NASA administrator suggests putting data centers in space

The suggestion from NASA Administrator Jared Isaacman to relocate data centers to Earth orbit is a bold and imaginative response to a very real crisis. As artificial intelligence expands, its energy appetite has become a major concern for utilities, policymakers, and environmentalists alike. Ground-based data centers consume massive amounts of electricity, often drawing from fossil fuels, and their cooling needs strain water resources. Isaacman’s proposal to tap into the sun’s unfiltered energy in space is technically sound—solar panels in orbit receive more constant and intense sunlight than any terrestrial site. It also sidesteps land-use conflicts and could theoretically offer a cleaner path for AI’s growth. However, the enthusiasm must be tempered with realism. Launching and maintaining orbital data centers would require a staggering upfront investment in rocketry, robotics, and thermal management. The cost per kilogram to orbit remains high, though reusable rockets like SpaceX’s Starship are lowering it. Yet even with cheaper launch, the logistics of repairing, upgrading, and networking space-based servers are daunting. Latency—the delay in data transmission between orbit and Earth—would be unacceptable for many real-time applications, such as autonomous vehicles or interactive AI assistants. For certain workloads, like batch processing or training models, latency is less critical, but the overall utility is limited. There is also a geopolitical dimension. Isaacman framed the idea as a way for the U.S. to “win” the AI race, but space is increasingly contested. Orbital data centers would become high-value targets for adversaries, raising security concerns. Who controls the infrastructure? How do we ensure equitable access? These are questions that extend beyond engineering into international law and ethics. The proposal is visionary, but it risks becoming a distraction from more immediate, pragmatic solutions like improving energy efficiency, advancing nuclear power, or building smarter grid systems. Still, we should not dismiss out-of-the-box thinking. History shows that transformative technologies often emerge from seemingly impractical ideas. The challenge is to balance ambition with feasibility. Isaacman’s role gives this idea legitimacy, and it may spur investment in orbital manufacturing and solar power infrastructure that could have broader benefits. As with any major policy proposal, we must weigh the potential rewards against the risks and costs. The debate itself is valuable—it forces us to confront the true cost of our digital future and consider all possible paths forward.