It was up to Elon Musk to revolutionize the automotive industry with Tesla and the electric car. Probably no one believed he could do it. Then he did the same with the aerospace industry with SpaceX, and that was more of the same: it seemed impossible. It may be many things, but the truth is that although Elon Musk promises many things and does not always fulfill them when he says (hello autonomous car), has achieved unimaginable things. That’s why when you talk about Terafab, maybe we should give it a chance. Because this seems almost as impossible as his other feats.
Terafab and Musk’s master plan. On Saturday night, from a power plant that has not been used for a long time, Elon Musk advertisement the last of the components of its master plan: Terafab. The objective is to create a chip factory in which Tesla, SpaceX and xAI will collaborate. According to Musk, this plant will be capable of manufacturing between 100 and 200 GW of computing capacity per year on earth, but it will reach 1 TW in space. The problem, as always with Musk, is distinguishing what part of the plan is engineering and what part is theater and fireworks.
He doesn’t do it just because. At that event, the magnate explained that semiconductor manufacturers do not produce enough chips for their AI and robotics needs. And since TSMC and the rest of the manufacturers cannot meet Musk’s demand, he has proposed manufacturing them directly. You need them for your robotaxis and your humanoid robots, Optimuswhich he hopes will end up multiplying by 10 or 100 the production rate of his cars. But it also needs chips so that xAI can compete in the field of AI, and SpaceX needs them for its satellites. That is, it actually needs a lot of chips. Many.
Chips from space. At Terafab they intend to create two types of chips. On the one hand, there will be those intended for autonomous vehicles or Optimus robots. On the other, the chips that already have their own name, D3, and that will be designed specifically for space, with products that use them that work in low Earth orbit and are powered by solar energy. For Musk, the idea “becomes an obvious decision”: there will come a point where putting payload into orbit is so cheap that host data centers in space It is cheaper than doing it on land because solar energy is practically unlimited there.
Too many unknowns. Everything was very nice and promising, but once the speech and promises were over, the questions began. Building a state-of-the-art semiconductor factory is a colossal challenge. It’s not just a matter of money: it’s that advanced chip manufacturing is in the hands of three companies around the world (TSMC, Samsung and Intel), and requires photolithography with UVE technology which is only manufactured by the well-known Dutch company ASML. And here’s the thing, that Musk:
- Did not announce any agreement with ASML
- It has not shown orders that demonstrate that it will have these equipment
- He has not named a technological partner for the project
- No estimated dates or calendar have been given.
- And he hasn’t talked about the budget either.
It’s all a gigantic unknown.
The most ambitious vertical integration in tech history. On several occasions Musk repeated how at Terafab they intend to cover the entire development, manufacturing, packagingtesting and improvement in the same facilities. If we fulfill that promise, we would be facing another unprecedented achievement, because the semiconductor industry has been doing just the opposite for decades: hyperspecialization by different suppliers: some design, others manufacture, others package… Musk wants to do it all, and if he succeeds he will become a direct rival for Samsung or TSMC, which a priori he would no longer need.
Promises and realities. This project seems especially diffuse, but with Musk anything is possible, as we have said. In recent years, yes, we have seen how several of his ideas or they have failedor they have been delayed, or they have been left in no man’s land. The robotaxis still haven’t arrived, the Cybertruck arrived late and it’s not settingand companies like The Boring Company or products like Solar Roof have had less reach than they promised, at least for now. Terafab seems like another impossible project from Musk. We’ll see if it ends up not being so.
Image | tesla
GIPHY App Key not set. Please check settings