Orbital data centers move the siting fight from the county board to the night sky

Astronomers led by Aaron Boley of the University of British Columbia posted a paper to the arXiv preprint server on August 3 modeling how bright the sky becomes if the orbital data-center constellations already filed with the U.S. Federal Communications Commission actually fly ^[1][2]. In the most extreme configuration they examined, and absent deliberate work to dim the hardware, the satellites would outnumber the stars visible to the naked eye by a factor of 100 ^[3].

The commercial logic is straightforward and worth stating plainly. Terrestrial data centers consume large amounts of electricity and strain local resources, and proposals for new ones are increasingly met with local opposition ^[4]. Putting solar-powered compute in orbit is pitched as a way around those objections and a way to move the environmental cost off the ground ^[5]. SpaceX, Blue Origin and the start-up Cowboy Space Corporation have filed FCC proposals for data-center constellations ranging from 20,000 to one million satellites ^[6] - a span of a factor of 50 between the modest and the maximal versions of the same idea ^[7].

The orbit that makes the economics work is the same one that makes the optics bad. The most efficient and therefore most sought-after orbit runs over the poles along the cusp between day and night, where satellites receive constant sunlight ^[8]. That is exactly the geometry that keeps large solar arrays lit while the ground beneath them is dark.

Boley and colleagues modeled three proposed layouts: Cowboy Space's single ring, Blue Origin's two crisscrossed rings, and SpaceX's crisscrossed rings plus an additional shell around Earth ^[9]. They also varied the spacing between satellites, which is a real trade: a looser spread lowers collision risk but sweeps over more of the sky, producing a broad blanket of light rather than discrete rings ^[10]. Regardless of configuration, the modeling showed the sky would be profoundly changed ^[11]. From lower latitudes the parade would peak in brightness at sunrise and sunset; nearer the poles the constellations would be bright and visible for most of the night in winter and around the equinoxes ^[12]. "People have lost the sense of the night sky because of the light-polluted cities, and now we're rewriting that story with artificial satellites," Boley says ^[13].

For observatories the damage is not confined to optical astronomy. The satellites would crowd out fainter objects and, per the study, could rival the Milky Way's prominence with their glare ^[14]. Antarctica is the sharpest case, where observatories exploit dry remote conditions and round-the-clock polar-winter darkness to study faint signals such as the cosmic microwave background ^[15]. Heat emissions would interfere with infrared work, satellite communications with ground stations and other spacecraft would affect radio astronomy, and space telescopes would find satellites in their fields of view ^[16][17][18]. There is precedent for mitigation: SpaceX and the National Radio Astronomical Observatory share data so NRAO telescopes can stop collecting when Starlink passes overhead ^[19]. The study's finding is that at some configurations the sheer number of objects makes that avoidance impossible ^[20]. The authors also flag effects on animals that depend on dark-sky cues ^[21].

Two things to watch. First, whether brightness mitigation shows up as a condition in FCC licensing rather than a voluntary courtesy, since the paper's worst case is explicitly the unmitigated one ^[3]. Second, whether the filings converge toward the 20,000-satellite end or the one-million end ^[6]; the difference is the difference between an annoyance and a permanent fixture.