Google unveils Project Suncatcher on 4 November 2025, announcing plans to test Trillium‑generation tensor processing units (TPUs) in space and to launch two prototype satellites with Planet by early 2027 to validate optical inter‑satellite links, radiation tolerance, and clustered low‑Earth‑orbit operations. Early tests show Trillium TPUs survive proton‑beam radiation levels and a bench demonstrator achieved 800 Gbps each‑way transmission, while the proposed dawn–dusk sun‑synchronous constellation targets tight formations and high solar power yield.

The project focuses on technical hurdles—high‑bandwidth free‑space optical links, thermal management, on‑orbit reliability, and launch economics—with initial findings indicating radiation resilience and plausible launch‑cost trajectories; Google says “the core concepts…are not precluded by fundamental physics,” and will proceed with the Planet mission as the next milestone. The work also explores AI‑specific infrastructure and long‑term scaling toward gigawatt‑class constellations, subject to regulatory and launch availability.

• What will Project Suncatcher test in orbit near Earth? Project Suncatcher will test TPUs, free‑space optical inter‑satellite links, and clustered constellation operations in low‑Earth orbit.
• When will the prototype satellites launch and with whom? Two prototype satellites are planned to launch with Planet by early 2027.
• How did Google test TPU radiation tolerance on the ground? Google exposed Trillium TPUs to a 67MeV proton beam to assess total ionizing dose and single‑event effects.
• What orbit and formation will the satellite constellation use? The design targets a dawn–dusk sun‑synchronous low‑Earth orbit with satellites clustered within kilometers, often hundreds of meters apart.