← All videos

Elon Musk Unveils Terafab AI Chip Project | Bloomberg Businessweek

Bloomberg Podcasts Published Mar 25, 2026 Added 5mo ago 11:58 383 views Open on YouTube ↗

Description

When Elon Musk took the stage on Saturday to unveil his plans to get into semiconductor production, he didn’t spare the superlatives. No one has ever proposed anything quite like what Musk calls Terafab. The project, as outlined, would be a massive operation to build cutting-edge semiconductors for artificial intelligence, robotics and space forays. He would not only try to take on the best chip manufacturer in the world — Taiwan Semiconductor Manufacturing Co. — he wants to do it at volumes far beyond the industry’s current capacity. The scale boggles the mind. The project would require something like $5 trillion to $13 trillion in capital spending, according to estimates from Bernstein analysts. That would fund 140 to 360 new factories, each making 50,000 wafers per month in order to reach the 1 terawatt of annual computing capacity he proposed.

Max Chafkin, Bloomberg Businessweek Senior Reporter, discusses how Musk, by far the richest person in the world, has accomplished what othe

Transcript

Read auto-generated transcript (2536 words)

Kind: captions Language: en Bloomberg Audio Studios podcasts radio news. >> You're listening to Bloomberg Business Week with Carol Masser and Tim Stenc on Bloomberg Radio. >> Well, Carol made you made this point on the call earlier today, Carol, that it seems like Elon's been a little quiet lately. >> We've missed Elon. >> Did you? But you just have to log into X and then you'll if you miss him, just log into X and look at what he's been [laughter] posting. Um, >> he's been busy. He has been busy. Uh he said his Terrafab project, it's a grand plan to eventually manufacture his own ships for robotics, AI and space data centers, will be built in Austin and jointly run by Tesla and SpaceX. Also breaking late Friday, a jury concluded that Musk defrauded investors when he disparaged the company back in 2022 in an effort to buy the social media platform for a lower price than his original $44 billion bid. >> All right, so let's get to it. Max Chaffkins in the house. Uh someone who knows all things Elon. He's Bloomberg Business Week senior reporter, co-host of the Everybody's Business Podcast, also author of The Contrarian, Peter Teal, and Silicon Valley's Pursuit of Power. He's here in studio. It's like we go from non-stop Elon [laughter] to not. And yet, there are some things going on. >> Yeah. I mean, this is potentially a gigantic mega merger between SpaceX and X, a massive IPO, and and this Terra Fab uh deal that Tim is referencing, you know, it it brings SpaceX and Tesla closer as well. So explain that. Go there. >> So what we heard over the weekend, Elon Musk had a big event in Austin. Uh he he loves to have these kind of nighttime party type launches and it's often uh very speculative, I'll say. So So the the latest one of these is they're going to build a massive fab that's a chip factory in Austin for chips that he is going to send SpaceX is going to send into data centers in space. So that would be a Tesla project with SpaceX managing the data centers, bringing these companies uh much closer together than they even are, even though they do share obviously a CEO. >> Okay, >> so >> makes sense kind of. >> Which part makes sense? >> I don't know. Are we [laughter] trying to convince ourselves that this >> they showed a rendering of a catapult lifting satellites from the moon into space? So it's it is very uh science fiction. I I will say on on Bloomberg Tech uh earlier today uh the um uh Ed and Caroline uh spoke to Karen Franchky, the head of global private equity investments at Fidelity. It's an investor in SpaceX. Uh she said conceptually it makes a ton of sense in the longer term to be thinking about data centers in space. The question of course becomes uh is it economical and frankly what we've been watching if we're investors is Starship. So, so let's we'll break down I want to talk about data centers in space first because I think a year ago if you would have said that you would have been laughed out of anywhere but now more and more people are talking about this and the feasibility here still I think people are like wait a second how does this work >> yeah I mean there is an extent to which people are talking about this in partly in part because Elon Musk a few months back started saying hey I think this is a thing I'm going to do and and uh you know a lot of people who pay attention to AI follow what Elon Musk says The the the thought process here is AI is growing very quickly. Lots of companies need data centers very quickly. You're also starting to see local opposition to data centers. You politically this is becoming this went from being kind of like a clear-cut jobs issue, I think, to an issue that now has some nuance where you're starting to see even Republican politicians, even President Trump in certain ways saying we need to build these things, but we can't have them driving up the electricity price and so on. So Elon Musk's pitch here is, hey, I've got the perfect location for your data center. It's outer space where the sun is always shining. And of course, he as as the owner, >> I will bring I will bring your data center to space for you. >> Plenty of room in space. And and of course, Elon Musk has a sort of pseudo monopoly in the launch business. Now his rockets uh the rockets that launch every day these um the the Falcon 9 rockets um do not carry you know would not be big enough for this plan and that's where Starship this this indevelopment rocket that he is trying to build would come in. Uh it's I would say that the development of Starship has been um you know uneven. They they they have yet to to sort of hit the milestones that they're talking about. They are a long way from achieving this project. We laugh, but it is Elon and he does do things that nobody could have imagined. >> But I still think, aren't there latency issues still with satellite communication and and how quickly you can get that data? >> Well, I you know, I'm not going to I'm not an expert in the physics of of data centers, but what what I will say is >> Starlink is amazing. We know that it works well. Okay. If you're if you're looking for challenges here, I I don't think the latency issue is the big one. The big one is that Elon Musk is proposing to become the world's largest manufacturer of semiconductor chips by a factor of I don't know like 10 or 20 or something. He he's he's proposing to build the largest semiconductor plant in the world in the history of the world bigger than all the factories put together and he has never met made a single one of these chips. So he's this is a company with zero exper like this is there's a lot of things that would have to happen for this to become real and and and yet you know Elon Musk is kind of a master at pointing way off in the future trying to find uh big opportunities opportunities that of course where there's where there's maybe a political constituency again talking about the the challenges of data centers and the fact that yes the AI industry is growing quickly and also there are you know there are challenges at Tesla This is a way for him to find growth categories at a time when the car business, the the the business that Tesla is actually in has not been growing. >> I don't know. >> And they've sent Teslas to space. Carol, >> all I can think of is there's a road. I know. I know. Okay. >> Tesla. It's working well, right? Like it's just it's just chilling. It's just chilling. Just orbiting. Yeah. >> Um can we go back to something that happened on Friday? And this [laughter] she's smiling. um a jury uh saying basically Elon Musk defrauded Twitter investors when he disparaged the company back in 2022 in an effort to buy it um for a lower price than the original $44 billion bid. Um is that significant? Does it matter? >> The reporting here from Bloomberg and others is that this may end up being worth as much as a couple of billion dollars. like that that that the the losses that that some shareholders incurred who sold their their Twitter shares early would be, you know, again, in the hundreds of millions, even a couple billion dollars. Elon Musk, of course, is worth a lot more than that. And um and what I'll say is what I think what is most kind of interesting here is this is the first one of these cases he's lost. He has he's constantly getting into trouble for things he says on Twitter now X. Um and even a couple of these have gone to jury trials and uh in general jurors have sort of bought an argument that his lawyers have made which is that Elon just says stuff and he he doesn't use Twitter like a normal CEO does. He just says what's on his mind. This one did not go that way. And and I think in part of the reason it didn't go that way because was because Elon was tweeting stuff out and then he was also at the same time pursuing a campaign in court to try to get out of the Twitter acquisition. So So I think that kind of complicated his what he was trying to do and and and in the end like he he tweeted something which is that the deal was on hold which was not true. The deal was not on hold because he had signed an agreement. It's not the first time that he's um been in a situation where he's said something uh about a deal potentially. I think back to the take private deal with with Tesla many many years ago. I mean, >> yeah, >> this was like what 2017 or something around then. But I look at Rich Go on the Bloomberg terminal. He's worth $641 billion, right? >> Uh >> a couple like is he does he pay what he pays a couple billion dollar fine? What like what's the what are the implications of of this? like what are the repercussions of this? >> There's there's going to have to be a whole process where investors who were harmed, you know, ask for damages and so on. So, this is going to this is going to take this is going to take some time. It's going to also be argued in court. Um, again, we're talking about potentially large amounts of money for a normal person, not necessarily large amounts of money for Elon. You know, it's probably worth saying this whole argument over, you know, Twitter was the Twitter and the bot problem, which which this was this was that issue at this case. You know, it ended up not mattering much. El Elon bought the company. He has now merged it with SpaceX and it's its valuation has has soared well, you know, well beyond the the original purchase price. So, so I don't know. It's it's one of these things where uh I I guess the the people who the the investors who sold are going to get some relief, although they they probably would have been better off holding on to their shares. >> There's another story out there, too, that XAI, that's uh of course Elon Musk's company, sending engineers directly into the offices of prospective corporate clients in an attempt to claw business customers away from rivals, OpenAI and Anthropic, just normal business. Is there something unusual about this or I don't know. >> You know, it reminds me a little bit of the p of the playbook. uh followed by Palunteer. Palanteer famously uses these forward quoteunquote forward deployed engineers in lie of salespeople uh to sort of help with customer integrations. That has been a big issue with Palunteer skeptics. They sort of see this as tantamount to a a a salesforce that is not being accounted for as a salesforce. So so yeah, I mean perhaps it raises some questions about the you know unit economics of what X is doing. On the other hand, this is a very competitive world. Like the the there are a handful of these large language model companies. They are all competing for the same hand, you know, big companies. It's not totally surprising that XAI would do this. I think if you're if you're taking sort of a skeptical point of view, you say, well, I guess the product isn't selling itself very well because you're having to send these very very valuable engineers into the field to kind of do something resembling a sales pitch there. Do we ever really know what data Elon Musk might have picked up from the government when he was head of Doge? >> I think we're going to spend a long time sorting out exactly what happened there. And there are there are continued cases and and and legal complaints and so on. And I and and I think one of the issues in sorting this out is that AI algorithms and these these large language models are are they're not black boxes, but they're pretty close to it. It's hard to know exactly what data data can leak into them in various ways. So, I I think this is going to be a a thing that, frankly, we spend a couple of years, maybe even longer, uh, talking about and and I I would not be surprised if, you know, should Democrats retake the House or or the Senate or both, that we may get some congressional hearings, we may have a political process. >> Come on, Max. Just ask AI. [laughter] >> You don't need to report on it. You can just ask the the bots. What do you have? >> If you tell the bot that you want it to answer a certain way, then it will. So, [laughter] >> give me this kind of answer, right? [sighs] >> Can't can't tell if um >> around the world around Elon. >> Yeah, I can't tell if we should be excited or scared about our our future. >> I don't know. Okay, >> TBD perhaps. Let's ask [music] Chad TBT. Hey, Max, thank you so much. Max Chapkin, Bloomberg Business Week, senior reporter, co-host of Everybody's Business Podcast. [music]

Related coverage