SpaceX's May 29 filing asks the FCC to greenlight a constellation 67 times the size of everything in orbit today. The numbers that will decide it are not the ones in the headline.
SpaceX told the Federal Communications Commission on May 29 that it wants to put roughly one million AI and data-center satellites in low Earth orbit, a constellation that would be about 67 times larger than the roughly 15,000 operational satellites circling the planet today, according to a CNET report on the filing. The filing is the news. What it actually asks for is the story.
Ten days later, on June 8, Elon Musk sat for a video interview on X and addressed the inevitable question with a one-liner: "space is really big." It is a comforting sentence and a wrong unit of measure. The binding constraint in low Earth orbit is not volume. It is the narrow bands of altitude where satellites can stay aloft without constant propulsion, and the collision-cascade dynamics that turn a crowded shell into a self-perpetuating debris field.
This is the mechanism that orbital-debris researchers have been warning about since Donald Kessler described it in 1978. Above a certain density of objects in a given shell, each collision produces fragments that trigger more collisions, a chain reaction that can render an orbital regime unusable for generations. The current ~15,000-satellite population is already large enough that SpaceX's Starlink satellites perform thousands of automated collision-avoidance maneuvers per year. A constellation of one million, even spread across multiple proposed shells, pushes that calculus into territory the field has not modeled at scale.
An astrophysicist who tracks the orbital catalog told CNET he finds the plan very hard to see. "Historically, betting against SpaceX has not gone so well," he added. That is a real position: a domain expert saying the company's track record makes him reluctant to call the technical bet impossible, while still flagging that the plan as described strains the physics he works with. Both halves of that quote matter. The honest read is that the FCC is being asked to approve something the orbital-debris community has not yet been shown a model for.
What is actually in the filing, and what isn't, is what the next 90 days are about. The FCC will scrutinize SpaceX's proposed shell altitudes, the cadence at which satellites are deployed and deorbited, and most pointedly the post-mission disposal lifetime: how long after a satellite stops working the operator has to bring it back down. Today's FCC norm is a 5-year post-mission disposal deadline. Anything longer means more dead satellites accumulating in shell. The filing's specific commitments on that number, and on collision-avoidance coordination with other operators, are the figures worth tracking in the docket.
The other thread, the one that explains why a company is pitching orbital compute in the first place, is terrestrial. Roughly 7 in 10 Americans oppose new data centers in their communities, and the electricity and water footprint of ground-based AI infrastructure has become a local-politics flashpoint. Musk's pitch in the X interview was, in part, that putting compute above the atmosphere sidesteps the terrestrial opposition. Whether the FCC reads that as a reason to expedite, or as a reason to demand unusually rigorous debris and atmospheric-chemistry answers, is a judgment call the agency has not yet made.
Reentry chemistry is the part the headline skips. Burning one million satellites, or even a meaningful fraction of them, over decades puts aluminum-oxide and other combustion products into the mesosphere and thermosphere. The atmospheric-chemistry community is still working out what sustained, large-scale reentry deposition does to ozone and to upper-atmospheric radiative balance. That is not a rhetorical question. It is a research gap the filing does not address and the FCC is not, by tradition, equipped to evaluate on its own.
So the watchlist is short and specific. The FCC docket will show SpaceX's proposed shell altitudes, the disposal lifetime the company is willing to commit to, the collision-avoidance architecture for the proposed scale, and the agency's own environmental and atmospheric review. The space-is-really-big line is a slogan, not a model. The model is in the filing, and the filing is now public.