Google to Launch First Project Suncatcher Orbital Data Center Test
Google is launching its first experimental orbital data center under Project Suncatcher on October 1. The payload features four Google Tensor Processing Units (TPUs) configured to run in brief 15-minute operational tests in space. Orbital computing represents a novel frontier for AI infrastructure that could harness continuous solar power without terrestrial land or cooling constraints. Successful testing could lay the foundation for scalable, space-based machine learning networks. The mission focuses on testing thermal management, cooling systems, and chip survivability in the harsh space environment. Operating in short 15-minute windows allows engineers to carefully monitor hardware stability while managing power constraints.
## BACKGROUND
Announced as a research moonshot, Project Suncatcher explores using networks of solar-powered satellites equipped with Google's custom AI chips to scale computing in orbit. Google's Tensor Processing Units (TPUs) are proprietary application-specific integrated circuits (ASICs) optimized for accelerating machine learning workloads. Moving AI hardware into space requires overcoming major engineering challenges such as extreme radiation and heat dissipation without air.