Google announced Project Suncatcher, a research initiative to deploy AI chips in space using solar-powered satellites equipped with Tensor Processing Units. The company plans to launch two prototype satellites by early 2027 in partnership with Planet to test the hardware in orbit and explore scaling machine learning compute beyond Earth.
Google launches its first experimental satellite, MVP, on October 1 as part of Project Suncatcher, an effort to build orbital AI data centers powered by solar panels. The refrigerator-sized satellite, built with Planet Labs hardware and containing four of Google's custom TPU accelerators, will validate technologies for a future constellation of space-based AI infrastructure.
Google's Project Suncatcher aims to deploy machine learning infrastructure in space by launching TPU-equipped satellites into low Earth orbit, where they can harness constant sunlight for up to eight times more solar power than on Earth. The first test satellite launches soon via SpaceX's Transporter-18 mission in partnership with Planet, designed to evaluate how Google's hardware survives radiation, thermal extremes, and the physical stresses of spaceflight. Key engineering challenges include component durability, radiation resistance, and developing new cooling systems for space environments.