The FCC filing for a million satellite AI data center network isn't a tech story. It's a governance stress test for an orbital commons that was never designed for this scale.
On May 29, 2026, SpaceX filed with the Federal Communications Commission to answer questions about its own proposal to deploy up to one million satellites that would, in the company's framing, host artificial intelligence data centers in low-Earth orbit. The CNET report on the filing treats the plan as closer to reality than to science fiction. The more important story is the institution being asked to evaluate it. The FCC was built to license spectrum and assign orbital slots, not to govern a roughly 67-fold scale-up of the orbital commons by a single private actor.
The numbers make the scale legible. About 15,000 active satellites currently orbit the Earth, a population dominated by SpaceX's own Starlink constellation. One million would be a 67x increase on what is up there now (a reporter calculation derived from the 15,000 active-satellite figure), and a step change in collision-avoidance maneuvers and end-of-life reentries. The same CNET reporting puts the operational burden squarely on the table: more satellites means more conjunctions to track, more propellant burned to dodge them, and more mass reentering the atmosphere each year.
Why push compute off-planet in the first place? The CNET piece points to ground opposition to AI data centers and the electricity and water load of terrestrial AI buildouts. The politics of AI siting are no longer hypothetical. The CNET reporting cites a poll finding that seven in ten Americans do not want data centers built near where they live, and documents community-level concern — including testimony from organizers near SpaceX's own launch footprint — about the localized effects of AI infrastructure buildout. That ground-level resistance is part of what is pushing the economics of orbital compute.
The scientific concerns are real, and they are not all the same. Orbital debris and Kessler-syndrome cascades are the most cited, and the CNET report gathers on-record expert warnings about a graveyard scenario if the math goes wrong. Atmospheric and ozone effects from a sustained, large increase in reentry ablation are a separate research question without a settled answer at this scale. Astrophysicist Jonathan McDowell, as quoted in the CNET piece, has cautioned against betting against SpaceX's execution while flagging the same sustainability constraints. The right read is to take the operational risk seriously without treating it as a foregone catastrophe.
The harder question is governance. The FCC licenses spectrum. The Federal Aviation Administration licenses reentry. The Commerce Department's Office of Space Commerce tracks debris. The International Telecommunication Union coordinates orbital slots across borders, with its own rules and a long queue. None of these bodies was built to evaluate, condition, or cap a single proposal that would multiply the active satellite population by nearly 70x. The CNET report describes the agency as being asked to evaluate the plan without a clear methodology for saying no. That is the structural gap the filing exposes.
Two things are worth watching. First, what the FCC actually does with the filing: a green light, a conditional grant, a request for more data, or a referral upstream. Each is a signal about how the U.S. intends to govern low-Earth orbit at scale. Second, what the international coordination regime looks like by the time the first tranche of these satellites is ready to launch. ITU priority, debris-mitigation standards, and deorbit-cadence rules will shape whether the proposal stays on paper or becomes a lived crisis.
The mitigations are not theoretical. The FCC mandated in 2022 that satellites be de-orbited within five years — a reduction from the prior 25-year rule — and SpaceX's filing proposes Earth disposal orbits and heliocentric orbits as alternatives to atmospheric reentry for satellites above 600 kilometers, where natural drag is too low to pull them back. Active debris removal, in-orbit servicing, more rigorous deorbit standards, and on-orbit collision-avoidance automation are in development across the industry. The question is not whether technology can keep up. It is whether the rules that govern the commons can keep up, and whether the public gets a seat at the table before the orbital neighborhood is built out around them.