NASA administrator suggests putting data centers in space

Jared Isaacman wants to put data centers in orbit and use the sun as a free power plant. Say what you want about the feasibility, but at least the man is thinking about the actual problem instead of just yelling about it. Data centers are eating up electricity and water in towns across the country, and the political response so far has mostly been complaining. This is the opposite of that. It's an engineer's answer to a policy fight.

There's a real reason to like the instinct here even if the execution is a decade away. The AI race with China isn't going to be won by moratoriums and zoning board fights. It's going to be won by whoever solves the energy problem first. If we can get server farms off the grid entirely and running on solar power that never sets, that's not a gimmick, that's a genuine end-run around the tradeoff everyone's been arguing about. No more picking between cheap electricity for families and compute for AI companies.

That said, somebody should ask the harder questions before this becomes a talking point. Cooling a data center is not trivial on Earth, let alone in a vacuum. Launch costs, radiation hardening, latency, repair logistics when a server fails a few hundred miles up. NASA under Isaacman has leaned hard into "move fast" energy, which is refreshing after years of the agency looking more like a jobs program than a space program. But there's a difference between an inspiring idea and a funded plan with an actual timeline.

Still, we'd rather have a NASA administrator throwing out big swings than one playing small ball while other countries build the future. This is the kind of ambition the space race used to run on, and it's good to see someone at the top willing to say the quiet part out loud: we should be trying to win, not just managing decline gracefully.