Project Suncatcher tests Google chips in orbit: Space data centers between abundant solar power and radiation challenges

Google is preparing to launch an experimental satellite next week to test its Tensor Processing Units in low Earth orbit, marking the first real-world test of AI chips outside the atmosphere. The exploratory step is part of a research project the company calls “Project Suncatcher,” with the experiment riding SpaceX’s Transporter-18 mission, in technical collaboration with satellite manufacturer and operator Planet, to verify how well AI-specific hardware can survive and operate reliably in the harsh space environment.

The experiment subjects the TPUs to a complex set of physical challenges, from the jolting forces and violent vibrations of rocket launch to the sharp temperature swings and direct exposure to space radiation. Radiation in low-Earth orbit poses a clear threat to fine-grained electronics by inducing random errors in data and computations, which has led the company to conduct prior ground tests, acknowledging that laboratory simulations cannot replace direct measurement in the actual orbital environment.Testing cooling in vacuum is one of the experiment’s most intricate aspects, as reliance on airflow used in conventional ground-based data centers is impossible in space.

The core motivation behind the project goes beyond merely testing silicon durability; Google is exploring the possibility of building and operating a large-scale computing infrastructure in space in the future, based on a superior energy equation offered by certain orbits that receive near-continuous sunlight. The company estimates that solar panels in low-Earth orbit can generate power equal to eight times what the same area can produce on Earth’s surface, theoretically opening the way to run fully integrated constellations of satellites that handle massive machine-learning workloads. Google plans to continue the research path by launching two additional satellites in 2027 to test high-bandwidth laser communication links between satellites, stressing that the current mission is intended to gather data and identify potential failure points rather than to inaugurate an operational data-center ready for service.

This research trajectory redirects attention to the real obstacle facing global AI expansion: energy scarcity and cooling costs.For the Arab region, and especially Gulf markets that are pouring large investments into massive data-center and AI complexes, this shift signals an early strategic cue. While ground facilities in desert environments face operational pressures on power grids and cooling resources, orbital-computing research shows that the pursuit of uninterrupted solar energy could eventually intersect with the advanced space sector in Saudi Arabia and the United Arab Emirates. The trend also compels infrastructure and network engineers in the region to monitor the evolution of radiation-hardened hardware standards and laser-link protocols, preparing for shifts that may no longer confine high-performance computing centers within traditional geographic borders.