Google Launches Test Satellite for Space AI Chips

Google, an Alphabet company, plans to launch a prototype satellite next week for the first orbital test of its "Sun Catcher" project. This research initiative aims to explore hosting large-scale artificial intelligence computing infrastructure in space. Companies like SpaceX and StarCloud are pursuing plans to deploy data centers in low Earth orbit. This approach seeks to leverage near-constant sunlight for powering energy-intensive AI computing, bypassing terrestrial electricity supply limitations. The new mission, scheduled for launch as part of SpaceX's upcoming Transporter-18 ride-share flight in partnership with Planet Labs, will assess Google's AI hardware's ability to withstand launch forces, radiation, and extreme temperatures in low Earth orbit. Following the launch, a prototype satellite carrying Google's Tensor Processing Units (TPUs) will be sent into orbit. The objective is to evaluate whether AI chips can operate reliably in space, where cosmic radiation can disrupt electronics or cause performance degradation, including data errors known as "bit flips." Google had tested its Tensor Processing Units while running AI workloads at a University of California, Davis facility. However, the company emphasized that orbital testing is crucial for fully understanding hardware performance in the real environment. The company will also evaluate its innovative design for cooling power-intensive chips in the vacuum of space, where traditional airflow cooling is impossible. The design combines heat pipes and radiators for thermal management. Experts, however, stated this concept is still years away from commercial viability due to high launch costs, complex engineering constraints, and satellite production bottlenecks. Google clarified that the first "Sun Catcher" mission is designed primarily to collect orbital data and identify potential failure points, not to showcase an operational orbital data center. The company later aims to launch two satellites in 2027 to test high-bandwidth laser links essential for connecting future computing clusters.