INAUGURAL EDITION

IN BOTS WE TRUST 2026

What five years of underwriting Physical AI says about the promise and risk of robots. Hundreds of thousands of data points from 392 robotics companies — the first report on robot risk built from money that actually changed hands.
$572M

of liability capacity deployed for Physical AI in 2025, seven times 2022

Top 1%

loss ratio significantly better than best-in-class among the 25 largest U.S. P&C groups

−69%

fall in the price of robot liability per dollar of revenue, from 2022 to 2025

Eight things nobody could measure in robotics before

Surveys capture what people think about robot risk. This report shows what operators, underwriters, and claims data reveal about how that risk plays out in the real world.

Top 1%

A loss ratio that beats the best of the largest U.S. insurers

In every policy year the program's loss ratio has been significantly better than the best-in-class result among the 25 largest U.S. P&C groups (48–52% on all lines), against an industry average of 57–67%. The lines robots touch run worse still: commercial auto liability at 71%.

Capacity grew seven-fold in three years

Aggregate liability limits deployed for robotics rose from $80M in 2022 to $572M in 2025, up 21% in the last year alone. Excess layers were the largest source of new capacity.

−69%

Robot liability got two-thirds cheaper

Premium per dollar of revenue in 2025 stood at 31% of its 2022 level. Price has stopped being the barrier to insuring robots; terms are where the friction moved.

6 → 700

The price ladder runs from software to dirt

With the autonomous mobile robot at 100, robotics software prices at 6, drones at 50, arms at 172, delivery at 191, and robots that clean, dig and move things in occupied spaces at 360–700.

34%

Cyber fraud leads the loss dollars

Email compromise and social engineering incidents account for a third of all loss dollars. Incidents in which a member of the public was hurt account for 12%. Only 4% of companies have ever reported an incident.

42% → 29%

Underwriters moved from excluding to pricing

Half of robot risks still carry a restriction, but exclusions fell from 42% to 29% of decisions, price adjustments rose from 11% to 36%, and outright declinations fell from 11% to 1%.

25–65%

Safety controls already earn lower rates

87% of applicants report a safe-stop system and 65% a safety framework. Companies reporting a control pay 25–65% less per dollar of revenue, yet only 6% publish a safety page. The terms discount is still on the table.

2019

The average robotics company was founded in 2019

With 21 employees and 25 robots in the field; 40% are pre-revenue. One in three is headquartered in California.

5

Predictions for 2027

Rates find a floor, excess and cyber grow with the fleets, warranties replace exclusions, and safety standards and compliance reports become underwriting inputs.

A look inside

Five years of robot underwriting, distilled into data-rich figures and tables. The full story is in the report.
The price ladder, segment by segment

23 robot segments priced against each other, from software to earth-moving

Unlock the report
Where robot insurance buyers are

32 states mapped, plus the sector, unit type and use case breakdowns

Unlock the report

Reviews

Samuel Reeves
Founder & CEO at FORT Robotics
"The report's data matches what we see in the field: robot safety is about much more than whether the hardware works. The focus has to extend to the entire system around the machine. Verifiable controls, secure communications, and effective human oversight turn unknown risks into manageable ones. The companies that evidence trusted, cross-industry safety best practices will capture the largest gains."
Grayson Brulte
Founder at The Road to Autonomy
"Physical AI does not scale in labs. It scales when the technology is commercialized and the risk is priced. Koop’s data shows robot liability is becoming cheaper, more capacity-rich, and underwritten at scale. That is the missing layer of the Autonomy Economy."

Written for the people who decide whether robots get deployed

Founders and operators
  • What your robot liability should cost relative to the benchmark, by segment
  • The six documents that pre-empt nine in ten underwriting conditions
  • Which controls already move price, and how to make terms follow
Investors and enterprise buyers
  • Diligence benchmarks: what the median vendor reports, and what the top decile can show
  • The price ladder as a market view of risk by segment
  • Why deployments arrive later than plans say, and what that does to exposure
Underwriters, brokers and policymakers
  • The hazards underwriters name and the levers they actually use
  • What goes wrong: root causes and lines of business, in shares
  • The standards map: ISO 10218:2025, R15.08, ISO 13482, UL 4600, the EU Machinery Regulation and more