Skip to content

The next great space race: Building data centers in orbit

Northeastern experts say that while early work on AI data centers in space may be ongoing, don’t expect them soon.

SpaceX’s mega rocket Starship makes a test flight from Starbase, Texas.
SpaceX’s mega rocket Starship makes a test flight from Starbase, Texas, Monday, Oct. 13, 2025. AP Photo/Eric Gay

Google, SpaceX and Blue Origin are reportedly racing to develop technology for AI data centers in space, but it will likely be years before we see them rocketing into the sky, experts say.

“I don’t think we’ll have an operative data center in space in the next couple of years, but we’ll start seeing some of the building blocks tested in the next couple of years,” said Josep Jornet, a computer and electrical engineering professor at Northeastern University and satellite researcher. 

The thinking goes that space isn’t bound by the same constraints as Earth, where energy demands and environmental impacts can be extraordinally  high. 

According to the International Energy Agency, or IEA, a typical AI data center on Earth consumes annually as much electricity as 100,000 households, and energy impacts are only expected to rise to accommodate the development of larger and more complex AI models.

There is concern that there isn’t enough energy to go around. In 2024, the IEA reported that energy demands from data centers and AI will likely more than double from 460 terawatt-hour (TWh) of electricity in 2022 to more than 1,000 TWh by 2026, which it said is “roughly equivalent to the energy consumption of Japan.”  

Space data centers, by contrast, would instead be equipped with solar-panel-equipped satellites designed to harness power directly from the sun, offset heat into space, and not be limited by available land. 

Fewer data centers on the ground would also potentially help lighten the load on communities like Loudoun County in Northern Virginia, which has been given the nickname “Data Center Alley” for being home to the largest data center hub in the country with over 250 operational facilities.    

“Think of space as the next frontier to conquer,” added Jornet. “There was a gold rush in the West. Now there is the space rush, and everyone wants to put their technology in space.” 

SpaceX’s and Blue Origin’s space data center plans were reported in December by The Wall Street Journal.

Blue Origin has a dedicated team working on creating the technology to support AI data centers in space, the Journal reported. While SpaceX is working on modifying and upgrading its rockets to be better to “host AI computing payloads.” 

SpaceX and Blue Origin have not publicly commented on the Wall Street Journal’s reporting and did not respond to Northeastern Global News’ request for comment. 

But both Elon Musk and Jeff Bezos have in the past talked about building AI data centers in space. 

They are not the only ones.

In November, Google, through its moonshot Project Suncatcher, announced its plan to launch two test satellites with its AI processing chips in 2027 in partnership with Planet Labs, an American satellite company. 

Smaller companies are also working in the space, including StarCloud, an Nvidia-backed startup that last month launched a satellite carrying an Nvidia H100 graphics processing unit, or GPU, which is now being used to run a version of Google’s open large language model, Gemma. 

However, experts say that while early work may be ongoing, there are a whole class of different considerations that should be taken into account as developments progress.    

One of the biggest hurdles will be securing enough power for these orbital data centers to even run in the first place, said Jornet.  

The sun may be a great source for power, but to adequately harness that power orbital data centers would either need massive solar panels miles in length or a constellation of smaller ones ranging in the tens of thousands, he said. 

Companies also need to contend with how their AI chips will hold up under space’s harmful radiation, as well as heat dissipation and cooling issues, he said. 

AI data centers on Earth require thousands of gallons of water to offset the massive amount of energy emitted by their GPUs. While temperatures in space might be low, there is no air in space for chips to cool off naturally and without the use of water. 

“Of course, there are people who say, ‘No, it’s going to be very cool, so we’re just going to radiate the heat,” he said. “But there are discussions from people who have been operating in space that it’s going to be very difficult to cool the amount of heat that even a single GPU generates.” 

Solving these problems will be an iterative process, he said. 

“When these companies say they want to launch things, it doesn’t mean they want to launch a data center right now,” he said. “What they want to do is maybe get a collection of high performance processors and not to use them for computing but to measure their performance, heat cycles, and energy consumption throughout the duration of whatever is the mission length. They are testing one component at a time.”  

But he has no doubt that companies have the incentive to act quickly and spend big to solve these particular challenges as they race for AI dominance. 

“Everyone wants to claim to have the first platform to reach this milestone,” he said. “So companies are throwing cash like there’s no tomorrow.” 

While AI data centers in space ideally should reduce environmental harm, they may very well do the opposite at least in the short term, as these projects will call for more rocket launches, explained Anncy Thresher, a professor of public policy and philosophy at Northeastern University who specializes in space policy. 

Rocket launches have been shown to endanger animals and hurt ecosystems and contribute to the Earth’s use of fossil fuels, she said. 

The Cape Canaveral Space Force Center, for example, which is the most biodiverse federal facility in the continental United States, launches more than 80 rockets each year, which have been found to kill fish in shallow waters nearby, and deposit harmful debris, she said. 

“I think that AI data centers have a lot of promise for solving a lot of the environmental issues here on Earth associated with data centers like those in Virginia, but in order for them to solve these problems, we’re going to have to do a lot more work on the environmental impact of rockets getting into space,” said Thresher.