Elon’s TeraFab Could Reshape US Chipmaking!

Kind: captions Language: en We're going to cover more about what Elon said about AI uh safety in a show we'll do tomorrow, but today uh we're going to look at what he said about Terra Fab. I'm not going to read all this because we're going to listen to him say it. Terrafab must be built and gives an update on the Tesla and SpaceX R&D chip fab in Austin. We're a little worried that maybe at some point um chips from Taiwan would would not be available for who knows what reason. Um but >> any number of reasons [laughter] >> for some reason at some point chips may not continue coming here from Taiwan. >> Okay. and um it would and that would really make things difficult um without if if we didn't have any chips. So, so that's an important reason to have Terrafab. Then, then long-term there's just a scaling challenge where if you want to really scale AI both at the service, you know, within server centers as well as at for you know for edge compute for humanoid robotics and cars um you you kind of run out of capacity of with the existing fabs. So as it is the this all the fabs are running at max capacity. So so I think we we we need to make sure we need we need there needs to be uh certainty of future supply of chips um even if uh things become challenged challenging geopolitically. Um and then even if they weren't challenging geopolitically, there's a scaling challenge with, you know, it's quite difficult to scale chip production. Um and you really need the logic, the memory, packaging, the whole works, uh in order to continue scaling. >> I want to pause it there for a second. So I thought that he was very articulate explaining the two reasons. One, geopolitical risk. He's always been very good Tesla of uh localized supply chain. They would have done this regardless, right? because they that's what they've done for the cars just in case anything changes and they were proven right by doing what they did. All the other automakers were suffered because of geopolitical issues, tariffs and others. Now he says though that uh also there's a scale. We just don't have enough. Um what's your thoughts on those two answers? >> Well, I have a video coming out probably this week on Terrafab's secret strategy because I think you know I'm not completely done. We are not completely done with the research but it's going to be interesting. Uh all I'm going to say so far because we need to tie up some loose ends is that people I think make a mistake when they think Terrafab is a competitor to Nvidia. I think you need to look at the manufacturing logic of semiconductors. You have to understand the the value chain but also the infrastructure value stack that you need to build these things. And uh the direction that this is going I think and why Jensen is so friendly towards Terrafab is this is much more an infrastructure approach than an Nvidia competitor approach because when you look at the Nvidia value stack and what they need TSMC packaging all kinds of things you need to onshore these things into the United States and no one is really doing that at scale and so I I think the way to view terrafab is not so much as oh they're just trying to replace Nvidia chips or Broadcom or something. I think it's a much bigger and more foundational approach which matches also what SpaceX is doing with the satellites. They are not just competing with you know nebas or something with some uh neoclouds. They are building a foundational infrastructure. And what I mean with that is it's more like a it's more like a government project where all the other players can actually then buy little booths in terra fab to you know use a simplistic example. So I think it's much more going to be a huge infrastructure that then invites Nvidia and Intel and all kinds of other people to then move in and do their own stuff. That's how I think this is going because it's way too complicated for one company to do the entire stack. >> That's a very good point uh positioning or thoughts about that. Uh it is a massive investment trillions of dollars, right? And then we know that they've already partnered with Intel. We know that they've already partnered with Nvidia. That's what you're saying. Do you think that there will be other partners beyond these two? Yeah, I think uh I think there is not a real appreciation with the commentators and analysts. I mean there's some knowledge about but there's not a com sufficient appreciation for what these chips are. If you you know have a data center or hold a GPU in your hand. I think most people don't understand what that is. This is the most advanced piece of technology ever created by humans these chips and no single company can produce them. So it requires a absolutely massive infrastructure to get to this chip and someone like Nvidia is just one piece of the puzzle. Maybe the central brain of the puzzle but not the whole puzzle. They need massive things that are not Nvidia to make it happen. And so I think Elon understands this entire structure pretty well. Others might too, but Elon takes action and says, "Okay, what is actually the real value here? What is actually missing in this value chain to onshore?" Definitely not Nvidia. Nvidia is on shore. So what are all the other pieces required? The giant infrastructure right the packaging the actual fab but then the tech of the fab is also not really missing. So what is missing? Well terra fab is missing and I think that's how I view it. This is kind of invi it's like a city that is missing and now all the little merchants and everyone else can move in to create the actual economy which is the chip in the end. And so I think sure you need some anchor tenants like Nvidia and Intel, but you might get more. You might have Samsung, you might have all kinds of people than using this thing. And of course, I don't know the details as well as they do. >> I'm sure there are some secret economics, what really matters, where the real bottlenecks are, and that's exactly what they're going to build. And then let they're not going to try to compete with Nvidia where they can't beat Nvidia. They're going to try to build the pieces that Nvidia simply can't do and that no one else can do. And I think it's genius. It's it's not just a savvy business strategy. It's much more a visionary civilizational strategy. It's like, okay, we need to remove these bottlenecks. We of course we get infinity money for it, but that's just a side effect. We unlock this entire thing for United States, >> you know. Um it when will the stock appreciate when will the investors appreciate that Tesla and SpaceX are actually doing terapab now it's just a it's a project right now they announced the project I get that next year the R&D which we're going to hear Elon talk about the R&D fab that's already being built at Giga Texas and then this is a future but imagine if a startup launched and they go we got a partnership with Intel we have a partnership with Nvidia we have billions of dollars we are building it now and here's our plans and are valued at zero. >> Well, I I think but I I think uh I think there's good and bad news. I think the bad news is they will never anticipate anything because this is way too new and complicated and it's way too not formulated by Tesla uh and SpaceX and I think they're in no rush to formulate anything and so I think it's one of these extreme ho hockey stick inflection points like cyber cap like optimism I think it's just until the last point it's not going to be priced in because no one understands this stuff and once it happens it's going to click like oh my god they're going to make 10 billion. Oh my god, they're going to make 100. They're going to make a trillion. How much could they make? It's like, oh, infinity. Oh, oops. And then his stock goes up, right? But I think >> it's just too hard to wrap your head around. >> Yeah. >> And the company, we know the company, they're not going to do anything to precommunicate anything that they are not completely sure because why would they? They don't care. So, I think it's going to be one of these things where you and me, we understand these things. We carve it out more and more. concern like the crowd us will like put together the pieces but the company is not going to do anything. They're just going to build it and then it's going to start to work and then everyone's going to wake up and then they can model the actual business model. So especially with something complicated like this it's it's going to be much more hockey stick than oh we anticipated we do milestones we price it into the stock. >> Yeah I agree with you. Okay let's listen to Russ what he says. So it's either build terra fab or or fail to scale >> right >> those are the two options >> and how sorry go ahead >> how deep due to uncertain um you know uncertainty on the geopolitical front >> and how deep have you guys gone in designing the facility you know is this fully scoped is it sort of an outline at this point uh to to what degree do you have like a an actual project plan on dates and deliverables and what's going to be up and running when >> we've got an R&D line that we're building first. So I think it's kind of crawl, walk, run. >> Yeah. >> Um >> I >> So there's there's an R&D fab that we're building in Austin. Uh that's a collaboration between Tesla and SpaceX. Um that's it's at the Austin Giga Texas campus. So that's um and and that that's it's like a pretty big R&D fab. Um, so we have all the equipment on order for that and we probably will be able to I think we'll probably be able to make something useful by the end of next year. Not not at scale, but as Gwin said, coral run. We we got to try to, you know, at least figure out how these machines work, you know, like we don't know how they work. We do a voltage up line before. >> I saw that you guys unbelievable. you had some job openings for lithography people and stuff and you know obviously all roads currently go through ASML but you probably would want to diversify andor vertically integrate I think that you you've shown a lot of capacity to do that >> um so is that part of the play as well Elon like you're just going to have to make sure that there's vendor diversity so that if for whatever reason the weather conditions in Taiwan are best that >> Yeah. Um yeah there's it really is crawl walking. So the first step is can we make anything like try to figure out if we can make anything. That's the crawl part and and then now now try to make useful chips at scale would be kind of the walk walk part and then run would be like like now let's make them at massive scale. So, it's hard to say how long it'll take us to do these things, but I think we'll get at least to the crawl part by the end of next year. >> Okay. >> I mean, there are many questions ASML and things like that. It's very hard to I mean, there are some things that are just very hard to do or impossible to do outside the core vendor. So, I don't know exactly how they're going to do it, but I'm sure when you look at the whole infrastructure that there are massive shortcomings that are that require massive investments and I think that's what Terra is all about. It's not like some giant chip factory where Elon sits there and then GPUs plop out and then he can compete with everyone. That's already clear. It's going to be with Nvidia, with Intel and all these guys. Maybe with M Micron uh because that would be the vision, right? You produce everything right there. So I wouldn't be I wouldn't be surprised if we also hear about memory like if there's also a micron built out in there. >> Oh, for sure. It's it's already they already said that it's uh everything including memory >> which I mean I mean not just memory not just sucky memory. It's not going to be Tesla memory. It's going to be potentially micron memory like the cutting edge stuff. So the same way Nvidia moves in probably SKH Highix or maybe Samsung. I mean they or maybe multiple of them are going to move in. I think the plan here is and I think that's my paradigm. I think this is about true infrastructure and true cutting edge which requires the top layers in the world to actually be in there from memory from GPUs from uh fabs uh from packaging. So I think that's how I view it more and more. It's kind of a market. It's like a city.