Google prepares to test AI computing in space with Project Suncatcher

14:58

By: Azzat Manal

Google prepares to test AI computing in space with Project Suncatcher

Google is preparing to take an unusual step in the development of artificial intelligence by testing specialized computing hardware in low Earth orbit. Through its Project Suncatcher initiative, the technology company is exploring whether space could eventually host large-scale infrastructure dedicated to running AI workloads.

The first experimental mission is scheduled to launch aboard SpaceX’s Transporter-18 rideshare mission in partnership with Planet. The test will place Google’s Tensor Processing Units, or TPUs, in orbit to determine how the hardware performs under the demanding physical conditions of space. Google describes the mission as an early step toward understanding whether AI computing can eventually be scaled beyond Earth.

The project is based on the idea of using interconnected satellites equipped with AI processors and powered by solar energy. In low Earth orbit, satellites can receive sunlight for much longer periods than solar installations on the ground, potentially offering a continuous source of energy for future computing systems. Google estimates that orbital solar panels could receive substantially more sunlight than comparable systems on Earth.

Before such a concept could become commercially viable, however, engineers must overcome several technical obstacles. The first mission will examine how TPUs respond to vibration and acceleration during launch, as well as radiation and extreme thermal conditions once in orbit.

Radiation is one of the main concerns for electronic systems operating outside the protection of Earth's atmosphere. Google says its preliminary ground tests have shown that its Trillium TPUs can withstand radiation levels exceeding those expected during a five-year mission. The company plans to use the orbital test to gather real-world data that cannot be fully reproduced in laboratories.

Cooling is another major challenge. On Earth, data centers rely heavily on air and other cooling systems to remove heat from processors. In the vacuum of space, there is no conventional airflow, meaning orbital computing platforms would have to rely on radiators and heat-transfer technologies to prevent AI chips from overheating. Google is testing these systems as part of the Suncatcher programme.

The longer-term vision goes beyond a single experimental satellite. Google is studying the possibility of deploying clusters of satellites carrying multiple TPUs and connecting them through high-speed laser communication systems. The company expects future tests to examine whether satellites can maintain the precise, high-bandwidth links required to distribute AI workloads across an orbital network.

The initiative reflects growing interest in alternative infrastructure as demand for AI computing continues to increase. Training and operating advanced AI models requires substantial processing capacity and electricity, prompting technology companies and researchers to explore new approaches to energy supply, cooling and data-center design.

Google first announced Project Suncatcher in 2025 as a long-term research effort aimed at investigating whether solar-powered satellites could eventually form an interconnected machine-learning infrastructure in space. Its initial research focused on satellite constellation design, communications and the ability of AI processors to withstand radiation.

For now, orbital data centers remain an experimental concept rather than an established alternative to terrestrial facilities. The upcoming test is intended to answer fundamental engineering questions, including whether AI hardware can operate reliably in orbit and whether satellites can eventually be connected efficiently enough to support large computing workloads.

If the technology proves viable, space could eventually become another location for specialized computing infrastructure, complementing rather than immediately replacing data centers on Earth. For Google, Project Suncatcher represents a long-term experiment at the intersection of artificial intelligence, renewable energy and space technology.