On Friday, SpaceX submitted a request to the Federal Communications Commission seeking approval to launch a constellation of one million data center satellites into orbit. While the FCC is unlikely to authorize such an extensive network, SpaceX's strategy is to use the approval request for an impractical number of satellites as a starting point for negotiations.
The application proposes establishing a network of solar-powered data centers in low Earth orbit, which would communicate with each other via lasers. The filing ambitiously frames the constellation within a science-fiction context, describing it as "the first step toward a Kardashev Type II civilization—one capable of harnessing the entire power output of its star."
Even if only a fraction of these one million satellites were to reach orbit, it would significantly increase the number of human-made objects in space. The European Space Agency currently estimates there are roughly 15,000 satellites orbiting Earth, the majority of which are Starlink satellites. (According to Jonathan's Space Report, the number exceeds 9,600.)
At a time when experts are already concerned about the proliferation of space debris and the risk of orbital collisions, a dramatic increase in orbiting objects may seem unwise. However, SpaceX argues that orbital data centers would be a more economical and environmentally friendly alternative to their terrestrial counterparts, which form the backbone of the rapidly expanding artificial intelligence industry. Instead of drawing water from local communities, polluting groundwater, or driving up electricity costs, orbital data centers could radiate excess heat into the vacuum of space, relying almost entirely on real-time solar power with limited battery backup.
Backlash against data centers is intensifying, with communities increasingly winning battles to block their construction. Consequently, it is unsurprising that major players in the AI sector are turning their attention to the few locations where there are no communities to disturb.
