(Credit: Planet Labs)
Will the next space race feature tech giants rushing to deploy data centers in the sky? Google today announced an effort to launch satellites equipped with the company’s AI chips—something SpaceX CEO Elon Musk and Amazon founder Jeff Bezos are also pursuing.
Google describes Project Suncatcher as a "research moonshot to one day scale machine learning in space." Like others, Google sees potential in harnessing the Sun to power a new class of data centers that orbit the Earth rather than consume energy here on Earth.
Google will initially launch two prototype satellites in early 2027. Each one will be outfitted with the company’s custom AI chips, called the Tensor Processing Unit, which are available through Google’s ground-based data centers.
"Our TPUs are headed to space!" CEO Sundar Pichai tweeted.
Google is enlisting satellite company and imaging provider Planet Labs to help build the Suncatcher hardware. In the meantime, Pichai noted that his company has already simulated how the TPUs might perform in space.
"Early research shows our Trillium-generation TPUs (our Tensor processing units, purpose-built for AI) survived without damage when tested in a particle accelerator to simulate low-Earth orbit levels of radiation," he wrote.
The company published more details in a preprint paper and research post, which reveal that Google envisions potentially using “fleets of satellites equipped with solar arrays” to tap the Sun’s energy in “near-constant sunlight.” To rival the computing capabilities of Earth-based data centers, the satellites could transmit data to each other with the help of space-based lasers, which SpaceX’s Starlink already supports.
Google will need to overcome some major challenges, though, including “thermal management, high-bandwidth ground communications, and on-orbit system reliability,” along with the costs. The paper largely sidesteps a major concern about orbiting data centers: The vacuum of space complicates cooling due to the absence of air to carry away heat. Google offers only a brief statement: “Cooling would be achieved through a thermal system of heat pipes and radiators while operating at nominal temperatures.”
Even so, the company’s research paper concludes “in the long run [space-based data centers] may be the most scalable solution, with the additional benefit of minimizing the impact on terrestrial resources such as land and water.”
Google won’t be alone in developing the technology. Musk tweeted last week about pursuing the same ambition, saying his company already has a foundation in Starlink, the satellite internet constellation. In addition, a startup called Starcloud successfully launched its first test satellite, outfitted with an Nvidia AI GPU, this past weekend, with the goal of developing its own network orbiting data centers.
Meanwhile, Google’s research paper notes it would need to rely on SpaceX, its future reusable rockets, and the lower launch costs to deploy its own orbiting data centers.
"Our analysis of historical and projected launch pricing data suggests that with a sustained learning rate, prices may fall to less than $200/kg by the mid-2030s,” the research post added. “At that price point, the cost of launching and operating a space-based data center could become roughly comparable to the reported energy costs of an equivalent terrestrial data center on a per-kilowatt/year basis.”