Three or four buildings within walking distance of Lloyd’s of London contain the people who functionally decide whether a satellite constellation gets built, whether a new launch vehicle gets a second customer, and whether a startup with a clever propulsion idea ever leaves the slide deck. They do not wear mission patches. They write policies, set premiums, and say no a great deal more often than they say yes.
The space industry likes to tell its story through engineers and founders. The actual gating function sits in underwriting rooms in EC3.
The quiet veto
In my time working in the space sector I watched dozens of mission concepts die. Almost none of them died because the physics was wrong. They died because someone in finance could not get a quote, or could only get one at a premium that turned the business case into a wet rag.
Space insurance is a small market. Industry estimates suggest it writes somewhere in the order of a billion dollars in premium a year, which is rounding error against energy or marine. But it is concentrated in a way those larger markets are not. A handful of syndicates at Lloyd’s, a few continental European reinsurers, and a thin layer of specialist brokers handle the bulk of launch and in-orbit cover. When two or three lead underwriters in London decline a risk, the risk often becomes uninsurable at any sensible price.
That is the quiet veto. It does not show up in press releases. It shows up as projects that quietly slip from Q3 launch to indefinite delay to cancellation.
Why London, and why so few people
London’s dominance in specialty insurance is not an accident of history so much as a compounding one. The same underwriting culture that reportedly grew up around marine cargo in the eighteenth century absorbed aviation in the twentieth and space in the late 1960s. The skills transfer cleanly: low-frequency, high-severity events with imperfect data and a need for syndicated capacity. Recent moves by major carriers continue to treat London as the central pillar of their specialty strategies, even as Singapore, Dubai and Houston grow as hubs.
The number of human beings who actually underwrite space risk is astonishingly small. Market observers estimate the number of senior space underwriters worldwide at well under a hundred. Many of them know each other. Most have worked at two or three of the same firms over their careers. They lunch together. They share a vocabulary, a memory of which booster failed in 2007, and a healthy distrust of optimistic loss models.
This is the room where commercial space gets its real safety review. Not the regulator. Not the customer. The underwriter.
What 2023 broke
For a long time the market was soft. Premiums reportedly fell through the 2010s as launch reliability improved and capacity flooded in. Then came a stretch of losses that made 2023 a genuinely bad year for the sector: a string of in-orbit failures, a couple of expensive launch losses, and the unwelcome discovery that some satellite buses had design flaws that only manifested after years of service.
Premiums hardened sharply. Capacity withdrew. Several syndicates exited space entirely. The ones that remained became more selective, and this selectivity in a market with five lead underwriters meant many operators suddenly found themselves shopping a risk that nobody wanted to lead.
That hardening has not really unwound. Operators launching today are paying premiums that would have looked punitive five years ago, and they are accepting policy structures with more exclusions, lower sublimits, and tougher conditions on configuration changes mid-build.
How the veto actually works
An underwriter rarely says no outright. The mechanism is quieter and more effective.
A broker takes a risk to the market. The lead underwriter, the one expected to set terms and take the largest line, asks for engineering documentation: flight history on the launch vehicle, qualification data on the satellite bus, the operator’s procedures for collision avoidance, the financial backing of the company. Then they quote a rate. If the rate is high enough that the operator’s business model collapses, the project dies in spreadsheet form, not in a meeting.
If the lead won’t quote at all, the broker tries another. If three or four of the obvious leads pass, the message has been received. The risk is uninsurable in the conventional market. Some operators self-insure or go without cover, which is increasingly difficult because launch providers, lenders and government customers all require minimum coverage levels.
This is how a few dozen people functionally referee the industry. Not by writing rules, but by pricing risk in a way that either does or does not let a project breathe.
The Vulcan problem
Look at what is happening with United Launch Alliance’s Vulcan rocket and you can see the dynamic operating at the institutional level. The Space Force is openly reconsidering its launch manifest after Vulcan suffered two solid rocket booster anomalies in less than two years, with Lt. Gen. Philip Garrant saying the experience will “absolutely” shape the next launch services competition.
The Pentagon is its own underwriter in a sense; it carries the risk on its own missions. But for every commercial customer thinking of buying a Vulcan ride, the calculus runs through London. A second nozzle failure on the same booster type does not simply raise the premium for the next Vulcan launch. It raises questions about every payload manifested on that vehicle for the next eighteen months. Brokers start asking for additional engineering reviews. Underwriters start attaching conditions. Some lines withdraw until the root cause is published.
None of that is announced. It shows up as a customer quietly choosing Falcon 9 instead, or as a satellite operator deferring a launch to wait for clarity. The insurance market is the transmission mechanism that turns engineering anomalies into commercial consequences.
What the underwriters actually look at
Spend any time around space underwriters and you notice they are less interested in the things that excite engineers and more interested in the things that worry auditors.
Heritage matters more than novelty. A booster with twenty successful flights gets cheaper cover than a new vehicle with a brilliant design and three flights, even if the new vehicle is technically superior. This is rational from an actuarial standpoint and infuriating if you are trying to bring something new to market.
Configuration discipline matters. Underwriters hate change. A satellite that flies the same bus the manufacturer has flown forty times before is straightforward. The same bus with a new payload, a new propulsion module and a new ground software stack is three separate sources of uncertainty bundled together, and the premium reflects it.
Operator track record matters. A first-time operator pays more than an established one for the same hardware. This is one reason new entrants struggle even when their technology is sound; they are paying a premium for not yet having a claims history that demonstrates how they handle anomalies.
Ownership structure matters in ways that are easy to underestimate. The growing complexity around foreign ownership mitigation in commercial space means underwriters now want to understand who actually controls a company, not just who is on the cap table. Cross-border ownership can complicate export control, regulatory standing and ultimately the enforceability of a claim. Risks that look identical on paper can price very differently depending on the corporate structure behind them.
The startup that never launches
The companies hurt most by this regime are the ones least equipped to talk about it publicly. Industry observers note that a small operator with a clever idea, a thin balance sheet and no flight heritage can face premiums that run into double-digit percentages of the insured value for launch alone. In-orbit cover for the first few years can add another several percent annually.
For a constellation operator with hundreds of small satellites, those numbers might be tolerable spread across the fleet. For a single bespoke satellite worth a hundred million dollars, the insurance line item can be larger than the launch cost. Investors do the maths. Boards quietly conclude that the unit economics do not work. The company pivots to a service model, or it folds.
I have seen this happen to teams whose engineering was excellent. The veto is not a judgement on their work. It is a judgement on the actuarial profile of doing something new with limited heritage and limited capital, and the underwriter is paid to be conservative about exactly that combination.
The state customer changes the equation
One reason the commercial market matters so much is that government customers have, until recently, behaved as though insurance were someone else’s problem. The Pentagon’s own analysis suggests it is happy to buy commercial launch services at scale but reluctant to budget meaningfully for non-launch commercial services, which leaves a lot of new entrants chasing a thin sliver of the addressable market.
That matters for insurance because government anchor contracts are how startups build the flight heritage that eventually lowers their premiums. Without a stable government demand signal beyond launch, smaller operators stay stuck in the high-premium tier longer, which keeps their cost of capital higher, which makes it harder to win the next contract. The flywheel runs in the wrong direction.
The companies that have escaped this trap have done so by achieving the kind of cadence that makes them effectively self-insuring. SpaceX flies the Falcon 9 often enough that its Starlink deployment programme functions as a continuous reliability demonstration. The underwriters who cover Falcon 9 missions today are pricing against a heritage almost no other vehicle can match. Cadence buys credibility, and credibility buys cheap insurance, and cheap insurance lets you fly more. It is not magic. It is a flywheel that very few operators have managed to start spinning.
When the rocket fails
The recent ISRO PSLV mission carrying sixteen satellites is a useful reminder of what insurance is actually for. When a launch vehicle suffers an anomaly, the question of who pays when a rocket fails turns into a multi-party reconciliation that takes months, sometimes years, and runs through the same London market.
Most launch policies are structured around launch plus one year of in-orbit life, with separate cover for subsequent years. A failure during ascent triggers the launch policy. A failure during commissioning typically falls into a defined post-separation window. A failure six months in falls under the in-orbit policy, which may have been written by an entirely different syndicate. Each of these has different exclusions, different subrogation rights, and different requirements for the operator to substantiate the claim.
What outsiders rarely appreciate is that the loss adjustment process itself shapes the next generation of policies. Every claim teaches the market something. Every claim that took too long to settle, or revealed an ambiguity in policy wording, gets fed back into the next renewal cycle. The market is in continuous, cautious learning. A disputed clause from a 2019 failure shows up as tighter language in a 2025 policy; a successful subrogation against a manufacturer in one case changes how underwriters think about manufacturer liability in the next. The institutional memory of the London market is not written down in any single document. It lives in the people, in the precedents they carry, and in the quiet adjustments they make each year to the terms they are willing to offer.
The cultural problem
Working at ESA, I sat in plenty of meetings where engineers and finance people talked past each other. The engineer wanted to fly the new thing because the new thing was better. The finance person wanted to fly the proven thing because the proven thing was insurable at a sensible rate. Both were right within their own logic.
The space industry has a romance with novelty that the insurance market does not share, and probably should not share. An underwriter who got excited about new technology would be a bad underwriter. Their job is to be sceptical, to ask whether the test campaign was sufficient, whether the failure modes have been characterised, whether the operator has the resources to manage an anomaly without going bankrupt mid-recovery.
That scepticism is a feature, not a bug. It is also, in aggregate, a constraint on what the industry can attempt. The frontier of what is technically possible runs ahead of the frontier of what is insurable, and the gap between those two frontiers is where a lot of interesting ideas die quietly.
What changes the picture
A few things are slowly shifting the dynamic.
Parametric insurance products, which pay out on a defined trigger rather than after a lengthy loss adjustment, are starting to appear for some space risks. They are still niche but they reduce the friction that makes traditional cover so expensive for smaller operators.
Captive insurance vehicles, where a large operator effectively insures itself through a wholly owned subsidiary, are becoming more common as constellation operators reach a scale where self-insurance makes actuarial sense. This pulls some of the largest risks out of the conventional market, which in turn changes the risk pool for everyone else.
Government-backed insurance pools, common in aviation and shipping, have been proposed for space but have not yet materialised in any serious form. A state-backed reinsurance facility for emerging launch vehicles could in principle break the cold-start problem for new entrants, though the moral hazard problems are obvious and unresolved.
And cadence, simply flying more, remains the most reliable way to lower premiums. Every successful flight is a data point. Enough data points eventually persuade an underwriter that the risk is what the operator says it is. There is no shortcut, only repetition.
The committee where ideas die
I have spent enough time around astronauts and the systems that support them to know that the human story of space is mostly invisible to the public. The underwriting story is even more invisible, and arguably more consequential for what gets built. As I wrote last week about the Rosalind Franklin mission, projects can survive political hostility if the contracts are in place. They cannot survive insurance markets that refuse to price them.
The committee where ideas die is not in Washington or Brussels or Paris. It is in a room above a Lloyd’s box, where two or three people read an engineering report, look at each other, and decide they are not going to lead this one. The broker takes the file away. The operator gets a phone call. The press release never gets written.
That is the unglamorous empire. It runs on actuarial tables and institutional memory and a deep, well-earned suspicion of optimism. It is one of the most consequential filters in the entire space economy, and almost nobody who works in space wants to talk about it, because acknowledging it means acknowledging that the gating function on human ambition in space is, quite often, a quiet conversation between people who have never been to a launch site.
The engineers will keep designing. The founders will keep pitching. And somewhere in EC3, a small number of underwriters will keep deciding which of those efforts get to find out whether they would have worked. The rest will join the long, silent catalogue of missions that were technically sound, commercially viable, and killed by a premium nobody could afford — vetoed not by physics, not by politics, but by the considered judgement of a few people in a few buildings near Leadenhall Street, doing the unglamorous work of deciding what risk actually costs.
Photo by Richard Harris on Pexels