On 31 October 2026, Google announced that its Project Suncatcher prototype satellite had run an artificial intelligence (AI) model in orbit and returned a complete response to Earth. CEO Sundar Pichai called it a “hello world” milestone. The satellite was developed with Planet and launched on 1 October aboard SpaceX’s Transporter-18 rideshare mission.
In-orbit AI test
During a ground-station pass, the team sent a prompt to Gemma 4, an open AI model running on a Tensor Processing Unit (TPU) aboard the satellite. Asked to write a haiku about space, it returned: “Blue marble below / Data streams in silent grace / Starlight feeds my core.” The response was received as the satellite passed over Svalbard, Norway. Google had also tested space-themed limericks and prompts about the satellite’s technical status.
Announced on 4 November 2025, Project Suncatcher examines the feasibility of scalable machine-learning infrastructure in space, where satellites could use near-constant sunlight for power. The next phase will test TPUs under orbital thermal stress and radiation exposure. Google plans to launch two interconnected satellites in 2027 to test optical links for larger AI workloads. The company has said satellites in low Earth orbit could generate up to eight times more solar power than on Earth.
Important Facts for Exams
- TPU architecture: Google’s TPU is an application-specific integrated circuit neural-processing unit developed for neural-network machine learning.
- Low-precision computing: TPUs are designed for high volumes of low-precision computation, delivering more input/output operations per joule than graphics processing units.
- Satellite power storage: Solar-powered satellites also need batteries because sunlight is blocked inside the launch vehicle and during periods of orbital night.
- Laser mesh links: Laser inter-satellite links can connect spacecraft into a space-based optical mesh, an approach used by satellite broadband constellations.
