What if some of the computing power needed for future AI systems is not inside a data centre on Earth but travels around the planet in orbit? That is the question behind Google Project Suncatcher, a long-term research project exploring whether space could eventually support large-scale machine learning infrastructure.
Google is now preparing to put the idea through an early real-world test. The company plans to launch a prototype satellite carrying its Tensor Processing Units (TPUs) to study how AI hardware performs in the harsh conditions of space. The mission is not about building a full-scale orbital data centre yet. Instead, it is designed to answer a much more basic question: Can AI hardware survive and operate reliably in space?
Why does Google want to put AI in space?
The idea is linked to one of the biggest challenges surrounding the growth of AI: computing infrastructure needs enormous amounts of power. Satellites in low Earth orbit have access to sunlight for much of their time and can potentially generate significantly more solar power than comparable solar installations on Earth. Google says satellites in orbit could access up to eight times more solar power than on Earth.
Read More | Google AI For The Planet Accelerator Picks Four Indian Startups For Climate AI Push
Project Suncatcher explores whether that advantage could eventually be combined with AI accelerators to create a network of solar-powered computing satellites. The longer-term concept involves multiple satellites working together in orbit. Each spacecraft could carry large numbers of AI chips, while clusters of satellites could be connected to handle increasingly demanding machine-learning workloads.
That vision, however, depends on solving several difficult engineering problems first.
Google’s first test: Can TPUs survive space?
The first challenge is simply getting the hardware into orbit in working condition.
Read More | Accenture, Google Cloud launch Gemini Enterprise group with 1,000 FDE workforce
A rocket journey to low Earth orbit lasts around 10 minutes, but that short trip exposes spacecraft to intense vibration and acceleration. Google says the spacecraft can experience sustained loads of up to 10 times the force of gravity, while individual components such as TPU chips can experience forces of around 50 to 100 g.
To prepare for this, Google’s engineers subjected the satellite to vibration testing across all three axes. The tests were designed to reproduce the frequencies and forces associated with a rocket launch.
The hardware successfully withstood the testing. Radiation is another major concern. Outside Earth’s atmosphere, electronics face exposure to solar events and cosmic rays that can cause errors in computing systems.
Google tested its TPUs at the UC Davis Crocker Nuclear Laboratory using a proton beam while running AI workloads. Engineers monitored how radiation-induced errors, including bit flips, affected the workloads.
The initial results were encouraging, according to Google. Its Trillium TPUs were able to withstand a total ionising radiation dose greater than the level they would be expected to encounter during a five-year space mission.
But laboratory testing cannot reproduce everything that happens in orbit. That is why Google is taking the next step by putting the hardware into space.
Cooling AI chips in a vacuum
Surviving launch and radiation is only part of the problem. AI chips generate significant amounts of heat, particularly when large computing workloads are concentrated in a small physical area. On Earth, cooling systems can rely on airflow and other conventional methods to move heat away from electronics.
Space is different. There is no air in a vacuum, which means heat cannot be removed through conventional airflow. Instead, spacecraft have to rely on radiators to release heat. For Project Suncatcher, Google is experimenting with a combination of heat pipes and radiators to manage the heat generated by the TPUs.
The company has already tested its cooling technology inside a thermal vacuum chamber designed to reproduce the thermal and vacuum conditions of space. The upcoming orbital test will provide another opportunity to see whether the system performs as expected outside the laboratory.
How will satellites communicate?
A large-scale orbital AI system would need more than powerful chips and solar panels. Its satellites would also need to communicate with one another at very high speeds. Google’s future concept involves clusters of satellites, with each spacecraft potentially carrying dozens of TPU chips.
For these satellites to work together, each one needs to know where it is and where its neighbouring satellites are. Google plans to use laser-based communication for this purpose. Laser communication is already used in space. The challenge for Project Suncatcher is different: the system needs very high bandwidth over relatively short distances.
Read More | Google Launches Ask Google Pay: Gemini AI Assistant to Help Users Track Spending and Save Smarter
Keeping two rapidly moving satellites connected also requires extremely precise pointing. Google compares the challenge to hitting a coin-sized target from miles away while both the target and the person aiming at it are moving. Google plans to test this part of the technology in 2027, when it expects to put two satellites into orbit.
When will Google’s space AI become a reality?
Project Suncatcher is still at an early research stage. The first mission is being developed with Planet and is scheduled to fly on SpaceX’s upcoming Transporter-18 rideshare mission. Its purpose is to collect real-world data on how Google’s TPUs respond to launch conditions, radiation and the thermal environment of space.
Google’s broader plan is expected to develop in stages rather than through one large deployment. The company first needs to establish that its AI hardware can survive in orbit. It then needs to solve cooling, communication and other engineering challenges before a large network of computing satellites could become practical.
The 2027 milestone will provide another important test, particularly for high-bandwidth laser communication between satellites. For now, Project Suncatcher is less about replacing terrestrial data centres and more about investigating whether space could become another location for AI computing infrastructure.
Indrani Priyadarshini is a journalist and editorial professional specialising in technology, artificial intelligence, smart cities, green energy, and digital transformation. With over four years of experience in tech journalism and digital media, she is known for turning complex industry developments into clear, engaging, and insightful stories. Her expertise spans reporting, editorial strategy, digital publishing workflows, and in-depth coverage of emerging technologies shaping the future.
