On Thursday, the European Southern Observatory published a study with a number attached: 100,000. That, the ESO argues, is the absolute ceiling on satellites in low Earth orbit if professional astronomy is to remain viable — a figure dwarfed by the 1.7 million spacecraft already proposed in various FCC and ITU filings. The study, submitted as a formal response to U.S. regulators, marks the first time a major observatory has translated the “megaconstellation problem” into a specific, quantitative demand.

It is good science. It is also a political document dressed in photometric units, and the politics deserve a harder look than most coverage has granted them.

The immediate villain in the ESO’s framing is Reflect Orbital, the California startup that wants to put 50,000 mirror-equipped satellites in orbit to sell “sunlight on demand” — beams five kilometers wide, aimed at paying customers on the ground. The ESO’s simulations suggest that Reflect Orbital’s full constellation alone would increase scattered background light across the entire sky, not just over illuminated patches, by enough to compromise wide-field survey telescopes. Add SpaceX’s ambitions, Amazon’s Kuiper, and the Chinese Guowang network, and the sky begins to glow with a diffuse anthropogenic haze that no remote mountaintop can escape.

Bad for astronomy. Genuinely. I am not here to dispute the physics.

I am here to ask who the ESO expects to enforce its limit — and what those enforcers are supposed to say to everyone else who wants the bandwidth.

The Cap Is a Demand Disguised as a Finding

The study’s headline recommendation is that satellites should be kept below naked-eye visibility and that no more than 100,000 should be permitted. This is stated as a finding about what astronomy requires. It is actually a demand about how the orbital commons should be allocated. And the institutions best positioned to impose that allocation — the FCC, the ITU, national telecom regulators — have mandates that do not remotely prioritize professional stargazing.

The FCC’s statutory obligations, to take the most relevant example, center on spectrum efficiency, consumer access, and national interest. The agency can weigh environmental and scientific concerns, but it has never treated them as dispositive. When SpaceX applied for its first 4,400 Starlink satellites, the FCC’s primary concern was interference with other radio services. Optical interference with ground-based telescopes entered the record as a footnote, not a blocking condition.

The ESO is effectively asking regulators to elevate an aesthetic and scientific claim — the integrity of the night sky — above competing claims for broadband connectivity, defense surveillance, and, yes, venture-funded gimmicks like orbital mirrors. That may be the right outcome. But calling it a “limit” doesn’t solve the enforcement problem. A limit implies someone will say no. Who, exactly?

“It’s a beautiful number,” an engineer at a competing launch provider told me in a Slack message Friday, on condition I not name the firm. “Zero chance the FCC uses it as a bright line. They’ll take the study, thank the ESO, and keep approving constellations on a case-by-case basis, same as always.”

Someone Has to Pay for the Dark

There is a more uncomfortable question the ESO study does not address. If the world decides that 100,000 satellites is the limit, what happens to the services the other 1.6 million would have provided? Some fraction are pure spectacle — Reflect Orbital’s mirrors are a solution in search of a problem, and a particularly obnoxious one at that. But many of the proposed constellations aim to deliver broadband to places terrestrial fiber will never reach, or to provide low-latency connectivity for aviation, maritime shipping, and remote industry.

The astronomy community has a habit of framing this as a trade-off between science and corporate greed. That framing works in op-eds. It works less well when you are a regulator fielding calls from a senator whose rural constituents just got their first usable internet connection via a satellite terminal that cost them nothing to install.

If we are going to cap orbital infrastructure, we need to be honest about what we are capping and who bears the cost. The ESO’s study — meticulous on photon counts, silent on economics — reads like the opening position in a negotiation where one side has not yet acknowledged it is at the table.

The Planetary Scale Problem Scientists Keep Avoiding

The final dodge is jurisdictional. Low Earth orbit is not governed by any single treaty with binding enforcement power over commercial actors. The Outer Space Treaty of 1967 requires states to authorize and supervise their nationals’ activities in space, but it says nothing about cumulative environmental effects. The Moon Agreement has fewer than 20 parties. The ITU coordinates spectrum; it does not regulate reflectivity or orbital debris by volume.

So when the ESO says the limit should be 100,000 satellites, it is speaking to a regulatory apparatus that doesn’t exist. The FCC can regulate U.S. licensees. The ITU can broker agreements. No single body can cap the total number of objects in orbit, and no existing treaty gives astronomy a veto.

This is the structural problem hiding beneath the photometric precision. The astronomers have done the math. They have not done the diplomacy — and they have not reckoned with the fact that the constituencies that benefit from orbital infrastructure are larger, more dispersed, and more politically legible than the constituencies that benefit from dark skies. A billion people with satellite broadband is a statistic. A Chilean observatory with degraded image quality is a line item in a European funding review.

No Cap Without a Constituency

The ESO’s study is rigorous, important, and doomed unless its authors build a political coalition that treats a dark sky as something worth paying for. That means lobbying not just the FCC but national governments. It means forming alliances with circadian-rhythm researchers, wildlife biologists, and public-health officials — coalitions that the Guardian reported in April are already forming. It means, crucially, showing up with an answer to the question “what about the broadband?” that is more sophisticated than “that’s a separate issue.”

I suspect the astronomers know this and are playing the long game — get the number into the regulatory record now, build the coalition over the next decade. Fine. But let’s not pretend the number itself does the work. A limit nobody enforces is just a wish with units.

Sources