SpaceX Wants a Million Satellites for AI: Is This the Future of Computing or a Space Disaster Waiting to Happen?

SpaceX has submitted a daring proposal to the US Federal Communications Commission (FCC) to launch a constellation of up to one million satellites into Earth orbit, not for Internet coverage like Starlink but as orbital data centers designed to power artificial intelligence applications on a global scale. The plan, if approved, could reshape how humanity processes data, runs AI models and thinks about computing infrastructure and it is already stirring excitement and controversy across tech and space communities.

SpaceX’s solar-powered swarm for AI brainpower

Unlike traditional satellites that relay signals, the proposed spacecraft would be solar-powered computing hubs, orbiting between roughly 500 km and 2,000 km above Earth in sun-synchronous and equatorial shells. Each craft is envisioned to act as a node in a massive orbital data center, capturing near-constant solar energy to run AI workloads more efficiently than on land-based facilities.SpaceX argues this is the “most efficient way” to supply the exploding demand for AI compute, a demand driven by ever-larger language models, autonomous systems and data analytics that now strain physical infrastructure on Earth. The company contends that solar-powered satellites could drastically reduce energy and cooling costs relative to traditional data centers, which consume huge amounts of electricity and water.

Why one million satellites? What’s SpaceX’s vision?

The sheer scale, one million, is what makes this plan so unusual and talked about. In a filing, SpaceX framed the initiative not just as a tech project but as a step toward a long-term vision of humanity harnessing the Sun’s power more directly, a concept sometimes linked to the theoretical Kardashev scale of civilisations.