Google is preparing to conduct the first in-orbit test of Project Suncatcher, a long-term research initiative exploring whether space could eventually support scalable machine-learning infrastructure. The company plans to launch a prototype satellite carrying its Tensor Processing Units (TPUs) to evaluate how the AI chips perform in the conditions of space.
The prototype is scheduled to fly aboard SpaceX’s Transporter-18 rideshare mission, with the satellite being developed in partnership with Planet. Google CEO Sundar Pichai announced the upcoming test on September 24, asking whether the company’s TPUs can survive and operate in space. The launch is expected on October 1, 2026.
Project Suncatcher is focused on examining the possibility of building AI computing infrastructure in low Earth orbit. According to Google, satellites operating in low Earth orbit can access near-constant sunlight and potentially generate up to eight times more solar power than on Earth. The company is exploring whether interconnected satellite constellations could eventually handle larger AI workloads while in orbit.
The upcoming mission is intended to collect real-world data on how Google’s AI hardware performs under the physical conditions of space. Google has already conducted ground-based testing, including radiation testing of its Trillium TPUs at the University of California, Davis’ Crocker Nuclear Laboratory. The company said initial results showed that the TPUs could withstand a total ionising radiation dose greater than what they would encounter during a five-year space mission.
Cooling the chips is another technical challenge being examined. TPUs generate significant heat, while conventional airflow-based cooling cannot be used in the vacuum of space. Google is therefore testing a cooling approach that combines heat pipes and radiators. The company has already evaluated the system inside a thermal vacuum chamber designed to reproduce the vacuum and temperature conditions of space.
Google is also studying how multiple satellites could eventually work together. Future designs could involve satellites carrying dozens of TPU chips and operating in clusters. These satellites would need to maintain precise relative positioning and exchange data through high-bandwidth laser links. Google plans to test this inter-satellite connectivity with two satellites in 2027.
The first Suncatcher mission, however, is an early research exercise rather than an operational space-based data centre. Google said the initial launch is intended to determine what works, identify potential failure points and gather information that can guide subsequent missions.
The concept of moving AI computing into space is also attracting interest elsewhere in the technology sector. SpaceX and other companies are exploring orbital computing concepts, while Google’s longer-term Suncatcher research focuses on interconnected, solar-powered satellites equipped with its own TPU accelerators.
For Google, Project Suncatcher represents an early step in testing whether AI computing can operate reliably beyond Earth. The results from the initial satellite mission are expected to provide data for the company’s planned 2027 tests and future research into scalable machine-learning infrastructure in space.
