Google is preparing to launch a prototype satellite next week as part of Project Suncatcher, its research effort to determine whether large-scale AI computing can eventually operate in space. The mission will send Google’s Tensor Processing Units (TPUs) into low Earth orbit for the first time.
The satellite will fly aboard SpaceX’s Transporter-18 rideshare mission, in partnership with satellite company Planet. Google will use the mission to collect real-world data on how its AI hardware performs under the harsh conditions of space.
Project Suncatcher is Google’s long-term research programme focused on building scalable machine-learning infrastructure in orbit. The company believes satellites could eventually use near-continuous sunlight to generate power for energy-intensive AI workloads.
Google says, “ Suitable low Earth orbits could provide significantly more solar energy than comparable systems on Earth. The broader concept involves multiple satellites working together as an orbital computing network.”
The first mission will examine whether Google’s AI chips can withstand launch vibrations, acceleration, radiation and extreme temperature changes in low Earth orbit. Radiation presents a particular challenge because it can interfere with electronic components and cause data errors known as bit flips.
The company will also test how to cool powerful AI chips in space. Conventional air-based cooling does not work in a vacuum, so the satellite uses heat pipes and radiators to move heat away from the processors.
AI data centres require substantial electricity and cooling capacity. Google is exploring whether solar-powered satellites could provide an alternative infrastructure model while reducing some of the constraints faced by terrestrial data centres.
The longer-term Suncatcher concept would connect satellites using high-bandwidth laser communications, allowing multiple spacecraft to operate as a distributed computing system. Google plans to test such satellite-to-satellite links with two spacecraft in 2027.
Also Read: Google Launches Gemini 3.8 Live Avatar for Real-Time AI Conversations