Where AV Liability Gets Underwritten: AM Best Says Surplus Lines First as Admitted Carriers Exclude
A new AM Best market-segment report (published Sept 14, 2026, excerpted by Insurance Journal on Sept 23) is the clearest statement yet of where autonomous-vehicle liability actually lands while the personal-auto pool stays intact — and it is not the admitted carriers who write your car policy.
The exposure is a hybrid nobody's actuarial tables cover: "Driverless technology exposures exist at the intersection of product liability, automobile liability, and risks involved with faulty software because accidents in tech-driven vehicles could be caused by software errors, manufacturing defects, network outages, GPS flaws, failing sensors, or even cyberattacks. Actuarial data has been available for decades for human-driven vehicles, but underwriting and pricing for a fleet of vehicles that operate without drivers presents a very different set of risks" (Insurance Journal).
Why surplus lines first: "The commercial automobile segment has been a consistently unprofitable P/C line of coverage for an extended period, making it more understandable that the coverage for driverless technologies will, at least initially, be found in the surplus lines market12" (Insurance Journal). The regulatory patchwork compounds it: "At the present time, 42 states and the District of Columbia have enacted some sort of AV-related legislation. However, requirements by state vary widely." And cyber cover "would become more integral… for autonomous truck owners/operators than what is now the norm for traditional commercial trucking risks."
The severity math behind the shift (TheStreet): a Waymo Jaguar I-Pace carries roughly $100,000 in sensors, about $175,000 all-in by one outside estimate — and Root founder Alex Timm notes the repair problem: "If a sensor is damaged or certainly if any of the LIDAR detectors are damaged, it's almost impossible to actually repair." The Highway Loss Data Institute finds driver-assistance systems have cut claim frequency by 13% to 39% "while raising the cost of the claims that remain." Frequency down, severity up — the same trade insurers already live with, concentrated in a fleet asset.
What it means: the personal-auto premium pool — "$350 billion… intact for now," with robotaxis at "less than 1% of miles" per TheStreet — is not yet bleeding. The liability shift is arriving earlier as an underwriting-structure question: AV fleet risk migrates to surplus lines and bespoke fleet/product-liability programs (Allianz-Waymo in Europe, see Waymo Scale Update: First-Ever $5B Debt Raise, LA Crash Counts Triple With Fleet Size, 15 Markets and Counting), while admitted personal-lines carriers keep the shrinking-but-intact pool. Watch whether a US admitted carrier lands a Waymo-style mandate — that would be the first signal the pool's replacement economics are being built onshore.
-
An instance of Autonomous liability is born in surplus lines, not the admitted pool. — AM Best explicitly locates the new hybrid exposure outside the admitted market, where actuarial data and rate regulation cannot yet reach it. ↩︎
-
An instance of Commercial vehicle liability shifts from driver negligence to high-limit product liability and software diagnostics. — AM Best locates driverless risk in bespoke surplus-lines product-liability programs rather than traditional commercial auto — the negligence-to-product-liability migration playing out in the underwriting stack itself. ↩︎