Google’s “moonshot” aspirations to expand its AI footprint are taking on a more literal meaning.
CEO Sundar Pichai said in a Fox News interview in December Google will soon begin construction of AI data centers in space. The tech giant announced Project Suncatcher late last year, with the goal of finding more efficient ways to power energy-guzzling centers, in this case with solar power.
“One of our moonshots is to, how do we one day have data centers in space so that we can better harness the energy from the sun that is 100 trillion times more energy than what we produce on all of Earth today?” Pichai said.
Google will take its first steps in constructing extraterrestrial data centers in early 2027 in partnership with satellite imagery firm Planet, launching two pilot satellites to test the hardware in Earth’s orbit. According to Pichai, space-based data centers will be the new standard in the near future.
“But there’s no doubt to me that a decade or so away we’ll be viewing it as a more normal way to build data centers,” he said.
The data center space race
Google isn’t the only company looking to the skies for an answer to improving data center efficiency. Earlier this year, SpaceX sought permission to launch as many as 1 million satellites into Earth’s orbit, part of a bigger goal of launching a solar-powered satellite network to “accommodate the explosive growth of data demands driven by AI,” according to a filing with the Federal Communications Commission.
In December 2025, Y Combinator and Nvidia-backed startup Starcloud sent its first AI-equipped satellite to space. CEO and cofounder Philip Johnston predicted extraterrestrial data centers will produce 10 times lower carbon emissions than their earthbound counterparts, even taking into account the emissions from launch.
While the cost of satellites used to test AI hardware in space has decreased drastically, putting extraterrestrial data center development within reach, the cost of building these solar-powered centers is still an unknown, particularly as earthbound data centers are expected to require more than $5 trillion in capital expenditures by 2030, according to an April 2025 McKinsey report.
Google, which catapulted itself back into the AI front-runner conversation with the recent release of Gemini 3, is one of several major hyperscalers pouring money into data centers to expand its computing capabilities. Alphabet, Google’s parent company, said in February it would spend $175 billion to $185 billion this year in capital expenditures, primarily to build out AI infrastructure.
Data center moonshots come back to Earth
All the while, speculation of an AI bubble threatens to create an oversupply of data centers, which could render the data center space race a dangerous over-investment. Moreover, with the technology quickly developing, there’s a risk data centers under construction now could have out-of-date equipment by the time they are completed.
Hyperscalers, including Alphabet, are taking an even greater risk by financing their AI buildouts with debt. In 2025, Alphabet, Amazon, Oracle, Meta and Microsoft issued $121 billion in new debt through bonds. That’s compared to $40 billion in new debt in 2020.
“The stakes are high,” the McKinsey report said. “Overinvesting in data center infrastructure risks stranding assets, while underinvesting means falling behind.”
Harnessing solar energy to power data centers has become increasingly appealing amid growing concerns about the sustainability of expanding AI compute, which requires an exorbitant amount of power. A December 2024 U.S. Department of Energy report on domestic data center usage found data center load has tripled in the past 10 years and may double or triple again by 2028. These data centers consumed more than 4% of the country’s electricity in 2023, and are predicted to consume up to 12% of U.S. electricity by 2028, according to the report.
Google alone has more than doubled its electricity consumption on data center use in the past five years, using 30.8 million megawatt-hours of electricity last year compared to 14.4 million in 2020, when it began specifically tracking data center energy consumption, according to its latest sustainability report released in June 2025.
Google has worked to reduce the energy needed to power its growing data centers, reporting it reduced its data center energy emissions by 12% in 2024, despite an increasing footprint. However, concerns about the feasability and timeline of extraterrestrial data center expansion remain.
Amazon Web Services CEO Matt Garman poured cold water on data centers in space at a tech conference in San Francisco in February: “I don’t know if you’ve seen a rack of servers lately: They’re heavy. And last I checked, humanity has yet to build a permanent structure in space. So … maybe.”
Others have warned about future sustainability concerns of an AI buildout expanding beyond Earth, indicating the AI space race don’t happen for decades.
“There is still much we don’t know about the environmental impact of AI, but some of the data we do have is concerning,” Golestan Radwan, United Nations Environment Programme chief digital officer, said in a 2024 statement following the program’s note warning of the environmental impact of AI infrastructure expansion. “We need to make sure the net effect of AI on the planet is positive before we deploy the technology at scale.”
A version of this story was published on Fortune.com on Dec. 1, 2025.
More on extraterrestrial data centers:
- Google’s plan to put data centers in the sky faces thousands of (little) problems: space junk
- Elon Musk is pushing to build data centers in space. But they won’t solve AI’s power problems anytime soon
- AI is running out of power. Space won’t be an escape hatch for decades
