Jeff Bezos just made his most audacious play yet — and it has nothing to do with same-day delivery.
Blue Origin, Bezos’s rocket company, filed a proposal with the FCC on Thursday for something called “Project Sunrise” — a constellation of up to 51,600 satellites designed to operate as data centers in orbit. That’s right. Not internet satellites. Not communication relays. Full-blown computing infrastructure floating between 500 and 1,800 kilometers above your head.
The filing argues that “insatiable demand for AI workloads” is outstripping what terrestrial infrastructure can deliver, and that space-based data centers represent “a complement to terrestrial infrastructure by introducing a new compute tier that operates independently of Earth-based constraints.” Translation: we’re running out of room on the ground, so we’re going up.
Blue Origin’s pitch centers on the built-in efficiencies of space: always-on solar power, zero land acquisition costs, no grid infrastructure headaches. The satellites would communicate through optical links and connect to Earth via Blue Origin’s planned TeraWave broadband service — another constellation that hasn’t launched a single satellite yet.
Here’s the reality check. Blue Origin has only flown its New Glenn rocket twice. TeraWave doesn’t exist yet. Gartner recently called orbital data centers “peak insanity.” And skeptics point out that the actual technology for running reliable compute workloads in orbit — with all the radiation, thermal cycling, and maintenance nightmares that implies — simply doesn’t exist at scale.
But this isn’t happening in a vacuum (pun intended). SpaceX filed for its own million-satellite data center constellation earlier this year, prompting Amazon to petition the FCC to block it. Now Bezos is jumping into the same arena with his own mega-constellation. The billionaire space race just got a brand new front.
For investors, the immediate implications are indirect — Blue Origin is private, after all. But the downstream effects are real. If orbital compute becomes viable, the winners will be companies supplying radiation-hardened chips (hello, Nvidia, which just announced space-rated GPUs), optical communications hardware, and launch services. The losers? Potentially the traditional data center REITs and power companies banking on decades of terrestrial AI buildout. This is still years away from reality, but the fact that two of the world’s richest men are betting billions on it means it’s no longer science fiction — it’s a business plan.
