may have more than a hundred million call fleet,
49:01
Speaker 1: and they'll have AI six AI seven, And if you're
49:04
Speaker 1: able to run a kilowatt of inference on one hundred
49:07
Speaker 1: million car fleet, now you've got a one hundred gigawatts
49:11
Speaker 1: of distributed inference with built in cooling and power electronics
49:15
Speaker 1: and distributed power. Probably the market's valuing that at zero
49:19
Speaker 1: point zero right now, so I guess, but it seems
49:21
Speaker 1: like an obvious thing to do if you've got distributed
49:26
Speaker 1: inference AI and you've got the power and the cooling,
49:30
Speaker 1: which is very difficult to do the power in the cooling,
49:33
Speaker 1: and one hundred gigawatts is a lot the average I said.
49:36
Speaker 1: The average power consumption in the US I think is
49:38
Speaker 1: around four hundred and sixty gigawats within that's the entire
49:42
Speaker 1: electric consumption of the US. So the if you do
49:45
Speaker 1: it one hundred gigawatts, that'd be a pretty big number.
49:49
Speaker 1: But yeah, it's basically something as a limiting factor and
49:53
Speaker 1: then we take actions to address the limiting factor.
49:59
Speaker 20: A quick follow up, Thank you for that very much.
50:01
Speaker 20: A request for you. So you guys just unveiled the
50:03
Speaker 20: cyber bear.
50:05
Speaker 2: Oh yeah, we'd like to. It's beautiful.
50:07
Speaker 20: We'd want you, guys or maybe do a cyber bull
50:10
Speaker 20: here in Giga Texas. My name is Herbert. I've got
50:13
Speaker 20: a brighter with Herbert channel on YouTube. And this is
50:16
Speaker 20: the Cyber Bulls. We are representing the Tesla Bulls and
50:19
Speaker 20: we stand with you, elon, but would it be cool
50:21
Speaker 20: to have a cyber bull right here in Texas?
50:24
Speaker 1: Like a cyber longhorn.
50:26
Speaker 20: Cyber longhorn.
50:27
Speaker 1: All right, we'll do a cyber longhorn for the factory.
50:31
Speaker 2: All right, good afternoon.
50:38
Speaker 22: I'm I'm very excited to ask this question. I've dreamed
50:41
Speaker 22: of giving away a Tesla for a very long time,
50:44
Speaker 22: and I finally wore Evy Jack down enough that they're
50:48
Speaker 22: willing to flip the bill for it. But it turns
50:51
Speaker 22: out to give away a Tesla, I have to have
50:54
Speaker 22: your permission to say we're giving way a Tesla, so.
50:59
Speaker 2: Just get away.
51:01
Speaker 22: You don't have to do anything Tesla does that. We're
51:03
Speaker 22: going to go through the normal channels. We'll buy it
51:05
Speaker 22: from a store that stuff if you.
51:07
Speaker 2: Can get totally cool. Yeah, yeah, that's certainly. You don't
51:13
Speaker 2: need my mission to give away a car.
51:16
Speaker 1: We'll take like maybe a couple more questions and call
51:18
Speaker 1: it a night.
51:20
Speaker 5: All right, Hilana, my name is Jonathan. You mentioned that
51:24
Speaker 5: the Roadster will have more tech than all the James
51:27
Speaker 5: Bond vehicles combined. Do you think this possible will be
51:30
Speaker 5: that any of that tech will make it into the
51:32
Speaker 5: current vehicle lineup? No, And to follow up on that,
51:40
Speaker 5: do you have any estimate of production or delivery timelines
51:44
Speaker 5: for the Roadster.
51:46
Speaker 1: So we're aiming for the product unveiled and will be
51:50
Speaker 1: of the Roster to which will be very different from
51:52
Speaker 1: what we're shown previously. That demo event will be April
51:55
Speaker 1: first of next year. I have some deniably because like
52:01
Speaker 1: I could say, I was just kidding, but we are
52:05
Speaker 1: actually tentatively aiming for April first for what I think
52:09
Speaker 1: will be the most exciting whether it works or not,
52:13
Speaker 1: demo ever of any product. And then I guess production
52:16
Speaker 1: is probably about twelve to eighteen months after that, So
52:22
Speaker 1: I think production is probably a year or so after that. Oh,
52:27
Speaker 1: I can't give away the secrets, but you won't be disappointed.
52:31
Speaker 1: All right, take one last question? Hey, all right, three
52:35
Speaker 1: last questions?
52:41
Speaker 2: Okay, hi Elan.
52:45
Speaker 23: I'm really excited about this a future of sustainable abundance
52:48
Speaker 23: that we're talking about. You're going to be saving a
52:50
Speaker 23: lot of lives with FSD, but the number of lives
52:53
Speaker 23: that would be saved and improved with this future vision
52:56
Speaker 23: you have is really inspiring and very exciting. So even today,
53:00
Speaker 23: you've mentioned though that in a post scarcity world, the
53:03
Speaker 23: role of money could diminish or become obsolete, given that
53:06
Speaker 23: much of today's power, including yours, is tied to wealth.
53:10
Speaker 23: Do you think achieving this abundance would require powerful people
53:14
Speaker 23: to relinquish their power? And how might we address resistance
53:18
Speaker 23: from those who hold power to make this vision a reality?
53:22
Speaker 1: I think actually, long term, the AI is going to
53:25
Speaker 1: be in charge, to be totally frank, not humans. If
53:28
Speaker 1: artificial intelligence vastly exceeds the summer human intelligence, it is
53:32
Speaker 1: difficult to mentionine that any humans will actually.
53:34
Speaker 2: Be in charge. So we just need to make sure
53:39
Speaker 2: that as friendly. Yep, thank you? Is that the question
53:45
Speaker 2: go ahead? Sure? Yes?
53:47
Speaker 18: Okay?
53:47
Speaker 24: Sorry, el okay, okay, yeah, I'm cybercat on X and
53:52
Speaker 24: I'm doing YouTube and the content creator. There's one thing
53:55
Speaker 24: I always heard about Tesla owners or people who want
54:00
Speaker 24: to buy Tesla complain about that is about super expensive insurance,
54:05
Speaker 24: right and the thing is Tesla I know Tesla also
54:09
Speaker 24: do insurance. However, it's not a cover every place. Based
54:12
Speaker 24: on Boston, I don't have too much options. It's super expensive.
54:16
Speaker 24: And the other point is the FSD is very safe
54:20
Speaker 24: right now. I've been using FSD for about four years.
54:24
Speaker 24: Right it's getting to a point it's like almost unsupervised. However,
54:29
Speaker 24: the insurance still does not take this into consideration. They
54:32
Speaker 24: don't ask you whether you have FFD or not, and
54:35
Speaker 24: they don't know how much you travel with FSD, and
54:39
Speaker 24: that is not a part of the risk prediction kind
54:42
Speaker 24: of thing. So I feel like, what's your thought about
54:46
Speaker 24: the insurance going forward, especially when we're getting close to
54:49
Speaker 24: the autonomy. What is PAS like either using your own
54:53
Speaker 24: with the external partnership, for example, there's a company called Lemonade.
54:57
Speaker 1: We need to not have the questions be super long.
55:00
Speaker 1: The Teslin insurance is trying to expand as as quickly
55:04
Speaker 1: as possible, but the regulatory structure for insurance is extremely
55:07
Speaker 1: complex and works on a state by state basis, so
55:10
Speaker 1: it's really somewhat of a racket, and the rules for
55:13
Speaker 1: insurance are different with every state.
55:15
Speaker 2: It's a very complicated thing.
55:18
Speaker 1: Yeah, so there's but I'm aware that insurance often is
55:22
Speaker 1: too expensive and doesn't take the right things into account.
55:25
Speaker 1: But so all I can say to that is, yeah,
55:27
Speaker 1: we'll keep expanding teslent insurance. When the car is operating
55:31
Speaker 1: as a cybercap Tesla will simply self insure, so that
55:34
Speaker 1: kind of solves that. But insurance is a yeah, a real.
55:39
Speaker 2: Pain in the neck for sure.
55:41
Speaker 1: But Okay, I do need to end this at some point.
55:46
Speaker 1: I'll take Yeah, one question, one question there.
55:50
Speaker 21: All right, thanks thelan for taking my question.
55:52
Speaker 2: I appreciate it.
55:52
Speaker 21: My question has to do with compute and what the
55:55
Speaker 21: build out or how much is necessary to train optimists
55:59
Speaker 21: and actually get them to a very household meaningful robot
56:04
Speaker 21: that can do things, and if the partnership with x
56:07
Speaker 21: A with XAI would help accelerate that.
56:12
Speaker 1: Yeah, there's there's a lot of training compute needed for Optimus,
56:16
Speaker 1: and because the AI tip in the robot is relatively weak,
56:21
Speaker 1: because it's really living in our power.
56:24
Speaker 2: You can make up with that with a lot.
56:25
Speaker 1: Of training to have a lot of training result in
56:29
Speaker 1: a very efficient AI that can run on a low
56:32
Speaker 1: power chip in the robot. So it is we will
56:36
Speaker 1: actually have to spend a lot of money on training.
56:38
Speaker 1: Ultimately it'll be like tens of billions of dollars on
56:40
Speaker 1: training compute, so it's a big number.
56:44
Speaker 21: What a partnership with Xai help accelerate that?
56:48
Speaker 2: Uh?
56:48
Speaker 1: Yeah, I think potentially, Yeah, I think that that could. Yeah,
56:51
Speaker 1: there's potential for accelerating that.
56:54
Speaker 2: Yeah. Did the Xai investment think get approved?
56:58
Speaker 3: No?
56:58
Speaker 2: Yes? Okay, okay, okay, okay, whatever. What it like?
57:04
Speaker 1: It's Teesla and some other company that I have been
57:06
Speaker 1: interested in, and it's like always quite complicated to do
57:09
Speaker 1: things has to go through a lot of hoops to happen.
57:13
Speaker 1: But I do think there's a lot of potential collaboration
57:15
Speaker 1: with Xai in the future and with space x A.
57:18
Speaker 20: Right, okay, this is the last one.
57:21
Speaker 17: Yeah, sorry, Thank you so much, Captain elion x. I
57:25
Speaker 17: support you. Thanks so much for everything you do. Very
57:28
Speaker 17: simple question. I'm from Israel. I don't represent a lot
57:31
Speaker 17: of people, but people do ask me and I'm going
57:35
Speaker 17: to ask you, I need a chance to have the
57:37
Speaker 17: app in other languages, like Hebrew for example. Some people
57:41
Speaker 17: struggle with me. I mean even the app is the
57:44
Speaker 17: app itself, Yeah, just the app.
57:46
Speaker 1: The app is not a Hebrew Oh and a lot
57:49
Speaker 1: of people don't speak English. Oh okay, shoot, I feel
57:52
Speaker 1: we had it in all languages. Okay, definitely the app
57:55
Speaker 1: needs to be in all languages, all right, all right,