Low Earth Orbit Just Got Too Expensive to Insure

In 2018, if every satellite operator in low Earth orbit simultaneously lost the ability to steer, engineers estimated it would take about 121 days before the first cascading collision. By mid-2025, the same calculation — insurers call it the "CRASH clock" — had compressed to somewhere between 2.8 and 5.5 days.

That is not a rounding change. That is the margin of safety in the most valuable band of orbital real estate collapsing by a factor of more than twenty in seven years. And the people who noticed first were not the rocket engineers or the regulators. They were the underwriters.

Space is getting crowded in a way the balance sheets have not caught up to. The European Space Agency's 2026 environment report now catalogs more than 40,000 tracked objects, with modeling that puts the true population above 54,000 pieces larger than 10 centimeters — each one large enough to end a mission. Below that threshold sit an estimated 1.2 million fragments in the "lethal but non-trackable" range, and roughly 140 million shards under a centimeter, any of which can punch through a spacecraft at orbital velocity. At 550 kilometers — the prime Starlink altitude — the density of dead debris is now approaching parity with the density of working satellites.

The operational cost of living in that environment is already visible. SpaceX told the FCC that its Starlink fleet performed roughly 300,000 collision-avoidance maneuvers in 2025, a 50% jump from the year before, and roughly 40 dodges per satellite per year across a constellation of about 10,000. Projections have that number crossing a million maneuvers annually by 2027. Every one of those maneuvers burns propellant, shortens a satellite's working life, and — this is the part the market is only now pricing — represents a near-miss that an insurer has to model.

The quiet repricing

Here is the shift that matters for money. The global space insurance market is not large — about $4.43 billion in premium in 2026, up from $4.06 billion the year before. Lloyd's of London remains the dominant hub, with European syndicates like Beazley and Hiscox controlling roughly a third of global capacity. For decades this was a boutique, clubby corner of the specialty market: a few dozen launches a year, big single-satellite policies, losses that came from rockets exploding on the pad rather than from anything that happened once you were on orbit.

That model was built for a sky with a few hundred active satellites. It is now being asked to underwrite a sky heading toward 100,000 — the number ESA modeling suggests LEO can sustainably hold, a threshold some analysts think we reach before 2050, not after. The risk has changed shape. It is no longer discrete and one-time (did the launch work?) but continuous and systemic (will this satellite spend five years dodging everyone else's hardware, and does a single cascade event wipe out an entire orbital shell at once?).

Briefings like this land in members' inboxes before the market prices them in. Join free →

Underwriters cannot diversify away a cascade. If the Kessler scenario — collisions begetting debris begetting more collisions — ever tips at 550 kilometers, it does not take out one insured asset. It takes out the entire orbital neighborhood the insurer has written policies across, all at once, with correlation of one. That is the definition of an uninsurable risk, and the market knows it. Premiums in the most congested bands are rising, capacity is tightening, and Munich Re, Swiss Re, and the Lloyd's syndicates have started wiring real-time orbital tracking and predictive collision models directly into their pricing — because a static actuarial table is now worthless against a sky that reshuffles every few hours.

When an insurance market starts pricing a risk it cannot fully lay off, one of two things happens. It withdraws capacity — or it goes looking, urgently, for whoever can sell it better data. Both are already underway. And that second move is where the actual trade is.