SpaceX is recruiting engineers to build AI clusters in orbit

SpaceX is hiring people to build AI supercomputers in space.

On August 1, 2026, Elon Musk posted that SpaceX is recruiting “exceptional engineering & skilled trades talent to build & operate the most powerful AI supercomputer clusters on & off Earth.” The application process is deliberately blunt: email three bullet points demonstrating evidence of exceptional ability, plus your resume, to datacenters@spacex.com.

Why orbit: the power case

“Off Earth” is the phrase doing the heavy lifting. The case for orbital data centers is power. In a dawn-dusk sun-synchronous orbit, a satellite rides the day-night line and gets near-continuous sunlight, while a normal low orbit is eclipsed for 35 to 40 minutes of every 90-minute pass.

Musk has repeatedly argued that electrical power is becoming the binding constraint on AI deployment, and space-based compute is his stated answer to that constraint.

The catch: cooling

Cooling is the part that gets harder in space, not easier. Deep space sits around 2.7 K, but a vacuum has no air to carry heat away, so radiation is the only exit.

A two-sided radiator near room temperature sheds roughly 600 watts per square meter. That works out to about 1.4 square meters of radiator for a single 700-watt H100, and about 80 square meters for one 40-kilowatt rack. In orbit you end up building close to a square meter of radiator for every square meter of solar array. The power comes free; getting rid of it does not.

From science fiction to hiring round

The post pulled millions of views within hours. A job listing that would have sounded like science fiction two years ago is now just another SpaceX hiring round, complete with an email address and a resume requirement.

Data centers in orbit: visionary infrastructure play, or the most expensive way possible to run a GPU? Either way, someone is now being paid to find out.