Introduction
Google said on October 30 that its Project Suncatcher prototype satellite had generated an AI response using a data center class TPU in orbit. The test ran the Gemma 4 model and returned its output to Earth. Google’s October update
From launch to a working response
The satellite launched on October 1 through Google’s partnership with Planet. Google described the mission as research into whether space could support machine learning infrastructure. The launch announcement
The new result moves that experiment beyond establishing contact with the spacecraft. A prompt sent from the ground produced an answer from the onboard model. The announcement dates that milestone to the week of October 30 without giving the exact day. What Google tested
What it could be used for
The October 1 announcement set out a hardware research goal: collect in-orbit evidence about how TPUs withstand spaceflight, radiation and temperature extremes. Those conditions matter when evaluating whether a future computing system could operate there reliably. The mission’s research purpose
Our take at araNET: this is useful infrastructure research to follow. A successful response answers one engineering question; it does not tell a business whether the system would suit a production workload. Reliability, access and cost would need their own evidence before a hosting decision.
Availability and next steps
Google says further tests will examine thermal stress and radiation exposure, with two interconnected satellites planned for 2027. Neither announcement offers customer access or commercial pricing. There is no announced service here for moving an araNET website or application into orbit. Next tests
