On February 10, 2009, at an altitude of about 790 kilometres above Siberia, two satellites met at a combined speed of roughly 11.7 kilometres per second. One was Iridium 33, an American communications satellite in active service. The other was Cosmos 2251, a Russian military satellite dead for fourteen years. Neither could move out of the way. The collision shredded both spacecraft and threw more than 2,000 pieces of trackable debris into orbits that other satellites use every day. It was the first accidental hypervelocity collision between two intact satellites — and nobody paid anybody anything. No court heard the case. No claim was filed. No fine was levied. Seventeen years later, that is still, in substance, the state of the law.

The traffic has not stood still. The European Space Agency's 2025 Space Environment Report counted about 40,000 objects tracked in orbit, of which roughly 11,000 were active payloads, and estimated more than 1.2 million fragments larger than one centimetre — each one big enough to end a mission. By October 2026, ESA's regularly tracked count had risen to about 47,000 objects. More than 300 launches in 2025 put over 4,000 new payloads into orbit, about eleven a day. Starlink alone had 11,207 satellites in orbit by November 7, 2026, according to astrophysicist Jonathan McDowell's independent catalogue, out of 13,096 launched. The constellation performed roughly 300,000 automated collision-avoidance manoeuvres in 2025 — about forty per satellite per year — and the number keeps climbing.

Every one of those manoeuvres is a cost somebody absorbs: propellant burned, observation time lost, operations staff on watch. The question the industry has deferred for a decade is who pays when avoidance fails — and the answer, so far, is that the legal system has barely begun to work it out.

The rulebook, such as it is

The most concrete rule in force is American, and it is four years old. In September 2022 the Federal Communications Commission adopted FCC 22-74, replacing a voluntary 25-year post-mission disposal guideline with a binding requirement: satellites in low Earth orbit must be disposed of within five years of the end of their mission. The rule applies to satellites launched after September 29, 2024, and to any foreign system seeking access to the US market. It is performance-based — the FCC does not tell operators how to bring a satellite down, only that they must — and it bites at the licensing stage, where a disposal plan that does not close can mean a denied application.

Europe moved in the same direction, harder. The European Space Agency's 2023 debris-mitigation requirements also adopt a five-year clearance for low Earth orbit, add a 90 percent disposal-success probability requirement, and feed into a Zero Debris Charter under which signatories commit to generating no new debris by 2030. The old 25-year guideline from the Inter-Agency Space Debris Coordination Committee still exists on paper. In practice, the five-year standard is now the price of admission to the two largest licensing regimes in the world.

The first enforcement, though, was small and strange. On October 2, 2023, the FCC fined Dish Network $150,000 — its first space-debris penalty — for failing to properly dispose of EchoStar-7, a direct-broadcast satellite launched in 2002. Under a plan approved in 2012, Dish had committed to raising the satellite 300 kilometres above its geostationary operating arc at end of life. It ran out of fuel in February 2022 and managed only 122 kilometres. Dish admitted liability and agreed to a compliance plan. The fine was a rounding error against a company of Dish's size, and the FCC's enforcement chief called it a breakthrough anyway. Both things were true: the agency had shown it would act, and it had shown the price of getting caught was trivial.

The treaty that has been tested once

Above national regulators sits a treaty regime written when orbit was nearly empty. The 1967 Outer Space Treaty makes launching states internationally liable for damage their space objects cause. The 1972 Liability Convention turns that principle into rules: a launching state is absolutely liable — no proof of fault required — for damage its object causes on the surface of the Earth or to aircraft in flight, and liable on a fault basis for damage caused in orbit.

The Convention has been formally invoked exactly once. In January 1978, the Soviet nuclear-powered satellite Cosmos 954 broke up over Canada's Northwest Territories, scattering radioactive debris across tens of thousands of square kilometres. Canada claimed C$6,041,174.70 for the search and cleanup. The two governments settled in April 1981 for C$3 million, with no admission of liability and no claims commission ever convened. The machinery functioned, after a fashion — and the fashion was diplomacy, not adjudication.

That distinction matters because an in-orbit collision is a fault case, and fault in orbit is almost impossible to prove. Two satellites meet at seven kilometres per second with no witnesses, no black box, and telemetry that belongs to the parties in dispute. Who was at fault — the operator that failed to manoeuvre, the one whose dead satellite drifted into the lane, the state that licensed them both? No in-orbit liability claim has ever been tested, let alone won. The 2009 Iridium-Cosmos collision produced no claim at all. Until one is, every constellation valuation carries an unpriced contingent liability: a collision that destroys a neighbour's billion-dollar asset, with the bill arriving by diplomatic note rather than insurance adjuster.

Orbit has traffic rules, a speed limit, and a fine schedule — but no court that has ever decided a crash.

The insurers have already priced it

If the law is silent, the insurance market is not. The global space insurance pool is tiny — a little over $650 million of premium in 2025, according to broker Aon — and it is brutally honest about risk. In 2023, claims of about $1.43 billion swamped roughly $550 million of premium. Two satellites did most of the damage: Viasat's ViaSat-3 Americas, a roughly $1 billion spacecraft whose antenna failed to deploy, generating a claim of about $445 million; and Inmarsat's 6-F2, declared a likely total loss after a battery failure, at about $348 million. Launch-plus-one-year rates that had been under 6 percent of insured value at the start of 2023 were around 10 percent by year-end. Three underwriting syndicates — Hiscox, Canopius, and Munich Re — exited the market in 2024.

The market has since steadied. Premium above $650 million against claims of about $503 million in 2025 produced a loss ratio of 77.4 percent, and 25 insurers were writing space risk in early 2026. But the structure of the market tells the real story. The most capacity any single satellite can buy in orbit is about $375 million. And fewer than 1 percent of active low-Earth-orbit satellites carry in-orbit insurance at all — roughly 300 of about 10,000. The constellations that create most of the collision risk self-insure: a Starlink satellite built for roughly $400,000 cannot justify a premium at 1 to 3 percent of its value. The risk never reaches an underwriter. It sits on the operator's balance sheet, unpriced and undisclosed.

Then, in late 2025, a particle of a few grams ended the mission of SpainSat NG II, a satellite reported to be insured for about $400 million. A fragment too small to track, too small to dodge, too small to regulate — and a claim large enough to absorb most of a year's premium on its own. That is the shape of the risk: not a dramatic collision between two named satellites, but a grain of debris, moving at orbital speed, meeting a spacecraft that did nothing wrong.

The honest counterargument

The case for calm is stronger than the debris alarmists admit. The system mostly works by prevention, not punishment, and prevention is improving: more than three intact satellites or rocket bodies fall back to Earth every day, and the five-year disposal rule means the new population is designed to leave. Low orbits clean themselves — atmospheric drag at 500 kilometres does in years what takes centuries at 800. The collision risk is concentrated in a narrow band around 550 kilometres, where the constellations fly, and that is also where the operators with the best tracking and the most manoeuvres fly. No Kessler cascade — the runaway chain reaction of collisions breeding collisions — has ever been observed. It remains a model, and models of complex systems have a habit of overestimating the worst case.

The regulatory gaps are real, but they may also be the price of a market that works. A $150,000 fine is trivial because the FCC's leverage is the licence, not the penalty: lose the licence and you lose the business. The Liability Convention has been invoked once in 54 years because states settle quietly rather than litigate — and quiet settlement is still settlement. And the insurance market's small size reflects a genuine fact: most satellites in orbit are cheap, replaceable, and self-insured by operators who can afford to lose one. The system is not elegant. It may not need to be.

There is also a cost to over-regulation that the debris debate rarely prices. A five-year disposal rule forces operators to reserve propellant, carry extra mass for drag devices, or fly lower than the orbit their business wants. For a small operator, that compliance cost is the difference between a viable business and none.

What it means for capital

For an investor, the debris question is a valuation question wearing a regulatory costume. Three things now belong on every constellation diligence list. First, the disposal plan: does the spacecraft carry the propellant and the reliable propulsion to come down in five years, and has the operator demonstrated it, or is the plan a paragraph in a licence application? Second, the insurance position: is the risk transferred, self-insured with real reserves, or simply unacknowledged? A constellation of thousands of uninsured satellites is a balance sheet with an unquantified tail. Third, the licensing jurisdiction: an operator licensed in a regime with a five-year rule and an active regulator faces a different risk profile from one licensed where the rules are guidelines.

The removal business is the other side of the same trade. Astroscale's ADRAS-J mission rendezvoused with a spent rocket stage in 2024, proving that a commercial vehicle can approach large debris. Governments are beginning to pay for removal as a service. But the market is still measured in demonstration contracts, not revenue, and the operator with a dead satellite it cannot move has no legal obligation to buy.

The bill for a crowded orbit will arrive in one of three forms: a collision claim that tests the Liability Convention for the first time in orbit, a regulatory penalty large enough to matter, or an insurance loss that reprices the whole market overnight. None of the three has happened yet. That is precisely why it is cheap to prepare for now, and why it will be expensive to ignore when the first one lands. The investors who price debris risk before the market does will own the constellations that survive it.

The category-defining name in space economy capital — spaceeconomycapital.com — is currently available for acquisition by the firm positioned to own this space.

Sources

European Space Agency, "Space Environment Report 2025," and ESA space debris statistics, October 2026 — via theworlddata.com and startup.ph. Jonathan McDowell, Jonathan's Space Report, Starlink statistics, November 7, 2026 — via contentlora.com. FCC 5-year deorbit rule, FCC 22-74 — incognito.space and viventine.com. Reuters, "DISH gets first-ever space debris fine over EchoStar-7," October 2, 2023 — kwsn.com; Foot Anstey, "Legal precedent in the stars," October 13, 2023 — footanstey.com. Orbital Radar, "The Liability Convention: Who Pays for Space Damage" — orbitalradar.com; LawStories, "Cosmos 954 Incident," 1978 — lawstories.in. Aon, "Space Insurance Market Report, Q1 2026" — via setcoin.group and spacelawjournal.com. New Space Economy, "The Orbital Insurance Market," March 29, 2026 — newspaceeconomy.ca.