Elon’s TeraFab: Build 1 Billion Chips a Year and Put Data Centers in Space. No Experience Necessary

Kind: captions Language: en Let's talk about Elon Musk, our favorite [ __ ] He said in Austin last week that he's going to start a chip production company in Austin. That's a joint effort between Tesla, SpaceX, and XAI, which is now owned by SpaceX, called the Terra Fab. The goal is to create a chip fabricator that can have all the equipment necessary to build chips of any kind. A very lofty and unspecific goal. His reasoning is that the semiconductor industry is too slow to keep up with surging demand for chips. And he said, "We either build the terra fab or we don't have the chips." And we need the chips. So, we build the terra. Sound logic. >> Perfect. I love it. >> For Elon, this is now my editorializing here. >> He needs chips to power robo taxi Teslas. They're humanoid robots called Optimus, which is a dumb name. uh terrestrial data centers and his proposed data centers in space and data centers of space was his reason for merging XAI with SpaceX. That was the whole idea of like wait, how do these two misfit puzzle pieces fit together? Oh, XAI needs data centers and SpaceX can make them in space. So, there's a perfect union, I guess. But really, the real reason the real reason for the merger was that XAI is burning through cash. like it literally just, you know, dumped fuel on it and set it on fire. And SpaceX actually is cash flow positive. And so basically merging it together is like it's like it's like taking a whole pile of manure and just putting it into a golden palace. And SpaceX is that golden palace. That's that's essentially what's happening here. And he always does this. His companies always intermingle in ways that public companies or only one of those public, but the other one's about to be public. In ways that public companies just just just shouldn't do. But that's what he does. He likes to [ __ ] where he eats and that's what's happening. But anyway, back to the terra fab. He wants to produce chips that can support a terowatt of energy production in in space data centers. How much is a terowatt? About one billion Nvidia B200 chips according to Barretts. >> One billion. One billion. This what Elon does. comes up with an extremely lofty visionary goal and and acts like everyone else is too stupid to do this and nobody's else has thought of this stuff and only he's thought of it and then basically uses that to continue getting investor money both from public markets and from private investors. >> Yeah. And so let's say you were to build a data center in space. Doesn't that just make like maintenance and upgrades or any repairs or replacement or upgrades of data or of chips something close to impossible? >> Yeah. So, let's let's let's talk about data centers in general and data centers in space. So, firstly, let me back up a sec just for earth living data centers, earth residing data centers. Bloomberg says that US data centers are expected to double energy usage by 2033. And they found that electricity costs are 267% more in a month compared to 5 years ago in areas near data centers. So this is sort of the genesis for why they need to bring data centers into space because nobody [ __ ] wants to live around a data center because of those reasons and others. It also creates like a toxic wasteland around it and it's ugly as [ __ ] And I know that Texas isn't known for its beauty anyway, but like it doesn't need help. No. So listen, I live in California. I can say that. I can say that. I can say that about any place in the world. Say for like Brazil >> and like I don't know, maybe New Zealand or something. Okay. But you were asking about why data centers in space. >> Yeah. A proposed solution to the data center problem on Earth is data centers in space which essentially has access to unlimited energy from solar um without damaging the planet and menacing society. So it seems like a good idea on the surface but it's not like in theory it's not impossible to put data center into space. There's a company called StarCloud that is proposing doing it with solar panels that are 4 kilometers long. Jesus. >> Four kilometers long. >> Yeah, >> that's like >> two and a half miles. About 2 and a half miles. >> 2.8 miles. Wait. Yeah. >> 2.4 miles. You're right. 2.4 miles. >> Okay. Yeah. >> Yeah. You were close. You were righter than I was. Anyway. >> Okay. All right. I'll take it. >> So, how the [ __ ] do you get that into space without causing a a 50 satellite pileup? I mean, I these are the kinds of questions that they have to solve because right now there are no rockets that can take something that big into space. SpaceX is working on a larger rocket that should be able to bring stuff like that into space, but it's still going to be a huge pain in the ass and and we don't know how far away that is. That rocket that SpaceX is building has been has seen some, you know, some struggles. Elon thinks that the Terrafab can start producing chips by early 2027, even though similar projects take a minimum of three years, you know. So, this is just typical Elon [ __ ] thinking that everyone else is so dumb and he's a narcissist and like nobody else is smart enough to speed up timelines even though there's unprecedented demand for chips. And honestly, like that's kind of been his MMO for forever since X.com. His idea is, I'm going to go into these big old dumb industries, bring in my intellect, and overhaul them. And and you know, that worked for X.com. It worked uh where it it revolutionized personal finance. And >> you're not talking about Twitter with Twitter that became X.com. >> No, this is Yeah. The one that merged with PayPal. >> Yes. Back in like the 90s, I think, or early 2000s. >> And same thing with the car industry and Tesla. I mean, granted, he he did wasn't a founder of Tesla, but the the idea of like, okay, these car companies don't get it. They don't see the value of this. They don't see the market for this. I see it. It has worked for him to do that. Even the space industry I mean that is like kind of the same it molds what I just described >> but this is semiconductor industry is different okay Nvidia is not the most valuable company in the world by accident you know not to mention >> even some of their contemporaries like Broadcom I mean Broadcom is I'm pretty sure as of today Broadcom's worth more than Meta it's literally like a one and a half trillion dollar company or more like these semiconductor manufacturers are are no joke surging demand creat hates that, but I get his need for more of it. But just the idea, the arrogance that oh, they're not moving fast enough. I could move so much faster, even though I've never done this before. It's just like, dude, shut the [ __ ] up. Like honestly, one like like one out of every 20 things he promises comes true. And he's been riding on the success of SpaceX for way too long. Okay? And and I give him credit for SpaceX, but like everything he promises, people I'm sure there's like prediction markets against his promises. There has to be. So anyway, so back to data centers in space. Elon thinks that you can put them there in four to five years. Jeff Bezos, Jeff Bezos, who also has, you know, Blue Origin thinks 10 to 20 years. Elon put out some images of his orbital satellites. This is like how he pitches things to investors to get them all, you know, to get them all aroused about the SpaceX IPO and and the potential for data centers in space and blah blah blah. He talks about it like it's already a done deal. You just need to build it. But every like all the science and everything is already in place and we just need to put it together and then it's good to go. >> You're right. And I I don't want to begrudge him that willingness to be like, "Hey, I have a lofty dream or a goal and I'm going to go after it." Because we do need people like that. We want to celebrate people like that. >> I I just I just think that he's a little bit more full of [ __ ] not just a little bit, a lot more full of [ __ ] than he realizes and that the world realizes about him. And that's where my sort of disdain comes from. It's like, yeah, yeah, okay. Like, why don't you follow through on any of the other promises you made like getting us to Mars and then talk about this other crazy [ __ ] But um one thing other things I want to mention about data centers in space is that the FCC put out like a application to to allow companies to apply to put these kinds of satellites which would basically be like data centers in low low earth orbit and they basically gave like 30 days for people to object to this which is a very abnormally short window and a consortium of astronomers objected to it. They were like, "Dude, this would ruin astronomy as a study because he wants to put a million a million of them in low earth orbit. So, this would ruin astronomy because there'd be too much [ __ ] in the sky." Number one, >> a million data centers in space. >> A million. Yeah. Um, it would pollute the orbit with space debris because what do you do when these 4 kilometer long solar panels break? You know, there'd be so much debris and cleaning that up is not easy or cheap. And then also another thing that I never thought of is that the launching of all the rockets that it would take to get to 1 million. We don't know what the pollution or the exhaust of those rockets will actually do to the ozone layer. And there are some early studies showing that that the launch of rockets is creating holes in the ozone layer which would obviously accelerate global warming. So this whole data centers in space thing may not actually be a solution. But you know we all know nobody gives a [ __ ] about the earth. they can save two seconds of latency uh you know off their Pornhub sex doll live stream or whatever. So I think that they're probably going to figure out a way to do it. I think 1 million is just like an ar like it's just just an arbitrary number but it is a but the idea of putting things more things in low earth orbit rather than like further away is a problem and the further away you put it the higher the latency is which creates problem a different kind of problem. Um, plus more costly, it's more difficult in terms of engineering and the science, materials, everything needed to make it real. I mean, I don't know what I'm talking about, of course, but I could just imagine the further away from Earth you want to build a zero gravity data center that needs to be kept at the correct temperature and that needs a constant energy supply. I mean, there's a bajillion things that could go wrong. Obviously, >> I think people always see like these beautiful images of the Earth from space, satellite imagery and whatever, and you just it just looks so beautiful. There's so much open space. The reality is like if you look at like satellites plotted around the Earth, it's insane. It looks insane already. And this is like with tens of thousands of satellites, you know? I I think SpaceX just got to 10,000. I can't even imagine with a million and they're like 4 kilometers long. But I do want to talk about Elon for a second because I don't want to sound like look he has a lot of contemptable parts of his personality and he embodies a lot of what's wrong with the world. The glorification and deification of founders of extreme wealth of being being completely inconsiderate of your fellow human being a bad father. >> Uh I mean there's just so many values that he does not possess. And anyone who's listening this thinking, "No, that's not true. You just don't understand, Elon. He really just cares about humanity. Like, prove that. How has he helped humanity? Tesla was supposed to be a green energy company. How's that going? They're not a [ __ ] green energy company. And you know it, whoever is arguing with me in my head. Like, they're not. That was a dream that he had that sounded cool. Oh, terraform Mars. Yeah. How's that going? >> It's not going anywhere. There's nothing he's doing to help humanity and data center. if he really cared about the Earth just with what I said, he would be like, "hm, maybe we should have put data centers in space because the ozone layer, but he's going to do it anyway." So there goes your theory, imaginary person that I'm arguing with. But aside from all that, that I have that aside there, >> I do admire his his willingness and ambition to take on projects, even if he's not always the first to get there, I think his his, you know, frenetic, impatient personality has accelerated the development of SpaceX rockets, the Tesla Supercharger network, but >> so much of his company's innovations are driven by him. At a company like Google or Amazon, innovation is built into the culture of the companies. And so you know they know that Sergey Brin cannot lead the innovation of the entirety of Alphabet. They build it in through the culture, through the management, through the performance incentives. And he has never figured that out. That's why he's never really had a company that manages without him. It's literally all him. Every executive in place is just there to do what he thinks they should do. They don't make their own decision. They don't create their own culture of innovation. And I don't think he's ever really learned how to build a company where he's not the driving force. And that's why this terra fab project, the Optimus robots, the space data centers that honestly I and I'm not rooting for this necessarily, but I think they're all going to fail >> because I think even with somebody who has a frenetic, impatient, do anything to make it work, workaholic personality like him, plus access to basically unlimited capital. You can't be good at that many things. and his track record precedes him. Yes, SpaceX is a success. >> What else has been a success? Don't say Tesla because I will tell you, you're flat out wrong. >> If Tesla didn't have If Tesla didn't have the tax credits that they sell to other automakers that is now going away because of the Trump tax bill, they would have lost money last quarter. Lost money in the red. >> Okay. So Tesla can't be a part of the argument in terms of >> like for the plus side of the >> doing well. They were doing well. They were a great company. They had a great brand. They had a lot of profitability. And then he [ __ ] it up with his with his insufferable personality. I mean that is literally what happened. They basically boycotted Tesla in Europe. It was a beloved company even, you know, I think until he showed his >> eared, he burned a lot of goodwill because he just had to get involved with the second administration and >> yeah and really like show who he is. My last point on this is that when this does fail, which I think it will because of what I said, not because he's not a great entrepreneur, not because he's not innovative, but because he is always the driving force of innovation and that does not scale. not the way that he thinks it does. Not not across Tesla, across SpaceX, across XAI, across Neurolink, across whatever boring company. >> You can't do Infinity things and be equally good at all of them, I guess. >> And it's not going to be an obvious failure when it does happen because he'll change the narrative just like he did by shrinking the ambitions of SpaceX. Like I said, we're not going to Mars anymore. We're going to terraform the moon. Okay. >> Is that something he said they were going to do? >> Yeah. Yeah. Yeah, he said the moon is just a lot more achievable in a shorter time span. I'm paraphrasing. And so we should we should do it on the moon or like he did with Tesla, not a car company anymore. Oh, an AI robotics company, lol. Which I put in my notes, lol. That's literally what I wrote. Um, and you know, he has to do these kinds of things to justify Tesla's ridiculous valuation. So if he says, "We're going to build a million Optimus robots," which is a dumb name. I'm going to keep saying it's a dumb name. Is there even a is there even like a defined use case for these robots? Do we know who's gonna [ __ ] buy these? No. Yeah, he just wants to build them because that's what he needs to trigger certain uh compensation incentives in his in his contract with with Tesla, right? That's that's literally what it is. And Tesla's ridiculous valuation is down almost 20% year to date. And that's like when I wrote my notes, probably down even more. Market selling off like crazy today. And so he just needs to keep coming up with these stories of moonshot bets that the Tesla's pursuing even if they never pay off. I mean, he used to talk about Tesla replacing the power grid with home charging. >> No one knows what happened to that Tesla fanboys.