Elon Musk's Technical SpaceX Update: Manufacture, Launch, and Operate AI satellites at scale

[00:00] I finally have some pretty interesting space news to share with you that SpaceX just dropped days away from their IPO this Friday. Elon Musk sat down and did an interview with employees in Bastrop, Texas, which is not far from Austin, in the Starlink factory that I, in fact, started covering their development back in 2022.

[00:22] But quite frankly, the first 10 minutes of this interview are, you know, Elon's sayings and words that I've heard him say many times before. So I want to get to the newer information.

[00:34] So if you want to hear him talk about a Kardashev-level civilization, if you want to hear him talk about how Starship, you know, is the holy grail for reusable rockets, and imagine if planes weren't reusable, yada, yada, yada, you can listen to the whole thing.

[00:50] But this is the part that I found most interesting. People probably struggle to visualize a little bit when you say, like, data center in space. Like, we're not going to slap engines on a building and fly it up there. Like, these actually look, like, pretty different.

[01:04] And so kind of walk through how you take something that's in a giant building on the ground and turn it into something that's functional in space. Yeah, I think it's pretty interesting. A lot of people don't actually know what the inside of a data center even looks like, right?

[01:18] Yeah. Yeah, and it's a, like, mythical place where the internet's in the cloud. Yeah, some people envision wires, some people envision boxes, but, like, effectively it comes down to a set number of chips

[01:30] and the things that we need to launch into space are actually quite small when we look at it. The more challenging part is figuring out how to get the power for it, and that's where a lot of what we've worked on for existing, like, Starlink technologies,

[01:45] the solar arrays are what we want to utilize that expertise to be able to build a satellite that can actually launch the critical components of the data center in space itself. We like to look at this and say like what is where's

[02:03] the actual engineering problem here and and it's really a combination of delivering power and then taking the waste heat and energy away and sending into the back of the space as you mentioned. Yeah.

[02:15] The AI satellite is actually much simpler than a Starlink satellite. A Starlink satellite has gigantic phased ray antennas. It's got parabolic antennas.

[02:29] It's got a lot of laser links. It's much more complicated than an AI satellite. An AI satellite is essentially a lot of solar cells, a radiator, and you still need some laser links,

[02:47] but you don't have all of the super complex antennas that you have on a solar link satellite. So, I mean, given the two, the easier one to design for is the AI satellite.

[02:59] Yeah. It's just a little bit bigger. It's bigger. It's bigger, yeah. I was like, so we've got, this is our AI-1, you guys want to walk us through. Yeah, so the first thing that we're really looking at here is, like,

[03:13] first you've got to make something compelling right. And we thought that the right place to charge is around the 150 kilowatt, like, peak power level. But as we look at the workloads with our experience with XAI,

[03:27] we get to actually see that we can also support about 120 kilowatts of average compute. What we're showing here is kind of a graph version of the version one of the SpaceX AI satellite, the AI1 I get at the core.

[03:44] And it seems like a reasonable place to start is 150 kilowatts peak power, 120 kilowatts sustained power. And to give you a sense of what does that actually look like in terms of the size of radiators,

[03:57] side of the solar panels the assumptions here are 250 watts per square meter for the solar array and about 1400 watts per square meter for the radiators so the radiators these are double-sided

[04:11] radiators are radiating both sides they're uh oriented maclages to the sun and uh and it's 1400 watts per square meter is a very achievable goal over time we think we can probably do above

[04:25] 150 watts per square meter and above 1400 watts per square meter for solar panels and radiators respectively. But this gives you like a pretty much what the satellite is going to look like.

[04:39] It's a lot of solar panels, radiator, and then everything else is pretty small by comparison. And these are like evolutions of of things that we have actually already launched in in our solar wind conservation to date. Yeah, that's that's really I think the cool part to me is that

[04:54] we're looking at solar technology that we already are going to use on the V3 Starlink vehicles. So I'm really excited to then just take those and make it bigger.

[05:07] Yeah part of what we want to convey here is that there not some magic that necessary that doesn exist for the AI satellites As you said, a lot of this is technology we've already made for the Stomach B3 satellites.

[05:25] So we don't think this is a super hard problem compared to things we already do. There will be quite something in the order of a terrible bit of connectivity of laser connectivity from the satellite.

[05:39] The 150 kilowatt peak power level roughly matches what say an NVIDIA GV300 rack would do. So a GV300 with 72 GPUs.

[05:52] It's C power, I think, is around 140 kilowatts. But it's rarely, it's almost impossible to get it to be at that speed power. A more reasonable operating envelope would be around 120 kilowatts average power.

[06:07] But it can peak up to 150. So it's basically thinking of it as a rack of compute in space. and then you can connect these racks of compute to either each other via the laser link

[06:22] or directly to the Stalink constellation. So you can close the link with the Stalink constellation and then Stalink can then send that data to the ground

[06:34] using the existing KA and KU antennas on the vehicle. It also has laser links to the ground as well. So, and this would not be at a particularly high latency.

[06:49] You know, we're joking about maybe being around 600 to 800 kilometers above the Earth. And light travels 300 kilometers per millisecond.

[07:02] So that's about, you know, three milliseconds away. It's not very far. Don't worry about that too much. Sometimes your water pick is going to go some high latency. Yeah.

[07:15] Lightning is pretty fast. Lightning is pretty fast. It's the only one. Yeah. I think the cool thing also is the radiators themselves are about the same size as the existing solar arrays for the V3 vehicle.

[07:29] Kind of in that realm where we're flying today. Yeah. I mean, they got about a 70-meter wing span, so they're fairly large. We're talking about building a lot of them. and putting them up there, but you say, like, space is in the name.

[07:45] Like, there's a lot of space up there. And so even when you're talking thousands or even, you know, up to a million satellites, you've got plenty of room to move around up there. Yeah, space is really big.

[07:57] So it's not like space is going to get crowded. Space is enormous. Like, if you zoom in close to the satellite, it looks big. But if you actually look at it relative to the Earth,

[08:09] these satellites are so tiny you can't even see them. So they're very, very tiny compared to Earth. And, I mean, we have about 10,000 star lengths in orbit right now.

[08:22] Yeah. We've got a pretty good idea of how to operate just really large constellations and do it safely now, right? We are the only operator that has any experience with that scale. it's a great thing that

[08:36] we have this background so we know how tightly we can pack the satellites and fly them safely. That's the number one goal when we look at the constellation. We're going to be building a lot of satellites and we're going to be building

[08:48] them here in Bastrop right? So we've got this which we're in that building kind of in the middle which we're in the middle right now. This is my first time here. The building is massive

[09:00] like you come around the corner you see it in the trees, you know, like, oh, wow, but we're about to kind of put this building to shame, aren't we? Yes, we're going to, in fact,

[09:12] we already have the solar manufacturing facility on the construction already, and then we will be building out the AI staff production building soon,

[09:24] and, yeah, so we expect to have the AI staff production, the solar production, and all of that operating at some reasonable volume by the end of next year.

[09:39] So if anybody wants to work on AI satellites, this is kind of going to become the hub of that. We're also, I mean, like right behind us, machines are humming. We're still making all of our user terminals for Starlink here.

[09:52] That's not going anywhere. In fact, we're turning on new production lines for new units, right? Yes. In fact, these are the new Starlink terminals, which we made in much higher volume than the current terminals.

[10:06] And also, we think there's probably going to be a few hundred million Starlink terminals out there. And then the Starlink direct constellation will connect directly to people cell phones and enable high communication directly from your phone to space All right We two limiting factors down We got Master Orbit we got putting solar in a few third ones chips

[10:32] Yes. So, at least in the beginning, we can obviously launch the chips that are already being made. So our current reference design is for NVIDIA Rubin chips, so either GB300 or Rubin chips.

[10:52] And we'll also have a reference design for GPUs, and essentially you can put up any existing chips into orbit. But the current industry seems to be

[11:09] It seems like it's gonna I don't know Get to maybe around a hundred gigawatts a year of AI compute but it That doesn't answer the question. Well, how do you get to a terawatt?

[11:22] That's why you need the terafap So we're looking at a step bigger. Yeah, yeah in order to get to the next order of magnitude you need a gigantic ship factory.

[11:35] To give you a sense of scale here, we expect that the TAR5 is going to be around 100 million square feet, which is 10 times the size of the Tesla Gigafactory Texas.

[11:49] And what, aside from just, you know, I'm going to need starship point to point to get from one end to the other, aside from just the size, what's going to make this unique difference from any other chip-building operation on the planet?

[12:04] Well, I think over time there's going to be a lot of technology evolution with the TerraFab, but fundamentally it's about scale. So even if there were no fundamental technology breakthroughs,

[12:16] you could simply scale the existing chip-making technology with a lot of difficulty to a terawatt of chip-out per year.

[12:28] If you look at it from the logic die standpoint, that's like having a billion chips per year with a kilowatt per radical. So it's a billion full radicals coming in chips, each doing a kilowatt, and then you're going to need a lot of memory to go with that.

[12:48] A lot of people today use them, I think, orbital data centers for like a decade away. Yeah, I think we want to try to give people a sense of the time frame. At least the time frame we're aiming for.

[13:02] I mean, people should take this with a grain of salt to some degree because this is just our best guess. So this is not a promise of what we'll do. This is what we are going to try to do and think we probably can do,

[13:16] which is to get to roughly an annualized rate of a gigawatt per year by the end of next year by in terms of space AI compute. And then aspirationally, scale that by an order of magnitude per year.

[13:32] So in two and a half years, hitting an annualized rate of 10 gigawatts a year of space, in two and a half years, maybe 100 gigawatts. And then depending upon what progress there is in chip making in the rest of the world

[13:46] and with the tariff lab, going beyond that to scale to a terawatt per year, which is 1,000 gigawatts. That's twice the electronic electricity consumption of the United States. I think there will be an appetite for that, but we'll see.

[14:01] It's a lot of satellites. I don't know what I was going to think about, but maybe a lot of simulations or something. So after we've, you know, worked through all the limiting factors,

[14:14] we've kind of topped out what we can do on Earth, what is the next step to, again, try and actually notch maybe some percentage points towards becoming Tardishev level two?

[14:27] Why stop there? Why pick small? So the terawatt actually is quite small. Why pick small? Why pick small? So there is, in order to get to another field of magnitude, so it's 1,000-inch from a terawatt per year,

[14:42] the only way that we can really see that you can achieve that is on the moon with a mass driver essentially where you do local production of portable tanks and radiators on the moon

[14:58] maybe you bring the chips from a boat you could conceivably make the chips on the moon and but you need most of the mass to be made on the moon so you don't have to transport it to the

[15:11] moon from Earth. And then because the moon has no atmosphere and only one-sixth Earth's gravity, you can act, you can get, you can accelerate the AI satellites into deep space without a rocket.

[15:24] So you can basically shoot them into space using an electromagnetic gun like a railgun type I mean just it basically a linear electric motor is the way to think about it I think we can show

[15:58] I mean, if that doesn't get you excited for the future, I don't really know what will.

[16:10] I'm far enough to see a mass driver on the moon. That would be very cool. Yeah. It's far for our future. Yeah, yeah. It would also mean that if we're bringing that amount of mass to the moon, it would mean that anyone who wants to go to the moon will be able to go to the moon.

[16:25] And I think that would be pretty cool. Yeah. I'm going to be jumping first in line to get up there. Yeah, I mean, everyone should have moved at least once. Yeah. Just once. Yeah. You can move there if you want. Just go live on the moon.

[16:38] And so there's some interesting information from SpaceX ahead of the IPO, as I mentioned, Friday, June 12th. Now, speaking of Elon Musk, another thing that I've been keeping my eye on is this plea essentially from a YouTuber who apparently built a set of like a moon.

[17:01] And it's somewhere in Austin. Of course, he can't disclose it because he's hoping that Elon will show up and do an interview with him. He's trying to interview him this week ahead of the IPO, and then it sounds like he's willing to embargo their interview until the IPO happens.

[17:21] I just think it's really interesting what he's doing. Number one, incredible set. I guess he had to fly to Ukraine to get these spaces as well. But it's like I'm not sure if it's going to work out.

[17:35] I think, like, he picked one of the busiest, craziest, most demanding, important weeks of Elon's life. However, you know, why not shoot for the stars? So it will be interesting to see if Elon actually commits to this interview, but it's been getting a lot of attention on X.

[17:52] A lot of people are saying, oh, my gosh, Elon, you should do it, this and that. Other people are saying, well, you know, you picked kind of a crazy week, and, you know, could this set a bad precedent if Elon does it? and then other people do extravagant things like this,

[18:06] and then everyone thinks that they should just get an automatic yes. I don't know. I think that it is a great lesson in the art of assuming the sell. And so I'm keeping my eyes on this one. I just wanted to let you guys know about this because if this happens,

[18:20] it could be a very, very interesting interview. However, I think that it is quite a long shot, quite a Hail Mary, before the IPO, trying to get it done before then.

[18:32] And, you know, maybe Elon will meet up with him after. I'm not sure why he's giving him kind of like a week deadline. But hopefully it works out for him. And then one more interesting thing I wanted to share with you from Prada, of all brands.

[18:49] They had teamed up with Axiom Space to work on, yes, a space suit. They wrote, on X, ahead of humanity's return to the lunar surface, Axiom Space and Prada unveiled the liquid cooling and ventilation garment.

[19:02] designed to be worn by astronauts inside the Axiom Extravehicular Mobility Unit spacesuit. This collaboration draws on Prada's expertise in design, pattern making, and advanced materials,

[19:14] resulting in a next-generation garment developed through advanced 3D modeling techniques that maintain cooling and ventilation while enhancing comfort during up to eight-hour spacewalks.

[19:26] So, there you have it. Prada, as we know, is an expensive brand. I've never had anything Prada because, number one, I wouldn't want it, quite frankly. I mean, this one I might want, but as far as a purse, I could care less.

[19:40] But I wonder how much this would actually sell for. Of course, you wouldn't need this unless you were trying to go to space, but it seems to me an interesting collaboration, and I just wanted to share this cool video with you.

[19:55] So thanks so much tomorrow. If you guys also want some more space news, and if you haven't heard, the Artemis 3 crew is going to be announced by NASA, so I will cover that as well. But I am just, you know, sitting here at home waiting to give birth, probably in like two weeks or so, and hoping that you guys are enjoying these news updates.

[20:20] I try not to just, I don't know, make pointless videos or force videos. I really want to share things with you guys that will be worth your time. So anyway, I hope that you guys enjoyed this update. Thanks so much for watching my channel, and I'll see you in the next video.