SES and Eutelsat scrapped roughly $900 million of satellites ordered from the same vendor in 2022, and the LEO replacement the industry is pivoting toward is not yet priced at scale.
The satellite insurance market is not recovering from a bad year. It is being forced to retire a business model.
In the last two months, SES disclosed it no longer needs two geostationary satellites, IS-41 and IS-44, that Intelsat originally ordered from Thales Alenia Space in 2022, and Eutelsat cancelled a third, Flexsat Americas, a month earlier. All three were slated for orbit within two years. Their combined build value runs around $900 million, per SpaceNews, based on the industry rule of thumb of roughly $300 million per GEO spacecraft on the launchpad.
The cancellations are not a cyclical wobble. They are the moment the market admits that the orbit class underwriting was built around is no longer carrying its weight.
The old GEO model was, on paper, the most elegant product in the industry. A single $300 million-plus satellite generated roughly $500 million in annual space-insurance premiums across a decade of launches, according to the same SpaceNews report on broker figures from Aon, whose U.K. space practice is led by Matthew Gleeson. The math rewarded patience: a small number of large, infrequent claims against a deep premium pool. When the 2023 loss year hit, with claims running roughly three times the annual premium pool, the model did not bend. It broke. Underwriters withdrew capacity, rates spiked, and a multi-year rebuild began.
That rebuild delivered two respectable years. 2024 produced about $390 million in underwriting profit on fewer claims. 2025 brought in roughly $650 million in premium on a GEO launch uptick, only to be offset by a $400 million claim in December, according to the same reporting on Aon's market read.
What the 2024-2025 data does not show is a return to the old shape. The 2026 launch and insurance calendar is set to run below 2025, and the new business that matters, the 2027-2028 insured-launch pipeline, looks fundamentally different from the GEO backlog that preceded it.
Eutelsat's Flexsat Americas cancellation is the cleanest signal of why. Eutelsat killed the satellite to free cash for the OneWeb low-Earth-orbit constellation, a strategic bet that the high-throughput growth story is moving down, not out. The industry is not adding GEO capacity; it is replacing it with smaller, cheaper spacecraft in LEO and with a new generation of smaller GEOs closer to dishwasher-sized than bus-sized. Viasat, the long-standing large-GEO champion, is signaling the same pivot after a ViaSat-3 contributed to the 2023 claim tally, as SpaceNews reported.
The problem for underwriters is that the new fleet is not a scaled-up version of the old one. It is a different product.
Relm, the specialty space-insurance carrier whose space practice is led by Andrew Bonwick, estimates that excluding Starlink, only about 5 to 7 percent of LEO satellites carry any insurance, compared with roughly half of GEOs, per the same SpaceNews report. The reason is mechanical. A traditional per-satellite policy does not survive contact with a proliferated LEO constellation of hundreds or thousands of spacecraft. Coverage is shifting, slowly, toward constellation-level and revenue-stream risk, where the underwriting question is no longer whether a specific satellite failed, but whether the operator's cash flow survived a batch anomaly.
That shift is also where the new money lives. Smaller-satellite coverage matters because it is the prerequisite for institutional capital that still treats space as exotic, according to Yuk Chan, founder and CEO of Charter Space, a California fintech mission-management and insurance-brokerage venture that launched a brokerage earlier in 2026 to streamline underwriting for smaller operators, as the same SpaceNews piece reports. Chan frames the brokerage as infrastructure for an underwriting market that does not yet exist at the scale of the LEO buildout.
Gleeson's 2027-2028 forecast, that insured launches are set to pick up significantly as the new constellations come online, is a conditional signal, not a guarantee. It assumes that by the time the launch cadence arrives, brokers and carriers have built the products to price mixed-orbit fleets as a portfolio rather than as a string of single-sat risks.
That is the gap the GEO cancellations expose. The old underwriting business was a single-orbit, single-customer-type annuity. The replacement is a portfolio business: LEO constellations, mixed-orbit operators, and small GEOs priced against revenue, regulatory, and supply-chain exposures that no underwriter has had to underwrite at scale. The market has roughly 18 to 24 months, the width of the 2027-2028 launch window, to build that framework before the volume shows up unpriced.
If the framework is not in place, the 2027-2028 launch wave will arrive into a market that has not learned how to price it. Smaller operators will not get the coverage that would let mainstream institutional capital underwrite their business. The growth the LEO buildout was supposed to deliver will run into a capital problem before it runs into a launch problem.
That is the watch item for 2027: not whether the rockets fly on schedule, but whether the underwriting model that was supposed to finance them is ready when they do.