Premiums in orbit

Published 1 October 2026

Sixteen chapters of this file have been about insurance at its most human - the widow, the farmer, the patient, the saver. This one is about insurance at its strangest: a global market of about two dozen specialist underwriters who price the risk of explosions on launch pads and dead satellites in orbit. It is worth a chapter for one reason. Stripped of salesmen, slogans, trust and tradition - stripped of everything except risk and price - space insurance is what the business looks like when it is honest. The premiums are absurd. The claims are paid. And the comparison with the machine back on the ground is the most unflattering one in the file.

Start with the prices, because they are magnificent. A commercial launch-plus-in-orbit policy for a large satellite has historically cost on the order of 15-25% of the sum insured - and above 30% in the years after the loss-strewn 1980s. Imagine your endowment agent quoting that: a fifth of everything, up front, no bonuses, and the rocket might still blow up. Yet satellite operators pay it, year after year, because the pricing is doing the thing insurance pricing is supposed to do: it reflects the risk, moves with the losses, and nobody pretends otherwise. There is no agent assuring anyone the launch is guaranteed. There is a rate, a deductible, a capacity market, and a claims record.

The claims record is the part that would embarrass the ground business. Aon's market report lays the last four years out like a cardiogram. In 2023, the market paid $1.43 billion of claims against about $550 million of premium income - a catastrophic year, losses nearly three times revenue. In 2024, with few claims, the same market earned a profit of about $383 million. In 2025, premium income rose past $650 million, claims ran to about $503 million - including the total loss of the SPAINSAT NG-II communications satellite, hit by a space particle on its way to orbit, and about $30 million each for two smaller failed spacecraft - and the market finished with a profit near $147 million, a loss ratio of 77.4%. Read those numbers as a policyholder: in this market, three of every four premium rupees go back out as claims in an ordinary year, and in a bad year the claims run to nearly triple the premiums. Nobody's claim is denied because the satellite failed to disclose a pre-existing orbit.

The market with no one to fool

Why does this corner of the business behave so differently? Because it has none of the machine's ingredients. The buyers are satellite operators with engineering teams and lawyers; the sellers are syndicates with actuaries; the product is a price, negotiated between professionals who both understand the risk; and the distribution cost is a broker's fee, not a third of the first premium plus a convention in Bangkok. When losses spike, rates rise and capacity shrinks; when losses fall, rates soften and capacity returns. It is volatile, occasionally irrational, and roughly honest - the five-year margin went negative for the first time since 2001 after the 2023 losses, and the report says so in print, because the readers of that report are the people pricing next year's rockets.

The same industry that sells a 5% endowment plan to a schoolteacher with a straight face also participates, in its professional wing, in a market where a 77% loss ratio is a decent year and the product's entire value is that claims get paid. The difference between the two is not actuarial talent - the industry has plenty. It is the customer. When the customer can read the risk, the product gets honest. When the customer is a schoolteacher being sold tradition by her neighbour, the product is what the earlier chapters described. Space insurance is the control experiment: proof that the machine's dishonesty is not in insurance itself, but in the distance it engineers between the person who pays the premium and the person who prices the risk.

Capacity: the honest thermostat

The mechanism that keeps space insurance honest deserves a name, because the ground industry has spent seventy years disabling it: capacity. Only so many underwriters will write rocket risk, and each declares, ahead of the year, how much loss it is willing to absorb. Aon's report tracks the number the way the ground industry tracks bonus rates - capacity per launch risk, new entrants adding $25 million here, $25.5 million there, total market capacity rising about $100 million year on year. When claims spike, capacity withdraws, rates harden, and buyers pay the true price of a bad year. When claims are quiet, capacity returns and rates soften. The price of risk breathes. Nobody calls a hard market a conspiracy or a soft one a scam; it is a market doing the one thing a market is for.

Now hold that against the ground machine. Indian retail insurance pricing breathes too - but in one direction. Medical inflation runs 13%, premiums rise to match, and the customer's sum insured halves in six years; the one number that never flexes is the distribution take, which the regulator's own figures show growing four times faster than the sales it generates. A rocket's premium falls when rockets stop exploding. A family's premium never falls when the family's risks improve, because the machine's first claim on the premium is not the risk at all - it is the cost of having been sold to. The space market's buyers would not tolerate that for a single renewal season; they have actuaries too, and alternatives, and the leverage to walk. The Indian household has none of those, and the machine's entire architecture - the agent, the lock-in, the surrender penalty, the thirty-year term - exists to make sure it never gets them. The distance between Mumbai's retail pitch and the London space market is not geography. It is leverage. And leverage, in insurance as in everything else, is the difference between a price and a tax.