On September 24, 2026, Waymo published its latest safety-impact update, covering more than 270 million miles of fully autonomous operation through the end of June 2026 across five metropolitan areas: Atlanta, Austin, Los Angeles, Phoenix and San Francisco. Compared with a human-driver benchmark for the same roads, Waymo says its vehicles were involved in 82% fewer injury-causing crashes, which it translates into 841 injury crashes avoided under a conservative assumption of at least one injured person per crash.
The company also reports 95% fewer crashes with serious injury or worse - about 20 times fewer - and reductions in injury crashes involving vulnerable road users of 93% for pedestrians, 86% for cyclists and 82% for motorcyclists. To give the mileage a scale, Waymo likens it to driving the length of Route 66 more than 110,000 times, or more than 300 human lifetimes of driving.
Why it matters: this is the largest real-world record any driverless operator publishes, and it is the dataset regulators, insurers and city governments cite when deciding whether to let robotaxis expand. The pattern - large relative reductions holding up as mileage grows into the hundreds of millions - is what Waymo leans on as it launches in new cities.
What it does not show: the figures are Waymo’s own analysis against a human baseline it constructs, the data stops at June 2026 and covers only the five most mature markets, and the “841 avoided” figure is an estimate rather than a count of events. It says nothing about how other operators’ vehicles perform.