A preliminary study estimates that SpaceX’s proposed orbital data-center constellation could sharply increase rocket emissions and pollution from satellites reentering Earth’s atmosphere. The figures are model projections for a proposed network of up to one million satellites—not measurements from an operating constellation. The repository hosting the study labels it not peer reviewed.
What the researchers estimate
In a version 1 preprint dated October 30, 2026, University College London researchers Eloise A. Marais and Connor R. Barker estimate that the proposal could raise first-order estimates of total launch emissions and reentry nitrogen oxides to 70–150 times contemporary levels. They estimate a 200–280-fold increase in reentry mass. These are separate measures, each with its own baseline; the preprint’s abstract does not give an absolute annual nitrogen-oxide tonnage.
Nitrogen oxides, often shortened to NOx, are gases produced by combustion and other high-temperature processes. When satellites burn up during reentry, their material can also introduce metals and other compounds into the upper atmosphere.
What the launch scenario assumes
A more detailed scenario reported by Scientific American assumes planned Starship Block 4 launches carrying at least 200 metric tons per flight. On that basis, deploying the proposed constellation could require at least 5,000 launches a year—about 14 a day. That is a modeled cadence, not a SpaceX launch schedule.
The same report gives an estimated 233,000–507,000 metric tons of reentry nitrogen oxides per year. The range draws partly on limited experimental measurements and estimates informed by NASA Space Shuttle studies, so it depends on assumptions beyond the preprint abstract’s headline multiplier.
The reported estimate for satellite material vaporizing in the atmosphere is about 200,000 metric tons a year if the 20 percent of satellites not moved to higher “graveyard” orbits burns up. The scenario assumes SpaceX plans to move 80 percent of the satellites into those orbits.
Why the estimates remain conditional
Rocket propellant use and satellite design affect what enters the atmosphere. The preprint says metal emissions from satellite ablation—the heating and breakdown of material during reentry—cannot be quantified from the available composition information. Scientific American also reports that an estimate of about 60,000 metric tons of methane escaping per year depends on Starship continuing venting behavior observed during test flights.
The study discusses possible effects involving ozone chemistry, stratospheric aerosols, climate, nitrogen in the mesosphere and air quality. Those pollutants can interact, making their combined planetary effects difficult to predict from a single emissions estimate.
The size of a projected emissions increase is only one part of the picture; the launch and disposal assumptions determine what that number describes.
The FCC’s October 6, 2026, order covers a separate SpaceX direct-to-device system of 15,000 satellites, whose application it granted in part and deferred in part; it is not an authorization for the proposed orbital data-center constellation.