Waymo's robotaxis had 95% fewer serious crashes than humans. Now it plans a fleet of 20,000 cars

Waymo’s robotaxis had 95% fewer serious crashes than human drivers over 271 million miles, according to its own data. The independent study by the Insurance Institute for Highway Safety (IIHS) found 81% fewer injury crashes. Waymo plans to expand its fleet from 4,000 cars to 20,000 by 2027, citing its own data. The data shows that Waymo’s rates are significantly lower than the human benchmark, with 55 fewer serious crashes in rider-only miles and 81% fewer injury crashes. The gap varies by city, with San Francisco having a larger gap than Phoenix and Los Angeles.

Waymo’s data is based on police-reported crashes in the counties where it operates, and it counts every crash, regardless of who was at fault. It reports 55 fewer serious crashes in rider-only miles and 81% fewer injury crashes compared to the human benchmark. The human benchmark has 0.21 serious-injury crashes per million miles, while Waymo’s rate is 0.01. The data also shows 81% fewer injury crashes and 85% fewer single-vehicle crashes for Waymo compared to the human benchmark. The gap changes with the city, with San Francisco having a larger gap than Phoenix and Los Angeles.

Waymo’s data is modeled against what human drivers would be expected to have in those areas, and the 55 fewer serious crashes are measured against it. The IIHS data stops at 2024 and covers about 50 million driverless miles in four cities. The data shows that Waymo’s rates are significantly lower than the human benchmark, with 81% fewer injury crashes and 85% fewer single-vehicle crashes. The IIHS also warned that the federal system is not ready for what comes next, as mileage reporting is voluntary and only Waymo releases its driverless miles. The data shows that Waymo’s rates are significantly lower than the human benchmark, with 81% fewer injury crashes and 85% fewer single-vehicle crashes.

AI-generated summary of the original article — see the source for the full story.