American company founded in 2024 to pursue orbital data centres. Its boss is Philip Johnston. In November 2025 Starcloud launched Starcloud-1, a fridge-sized satellite containing an Nvidia H100 GPU, into space. It used the satellite to train a small AI language model, NanoGPT, on the works of Shakespeare, and to run Gemma, an open-source large language model made by Google—providing data on the reliability of AI chips under orbital conditions.
Starcloud is aiming for a specific power of 70W/kg for its forthcoming satellites, based on conservative assumptions. At that level, and with satellite costs of less than $5 per watt (excluding GPUs), the firm estimates an orbital data centre becomes cost-competitive with terrestrial equivalents—assuming launch costs fall to around $200-500/kg with fully reusable rockets such as SpaceX's Starship.
Starcloud-1 could not run its GPU continuously owing to overheating. The firm plans to launch Starcloud-2 to test an unfolding radiator design—what it says will be the largest commercial deployable radiator in space, providing ten times as much heat dissipation per kilogram as the International Space Station's radiator.
If some people didn't tell you, you'd never know they'd been away on vacation.