WASHINGTON, Aug. 27 (DC Times Online) — NASA Administrator Jared Isaacman has proposed putting data centers in Earth orbit, where solar panels could supply power without using land, water or electric-grid capacity on the ground.
The idea is still largely a proposal, not a commercial service. No full-scale orbital data center has been shown to be operating, and major questions remain about cost, construction, cooling, communications and space debris.
Isaacman discussed the concept in an interview with Nexstar DC. He said orbital facilities could use the sun’s energy while helping the United States compete in artificial intelligence and space.
“Let’s take our data centers, put them in Earth orbit, take advantage of that free fusion reactor that’s out there for us and use that to ensure we do win the space race for sure, but the AI race, too,” Isaacman said.
Why put a data center in space?
A data center is a facility filled with computer servers. AI systems require large amounts of computing power, and those servers consume electricity and produce heat.
On Earth, data centers can compete with homes, businesses and industry for power. They also need land, water and connections to the electric grid. The facilities can become a source of local opposition when communities are asked to support new transmission lines, power plants or water use.
An orbital facility would operate differently. Solar arrays could generate electricity above the atmosphere, and the facility would not need to occupy land or draw cooling water from a local supply. Isaacman has described that separation from communities and terrestrial infrastructure as a central benefit.
Space also offers continuous exposure to sunlight in some orbits. But solar power in orbit is not automatically simple or inexpensive. The arrays, computing equipment and support systems would all have to be launched, assembled and maintained in space.
What companies are proposing
Google has described Project Suncatcher as a research effort involving its tensor processing unit, or TPU, chips on solar-powered satellites linked by laser communications. Google has said a demonstration mission is planned for 2027.
Starcloud is among the companies working on on-orbit computing. Scientific American reported that the company launched a 60-kilogram satellite carrying an NVIDIA H100 graphics processing unit. That is a small demonstration compared with a commercial data center.
The Aerospace America publication has also reported proposals from SpaceX and other companies for large satellite networks designed to process AI workloads. It said companies in China have announced on-orbit AI plans and that several U.S. companies are seeking regulatory approval for orbital data centers.
The Federal Communications Commission would be involved in licensing relevant satellite systems. The European Space Agency has announced a contract with Edge Aerospace to study orbital data centers.
What makes the idea difficult?
The first problem is transportation. A space-based data center would require many launches, and the launch price would have to fall sharply for the concept to compete with facilities on Earth. Wen, an expert quoted by Aerospace America, said launch costs would need to reach about $200 to $300 per kilogram.
The second problem is heat. Computer chips turn much of the electricity they use into heat. On Earth, data centers can move heat into air or water. In space, there is no surrounding air, so heat must be released through large radiators that emit infrared energy.
The equipment would also need protection from radiation. Laser links could move data between satellites, but those systems would have to provide fast, reliable connections. Repairs would be difficult and could require additional launches or robotic servicing.
The scale could be enormous. An arXiv research paper said producing 5 gigawatts of power in orbit would require solar arrays measuring about 4 by 4 kilometers. The paper also raised concerns about the appearance of large satellite groups in the night sky and the risk that collisions could create debris.
Could it reduce emissions?
Researchers have pointed to solar power and radiative cooling as possible advantages. Benjamin Lee of the University of Pennsylvania said continuous solar energy could help address the challenge of powering computing in a carbon-efficient way.
That does not settle the environmental question. Launches produce emissions, and manufacturing millions of satellites would require substantial materials and energy. A large satellite population could also add to congestion in orbit and increase the chance of collisions.
The arXiv paper said a SpaceX filing with the FCC described plans for as many as 1 million solar-powered orbital data-center satellites and 100 gigawatts of AI computing capacity per year. The underlying filing was not included in the available materials, so those details could not be independently reviewed here.
How soon could it happen?
Small satellites can demonstrate individual computing components, but that is different from operating a large, commercial data center in orbit. Google’s planned demonstration, company projections and proposed regulatory filings point to continued interest, not proof that the technology is ready for broad deployment.
A separate effort by Caltech and Sophia Space involves a patented space-based cooling system that the companies say could support solar-powered data centers, potentially in the early 2030s. That timeline is a target, not a confirmed launch date.
For now, orbital data centers remain a possible answer to a terrestrial infrastructure problem. NASA’s administrator has given the idea a prominent federal voice, but the technology must still overcome the basic economics and engineering of building, powering, cooling and protecting computers in space.
