Google has demonstrated an artificial-intelligence workload running on computing hardware in orbit, according to a report published on October 31 by The Information Machine. The test involved a satellite carrying a Google Trillium Tensor Processing Unit and a Gemma model: a prompt sent from Earth triggered an inference run in orbit, and the system returned a response.
The experiment is an early technical demonstration, not evidence that orbital data centers are ready to replace terrestrial infrastructure. Space-based computing faces difficult engineering constraints, including launch costs, radiation exposure, thermal management, communications capacity, maintenance and the limited lifetime of satellites.
Google has also described plans to explore optical links between satellites, which could help test how orbital systems communicate and share workloads. If the technology develops, potential applications could include processing some data closer to where it is collected, although the economics and practical use cases remain uncertain.
The broader significance is that technology companies are exploring new ways to expand computing capacity as AI workloads grow. For now, the demonstration should be viewed as research progress rather than a near-term commercial service.
Source: The Information Machine, October 31, 2026 — https://www.theinfomachine.fyi/daily/2026-10-31
