Study: Proposed 1‑Million‑Satellite Orbital Data Centers Could Severely Disrupt Vera C. Rubin Observatory

A newly posted arXiv preprint forecasts that a proposed orbital data center constellation on the scale of roughly 1 million satellites could severely disrupt observations by the NSF-DOE Vera C. Rubin Observatory. The study is a simulation, not a report of effects from an operating system, but its modeled results are stark: a synthetic constellation built from public filing parameters would interfere with a large share of Rubin’s planned sky survey observations.

In the paper, posted Oct. 22 as arXiv:2610.31772v1, the authors write that “On a given night, up to 80% of LSST pointings would contain at least one illuminated ODC satellite streak.” Over a full year, the average share of contaminated exposures is closer to 20%, because visibility is strongly seasonal and worst near twilight. The study also says the satellites would raise sky brightness “by more than 1000% near twilight with no Moon,” and it estimates typical nightly mean pixel losses near twilight of 5% to 10%, even with multiple mitigation strategies.

That matters because Rubin, a joint project of the National Science Foundation and the Department of Energy, is designed to conduct the Legacy Survey of Space and Time, or LSST. The observatory will repeatedly image wide areas of the sky to detect faint objects and changing events over time. That wide-field, high-cadence strategy makes it especially vulnerable to bright satellite streaks crossing images and to a brighter background sky that can wash out faint signals.

The preprint, “The Impact of Orbital Data Centers on the NSF-DOE Vera C. Rubin Observatory,” is by Nayan Jangid, Meredith L. Rawls, Peter Yoachim, Connie Walker and Siegfried Eggl. Affiliations listed on the paper include the University of Illinois Urbana-Champaign, Vera C. Rubin Observatory and the University of Washington, and NSF NOIRLab. The researchers modeled one year of Rubin observations against a full constellation of 998,240 satellites, based on public FCC filing parameters cited in the paper, including a SpaceX orbital data center application and a summary by astronomer Jonathan McDowell.

Orbital data centers, in this context, are proposed constellations of satellites meant to provide computing or data-center capacity in space. Astronomers have raised concerns throughout 2026 that systems built at that scale could create a new level of optical interference for ground-based telescopes.

The analysis uses a six-stage simulation pipeline that tracks satellite orbits, checks whether they cross Rubin’s field of view, evaluates sun and observer geometry, models satellite brightness, calculates added surface brightness and estimates lost pixels. The authors also say their code and pipeline are publicly archived on GitHub and Zenodo, allowing other researchers to inspect or reproduce the work.

Among the paper’s more notable findings is that contamination is not evenly distributed across the simulated system. The authors say sun-synchronous orbital shells account for a disproportionate share of the problem because their geometry keeps satellites illuminated more often from Rubin’s point of view. The paper also reports a multi-modal brightness distribution, with specular reflections from solar panels appearing as bright as 2 to 4 AB magnitude, while more diffuse reflections are typically around 6 AB magnitude.

The authors also stress that the modeled constellation is synthetic and idealized. It is a shell-based, steady-state orbital catalog rather than a prediction of an actual moment-by-moment deployed fleet. The model does not include every operational behavior, such as deployment transients or maneuvers, and the paper explicitly says it should not be read as a forecast of an exact operational configuration. Even so, the authors conclude that “Even at 10% deployment, the constellation would still produce widespread visible streak contamination throughout the year.”

The regulatory backdrop is still unsettled. SpaceX filed an FCC application for an orbital data center system in January 2026, and the Federal Communications Commission accepted it for filing on Feb. 4, opening a public comment process. Acceptance for filing is not approval. As of mid-October 2026, according to the research summary cited in the paper, no final FCC authorization had been posted for the proposed 1 million-satellite system.