Orbital plans space data centres to power AI, seeks FCC clearance for 100,000 satellites

It sounds like science fiction, but US-based satellite startup Orbital is looking to set up data centres in space that may power responses given by chatbots as artificial intelligence (AI) drives an unprecedented demand for computing power.

Orbital founder and CEO Euwyn Poon told ET it plans its first data centre satellite launch next year on a shared payload. The company has filed an application with the Federal Communications Commission (FCC) seeking clearance to have 100,000 satellites in the low earth orbit (LEO) between 500 to 800 kilometres to deliver 10 gigawatts of compute to help process AI workloads.

Orbital plans to scale up deployment towards the end of the decade when the SpaceX-owned Starship will come online and significantly reduce the launch cost, Poon said.

“I actually agree that large-scale infrastructure, the kilometres-wide kind, is still science fiction in space today,” he told ET. “But what's practical now is what Starlink already demonstrates: you can put up small satellites — 10,000, 50,000, 100,000 of them — as Starship brings launch costs down. They’re large, yes, but not crazy large. So, the confusion is really about scale.”

Harnessing Solar Energy

Technology giants including Elon Musk’s SpaceX, Google and Jeff Bezos’ Blue Origin besides startups such as Starcloud (formerly Lumen Orbit) and Kepler Communications are looking at either building or facilitating satellite-based compute that could harness uninterrupted solar energy and eliminate the need for millions of litres of water to cool down processors. In fact, Musk has announced plans to launch one million satellites for AI data centres in space and has filed an application with FCC.
Poon expects orbital compute to be complementary to terrestrial ones for the next 10 years or so.

“If total data centre capacity today is on the order of 100 gigawatts on earth, we can offload a meaningful chunk. Our 100,000 satellites get to 10 gigawatts — and with multiple companies, that alone could roughly double today’s capacity,” he said.
Orbital’s compute in space will be primarily for inference, where trained AI models make predictions on new data, Poon said.
“We're betting inference will be the bulk of AI workloads going forward, because the models have reached maturity and consumers and applications are now discovering the use cases,” he said.

The satellites that will be used for AI data centres would be smaller — roughly the size of a fridge — from where the solar panels and radiators extend out. “It's really ‘server racks in space’ — fridge-sized units at scale,” said Poon who previously built a micro-mobility infrastructure company and sold it to Ford.