A California startup plans to launch a constellation of up to 100,000 satellites that will run artificial intelligence (AI) workloads in space and the company, Orbital Compute Inc., has filed for this platform with the U.S. Federal Communications Commission (FCC).
When deployed, the 100,000-satellite data center would be capable of delivering 10 gigawatts of compute power without drawing on terrestrial electricity, land or water. Instead, the space data center would generate electricity via solar panels and use the vacuum of space to cool it.
Each satellite functions as a single high-density rack powered by 100 kilowatts generated by solar arrays. Each spans about 100 meters and weighs about two tons.
Orbital said the advantage of putting compute in continuous sunlit orbit offers cost advantages over terrestrial data centers while sidestepping grid constraints, permits and community impacts. All of these, the company said, are impacting data center growth on Earth.
"The demand for AI compute is outrunning what we can reasonably build on the ground — we're short on power, land, and water all at once,” said Euwyn Poon, founder and CEO of Orbital Compute Inc. “Space solves all three. Sunlight is constant, cooling is free, and there's no neighborhood to disrupt. We think the next generation of data centers won't be built in the desert — they'll be built in orbit."
Deployment schedule
Orbital plans to phase out the rollout of the 100,000-satellite constellation with a demonstration payload carrying a single GPU via a Falcon 9 rocket in 2027. The company will then launch Orbital-1 in 2028, its first purpose-built satellite.
This deployment schedule is fast considering the Los Angeles-based company was founded in 2026 and earlier this year raised $5 million pre-seed funding designed to accelerate the rollout of its data center satellite constellation.
