Waymo put 270 million driverless miles behind an 82 percent injury-crash reduction

Waymo published an update to its safety analysis on September 24, covering more than 270 million miles of fully autonomous operation through the end of June 2026 across Atlanta, Austin, Los Angeles, Phoenix and San Francisco.

Measured against human benchmarks and regardless of fault, the company said its driving system was involved in 841 fewer injury-causing crashes, an 82 percent reduction. It put serious-injury-or-worse crashes at 95 percent fewer, and reported reductions of 93 percent in injury-causing crashes involving pedestrians, 86 percent for cyclists and 82 percent for motorcyclists.

The post’s opening claim, that at least 841 people avoided injuries they would otherwise have sustained, carries a footnote. It is based on the 841 estimated crashes prevented, “assuming a conservative baseline of at least one injured person per crash”. The prevented-injury count is a modelled figure resting on a stated assumption rather than a count of people.

Two further limits sit on the page rather than behind it. The data runs to the end of June, roughly three months behind publication. And both the crash counts and the human benchmark they are measured against are the company’s own, drawn from its own service areas.

The post also pointed at the assurance side of the work. It said the driving system undergoes a formal safety review before deployment on public roads, and linked to a paper setting out the acceptance criteria used to judge whether it is ready.