Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
On Friday, when SpaceX filed plans with the Federal Communications Commission (FCC) for a million-satellite data center networkYou probably thought Elon Musk was having fun with us. But after a week, it is clear that he is very dead.
The most obvious part, of course, is the integration between SpaceX and xAI there went ahead on Mondayofficially bringing together Musk’s space and AI services in a way that makes sense if there is some kind of project to integrate infrastructure.
But even beyond integration, we’re starting to see the idea of ​​orbital AI data clusters — specifically, networks of computers operating in space — interacting with the real system. On Wednesday, the FCC approved the filing and set a schedule for public comment. It’s the usual pro forma section, but FCC Chairman Brendan Carr did something unusual file sharing on X. Throughout his tenure as chairman, Carr has proven himself eager to help Trump’s friends and punish his enemies – as long as Musk is still on Trump’s good side, the idea can go ahead without a hitch.
At the same time, Elon Musk has begun to explain the debate on public data centers. At a a new section On Stripe co-founder Patrick Collison’s podcast “Cheeky Pint,” which also featured guest Dwarkesh Patel, Musk made a big case for moving more of our AI computing power into space. Basically, solar panels generate a lot of energy locally, so you can reduce one of the biggest costs of running a data center.
“It’s harder to get on the ground than it is to get into space,” Musk said in a podcast. “Any given sun will give you five times more energy than what’s on the ground, so space is very cheap.”
Attentive listeners will notice that there is a slight difference in opinion here! It’s true that solar panels generate a lot of energy in space, but since energy isn’t the only cost of running a data center and solar panels aren’t the only way to build a data center, it doesn’t follow that it’s cheaper to do the whole thing in orbit, as Patel said in the podcast. Patel also raised concerns about using GPUs that fail during AI model training, but you should listen to the rest of the session.
Overall, Musk hasn’t let up, listing 2028 as the year to reach an orbital data center. “You can put my words, in 36 months but maybe closer to 30 months, the most suitable economic place to put AI will be land,” Musk said.
Techcrunch event
Boston, MA
| |
June 23, 2026
He didn’t stop there. “Five years from now, my prediction is that we will be deploying and operating annually more AI in space than there is on Earth,” Musk continued.
In fact, from 2030, the number of data centers in the world will be about 200 GWwhich is a trillion dollar infrastructure when you are just putting it down.
Of course, SpaceX makes its money by launching things around, so this is all very easy for Musk – especially since SpaceX has an AI company attached to it. And with the new SpaceX-xAI team headed for an IPO in just a few months, you can expect to hear more about the orbital space station in the coming months. It’s a technology company they are still pouring hundreds of billions of dollars into data center investments every yearthere is a real chance that all the money will not remain in the world.
