Google Tests AI Chips in Space to Revolutionize Orbital Computing

Google is set to test its in-house AI computing hardware, Tensor Processing Units (TPUs), in space as part of Project Suncatcher. The mission will take place aboard SpaceX's Transporter-18 rideshare mission with Planet and aims to determine whether the TPUs can handle launch conditions and operate amid radiation and temperature swings in low Earth orbit.

The test is a crucial step for Google, which has been working on developing its Trillium TPUs. Prior testing of these chips found that they could withstand radiation levels above those anticipated over a five-year mission. However, managing heat remains a technical challenge due to the lack of standard air-based cooling in space.

Google is studying various thermal control methods, including heat pipes and radiators, with longer-term plans involving multiple satellites linked by lasers. A two-satellite communications trial is targeted for 2027, but the initial mission focuses on gathering engineering results rather than proving a commercial service.