Elon launched Terafab. Largest chip factory on Earth. Or Galaxy?!

Elon launched Terafab. Largest chip factory on Earth. Or Galaxy?! Elon Musk announces plans to build Terafab, the world's largest chip factory. It's the largest chip manufacturing facility ever built by a massive margin. And for this project he's bringing together Tesla, SpaceX, and xAI all on one joint venture for one goal. They're going to build one terawatt of AI compute per year. That's to give context, every chip fab on Earth only produces 20 gigawatts per year. And so Elon wants 50x that. And so you're thinking to yourself, why does he even need this much compute? Well, he explains it here. This chart explains why we need to build the Terafab, because 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 Terafab project. We certainly want our existing supply chain expand as quickly as they can, and we will buy all of their chips. We need the chips, so we're going to build Terafab. Now in this presentation, he also gives context to his ambition with the Kardashev scale, and we're not even Type 1 yet officially. And so what his point is, is that the compute wars isn't about who has the best models anymore, they're about who controls the infrastructure to run a civilization at scale. Terafab's going to happen all within one massive building in Austin, and it's going to handle every stage of chip production all under one roof: design, fabrication, packaging, testing, and that'll let them iterate in days rather than months. And they have two chip lines, one that's doing inference chips for Optimus and Tesla vehicles, and another chip line that's going to be about harnessing compute in space. And because 80% of Terafab's compute is going to go into orbit, they're going to have solar-powered AI satellites connected by lasers running 24/7 with no grid connection. His projection is that space-based AI will become cheaper than terrestrial AI within the next two to three years. And when you really do the math, it really comes down to the costs for the tonnage that you could bring up to orbit. And since SpaceX controls 90% of the up-mass, they control that lever. Then he starts talking about the Moon mass driver. It's this massive electromagnetic launcher using the Moon's low gravity, and so they don't need rockets or fuel, and it's just shooting AI satellites into orbit. So a terawatt of energy is the start, but they can eventually go to like petawatts, which would be like thousands of times more. And now this is starting to sound like serious sci-fi, which makes sense because Elon always references Iain Banks the Culture series. Post-scarcity, galactic civilizations, humans free to do whatever they want. And so of course I started reading the Culture series to find out what this looks like. And if you read it already, can you please put in the comment and let me know, is this a good future? To be honest, for founders, just seeing someone with this level of ambition should be inspiring. I mean, realizing that compute is no longer about who has the best models and being able to play it all the way out to harnessing the Sun's energy in order to drive down the cost of compute. Nvidia owns the chips today, Big Tech owns the cloud, Elon is now trying to own this third layer all the way from Austin to orbit to the Moon. And I think now it's in everyone's interest to start reading Iain Banks the Culture series, because I think we're going to need the reference material.