Elon Musk is Fixing The Semiconductor Industry | Terafab Documentary | Kepler

Kind: captions Language: en We've all seen the promises of the future. Self-driving cars, humanoid robots, AGI doing everything humans did in the past. It looks incredible. Feels almost inevitable at this point. But there is a wall standing in the middle of this future. Everything depends on one critical piece of technology. A silicon chip. These chips process millions of data every second, letting machines make real-time decisions. Building these chips is one of the hardest and most complex manufacturing processes in human history. Right now, every chip factory on Earth is running at their maximum capacity. According to Elon, if you add every advanced chip made by every supplier on the planet today, it only covers about 2% of what his companies alone will eventually need. >> All of the rest of the output from Earth is about 2% of what we need. So, if you add up all the fabs on Earth combined, they're only about 2% of what we need for the for the terawatt project or terrafab project. So, >> 2%. Think about that. Existing supplies simply cannot make enough silicon for the future. Unless someone reinvents how these chips get built and how many can be built at once, this future of abundance is still very far away. >> Um and uh but there's there's a maximum rate at which they're comfortable uh expanding, but that rate is uh much less than we would like. And so, we we either build the terrafab or we don't have the chips. And uh we need the chips, so we're going to build terrafab. >> Elon decided to skip the traditional chip industry entirely. If the industry couldn't supply the chips he needed, he would build them himself. To see why this is such a big deal, look at how chips are normally made. Today, the semiconductor supply chain is scattered across the planet. Design the chip in America, manufacture it in Asia, test it somewhere else entirely, and ship it back home to put it in a car. It takes months and costs a lot of money. And if one link in that chain breaks, everything just stops. This new plan skips that system, putting everything under one roof. Terafab is a fully self-contained chip factory, which is vertically integrated. Raw materials go in one door and a finished shiny silicon chip comes out the other end. Design, manufacturing, testing, packaging, everything in one complex. Because everything happens under one roof, chips get built faster and cheaper and custom built for the things they'll actually be used for. >> Uh we will have all of the equipment necessary to make chips of any kind, logical memory. And we also have all of the equipment necessary to make the lithography masks. So, in a single building, we can create a lithography mask, uh make the chip, uh test the chip, make another mask, and and and have an incredibly fast recursive loop for improving the chip design. Uh to the best of my knowledge, this doesn't exist anywhere in the world. Where you've got everything necessary to build logic, memory, and do packaging, and test it, and then do the masks, improve the masks, and and just keep looping it. And we're not just going to do conventional compute in this. I think there's some very interesting uh new physics. So, this is going to really going to push the limit of physics in in compute. And we're going to try a bunch of wild and crazy things, which you can do if you've got that fast iteration loop. That I can't emphasize enough the importance of being able to make a chip, test it, change the design, do another one, and have that in a single building. I I think that our recursive improvement with that situation is probably an order of magnitude better than anything else in the world. We expect to make two kinds of chips. One will be optimized for uh edge and inference. So, that'll be used primarily in Optimus and in the cars. And then we need uh a high-power chip that is designed for space. Uh that takes into account more difficult environment in space, where you've got high power, you have high-energy ions, photons, you have electron buildup. It's It's a hostile environment in space. So, you want to design the chip, you want to optimize it for space. >> Fitting an entire global supply chain, one that normally spans three continents under a single roof, takes a very large roof. You don't build the the factory on Earth overnight. So, the project was split in two. Tesla took the research site first, breaking ground on a prototype lab in Austin, Texas. This is where the physical chip layout gets designed and the manufacturing process gets worked out. Once that process is ready, it has to run at scale. That is where SpaceX comes in, taking over mass production at scale. Out in Grimes County, Texas, northwest of Houston, they're building the full-scale Terafactory. When finished, it'll be the largest building on Earth by floor area, 100 million square feet, 1,736 football fields, 10 times the size of Giga Texas, 50 times the size of the Pentagon. The blueprint features a symmetrical pair of rectangular factory wings under a curved roof with a long central highway running through the middle of the complex. To keep these equipments from flying, the facility taps directly into the nearby Gibbons Creek Reservoir for cooling. The project starts with a $16.8 billion investment and is expected to scale to 119 billion dollars. If the only goal was powering self-driving cars or robots doing chores, you wouldn't need a factory this size. To understand why Terafactory is this large, look at the real target, 1 terawatt of compute. >> And um the for the space compute, my guess is that is uh the vast majority of the compute because you you're power constrained on Earth. That's why I think it's probably 100 to 200 gigawatts a year of terrestrial chips. Um and probably on the order of a terawatt of chips in space. >> To understand why a terawatt matters, look at how progress is measured for civilizations. It's called the Kardashev scale. A type 1 civilization harnesses all the power of its home planet. A type 2 harnesses all the power of its star. Right now, humans barely register for the galactic competition. Earth only captures about half a billionth of the sun's total energy. Terrestrial data centers face a great amount of problems like power, cooling, land, permits, environmental concerns, and yes, a shitload of protesters. Scaling up computer this much on the ground is not practical for the future. That is part of why SpaceX is leading the charge. A share of the chips built at Terrafab are meant for AI one satellites, AI data centers in space with 70 m solar wings flying around. In orbit, there is no power grid to worry about. There is constant 24/7 sunlight for energy and radiative panels in the vacuum of space to keep the chips cool. >> The cost of AI and deploying AI in space will drop below the cost of terrestrial AI much sooner than most people expect. I think it may be only two or three years before it is actually lower cost to send AI chips to space. The solar power you can you're going to get at least five or more times the solar power you get in space versus the ground because you don't have atmospheric attenuation or a day-night cycle or seasonality. Um and you're always normal to the sun. So, you're really maximizing the the the solar power at that point. And this this uh space solar actually costs less than terrestrial solar because you don't need heavy glass or framing to protect it from extreme weather events. Um so, as soon as the cost to orbit drops to a low number, it immediately makes extremely compelling sense to put AI in space. It it becomes a no-brainer, basically. Um moreover, as you go to space, you get increased economies of scale and things get easier over time, whereas as you try to put more and more power on the ground, you run out of space. And and you you you you start using up the the the easy spots, and then you get next-level NIMBY. Uh nobody wants the thing in their backyard. So, then so actually increasing power on Earth has becomes harder over time and more expensive over time, but in space, it becomes actually cheaper and easier over time. These these are very important points. >> There are a great deal of engineering challenges to face for this exercise, but if this works out, it'll be one of the most amazing achievements of humanity. >> What you're asking what's on your mind is what do you do after a Terrafab? >> Uh Don't think small. Uh well, yeah, good point. So we you know how do you get to a petawatt? Is is the obvious next question. And you get there by having an electromagnetic mass driver on the moon with robots with octomay and also lots of humans and with that you can send a petawatt. You can get a petawatt of compute and send that to deep space.