Google to put datacentres in space to meet AI demand
Google plans to put AI datacentres in space, with its first prototype satellites set to launch in early 2027 to meet rising demand for artificial intelligence.
The US technology company’s engineers believe around 80 solar-powered satellites could orbit about 400 miles above Earth, equipped with processors to handle AI workloads.
Research released on Tuesday suggests that by the mid-2030s, operating costs for space-based datacentres could match those on Earth, driven by falling rocket launch prices.
Google engineers say the orbiting facilities would use solar panels up to eight times more productive than Earth-based ones. The satellites would transmit data back through optical links using light or laser beams.
The company said: “In the future, space may be the best place to scale AI computers.
“Working backward from there, our new research moonshot, Project Suncatcher, envisions compact constellations of solar-powered satellites, carrying Google TPUs and connected by free-space optical links.
“This approach would have tremendous potential for scale, and also minimises impact on terrestrial resources.”
TPUs are processors optimised for training and the day-to-day operation of AI models.
The orbiting datacentres are part of Project Suncatcher, which aims to reduce pressure on land and water resources needed to cool Earth-based facilities.
By the mid-2030s, the company’s research suggests the cost of running such space-based operations could rival those on the ground.
However, launching a single rocket produces hundreds of tonnes of CO2.
Astronomers have also raised concerns about growing satellite numbers in low orbit, describing them as “like bugs on a windshield” when observing the universe.
Major technology companies are projected to spend US$3tn (£2.3tn) on Earth-based datacentres from India to Texas and Lincolnshire to Brazil, raising concerns about carbon emissions if cleaner energy is not found to power the sites.
Google plans to launch two prototype satellites by early 2027 and said current research results mark a “first milestone towards a scalable space-based AI.”
But the company added a note of caution: “Significant engineering challenges remain, such as thermal management, high-bandwidth ground communications and on-orbit system reliability.”
Elon Musk, who runs the Starlink satellite internet provider and SpaceX rocket programme, last week said his companies would begin scaling up to create datacentres in space.
Nvidia AI chips will also be launched later this month in partnership with the startup Starcloud.
“In space, you get almost unlimited, low-cost renewable energy,” said Philip Johnston, co-founder of the startup.
“The only cost on the environment will be on the launch, then there will be 10 times carbon dioxide savings over the life of the datacentre compared with powering the datacentre terrestrially.”
The concept could reduce reliance on land and water resources currently needed to cool conventional facilities, though the environmental cost of increased rocket launches remains an issue.
