Musk’s xAI–SpaceX Merger Revives Long-Held Dream of Data Centers in Space

BY COMFORT OGBONNA

Seventy-five years ago, the notion of harnessing the vast power of space for human industry belonged largely to the realm of science fiction, imagined by writers such as Arthur C. Clarke and Isaac Asimov. This week, Elon Musk’s sweeping merger of his artificial intelligence venture xAI with SpaceX has pushed that once-distant vision closer to reality, reviving serious discussion about moving large-scale computing infrastructure beyond Earth.

For nearly two decades, engineers and technologists at NASA have debated the possibility of relocating energy-intensive computing to space, where solar power is abundant and unconstrained by Earth’s geography. In recent years, the idea has also caught the attention of major technology players, including Alphabet and Jeff Bezos’ Blue Origin. While the underlying physics have long appeared sound, the technical, economic and logistical barriers were widely viewed as overwhelming.

Musk, however, has built a reputation on pursuing ideas many once dismissed as unrealistic and turning them into operational systems. With control of the world’s most active satellite launch fleet through SpaceX, a rapidly growing AI company in xAI, and an appetite for building infrastructure that spans from Earth’s surface into orbit, he may now be positioning himself to attempt what others have only studied. Speaking on Monday, Musk argued that space-based AI infrastructure is ultimately unavoidable if computing is to scale far beyond current limits, pointing to the immense, largely untapped energy of the sun as the decisive factor.

The merger has sharpened investor focus on how Musk could knit together rockets, satellites and AI systems into a single ecosystem capable of supporting data centers in orbit. This renewed attention comes as SpaceX prepares for what could become a landmark initial public offering, with valuations discussed as high as $1.5 trillion, underscoring the scale of ambition behind the strategy.

SpaceX has already taken concrete regulatory steps in this direction. The company has sought approval to launch as many as one million solar-powered satellites designed to function as orbital data centers, a figure that would dwarf any existing satellite constellation. In a filing with U.S. regulators, SpaceX outlined a concept for a solar-powered, optically linked “orbital data-center system,” though it stopped short of detailing how many Starship launches would be required to deploy such a network at operational scale.

Supporters of the idea argue that data centers in space could eventually be cheaper and more efficient than their Earth-bound counterparts. Continuous access to solar energy and the ability to radiate heat directly into space could eliminate some of the most pressing constraints faced by terrestrial facilities. Yet even advocates acknowledge that the concept faces formidable obstacles, including radiation exposure, orbital debris, heat management, communication latency and a cost structure that remains difficult to predict.

Industry analysts caution that commercial viability is likely many years away. The financial modeling of such projects is especially challenging because so many technical variables remain unresolved. Still, even skeptics concede that Musk’s track record makes it difficult to dismiss the concept outright, noting that much of the confidence surrounding the effort rests on his history of executing ambitious, capital-intensive projects.

The physics behind space-based infrastructure are not new. Research into orbital solar power dates back to the Cold War, when U.S. government agencies explored the idea in the 1970s before concluding that launch and materials costs rendered it impractical at the time. What distinguishes Musk’s approach is the degree of vertical integration his companies possess. SpaceX controls launch capabilities, satellite manufacturing and communications links, while Musk’s broader business empire could help generate demand for large volumes of AI computing.

Despite these advantages, significant technical challenges remain. Data-center hardware in orbit would be continuously exposed to cosmic radiation, which can degrade or destroy conventional chips. Historically, space-grade electronics were specially hardened to withstand such conditions, but they lag far behind modern AI processors in performance. Cooling presents another major hurdle. Although space is cold, it is also a near vacuum, meaning heat cannot dissipate through air or liquid as it does on Earth. Instead, powerful processors would need to channel heat into large radiators that emit infrared energy, adding mass, complexity and cost.

SpaceX has indicated that its proposed systems would rely on passive heat dissipation into the vacuum of space and has described mechanisms to rapidly de-orbit satellites that suffer failures. Other technology companies are also experimenting with related ideas. Alphabet, for example, has subjected its AI chips to radiation testing in laboratory conditions to evaluate whether they could survive multi-year missions in orbit as part of a research initiative aimed at networking solar-powered satellites into an orbital AI cloud.

Experts widely agree that even under the most optimistic scenarios, fully functional data centers in space are unlikely to become a reality for at least another decade. Still, the convergence of cheaper launch costs, advances in satellite manufacturing and the explosive growth of AI computing has brought the concept out of the realm of pure science fiction and into serious strategic planning. With the xAI–SpaceX merger, Musk has made clear that he intends to be at the center of that transformation, betting that the future of computing may ultimately lie far above the Earth itself.