Waymo’s growing driverless service is producing a larger public record of crashes, injuries and disputed responsibility. That record does not show a crash-free technology. It shows something more specific: after more than 50 million miles of driverless operations, Waymo vehicles appear to be involved in far fewer serious crashes than comparable human driving.
The distinction matters because Waymo will almost certainly appear in more crash reports as its service expands. Some of those crashes may involve serious injuries or death. The key question is not whether crashes happen, but how often they happen, how severe they are, and who appears to be responsible.
The fatal crash that put the issue in focus
The first ever fatal crash involving a fully driverless vehicle occurred in San Francisco on January 19. The vehicle belonged to Waymo, but the reported sequence does not place blame on Waymo.
A driverless Waymo with no driver or passengers was stopped at a red light. Another car was stopped behind it. According to Waymo, a human-driven SUV then struck the stopped vehicles at high speed, creating a six-car pileup. One person was killed, five others were injured, and a dog also died.
A separate major crash in October, also in San Francisco, followed a similar pattern in one important respect: the Waymo was stopped for a red light. According to Waymo, another vehicle crossed the double yellow line, hit an SUV that was stopped to the Waymo’s left, and pushed that SUV into the Waymo. One person was seriously injured.
Those two incidents produced the worst injuries among Waymo crashes in the last nine months covered by the source article. They also illustrate a recurring point in the reports: many serious Waymo-involved crashes begin with another road user making the dangerous move.
Why the denominator matters
Crash reports can look alarming when read one by one. A long list of incidents involving autonomous vehicles naturally raises concern, especially when injuries are mentioned. But safety comparisons need mileage as context.
Since 2020, Waymo has reported roughly 60 crashes serious enough to involve an injury or airbag deployment. Those crashes took place across more than 50 million miles of driverless operations. The source article compares that distance to roughly 70 lifetimes behind the wheel.
Federal regulations require Waymo to report significant crashes whether or not the Waymo vehicle was at fault. A report can be required even when the Waymo is not moving. That means the public record includes incidents where a Waymo was struck while stopped, hit from behind, or caught in a chain reaction caused by someone else.
This reporting approach is useful, but it also makes simple crash counts misleading. A database entry involving a Waymo does not automatically mean the autonomous system caused the crash.
What the recent crash reports show
The source article reviewed Waymo crashes from July 2024 through February 2025. During that period, Waymo reported 38 crashes serious enough to involve an alleged injury or an airbag deployment.
Based on that review, only one crash clearly appeared to be Waymo’s fault. Waymo may have been responsible for three more, but the information was not enough for a firm judgment. The remaining 34 crashes appeared to be mostly or entirely caused by others.
The 34 crashes included several recurring patterns:
- 16 crashes where another vehicle hit a stationary Waymo or caused a multi-car pileup involving one.
- 10 rear-end crashes, three side-swipe crashes, and three crashes involving a vehicle crossing the center line.
- Eight crashes where another car, or in one case a bicycle, rear-ended a moving Waymo.
- Five crashes where another vehicle moved into a Waymo’s right of way, including a car running a red light, a scooter running a red light, and a car running a stop sign.
- Three incidents during passenger drop-off, when a passenger opened a door into a passing car or bicycle.
Waymo has a “Safe Exit” program meant to warn passengers and reduce dooring incidents, but the reports show it does not eliminate that risk.
There were also two reported incidents where it is unclear that any collision occurred. In one, Waymo says its vehicle slowed and shifted slightly left within its lane while preparing to change lanes because of a stopped truck ahead. An SUV driver in the next lane reacted by steering left and hitting the opposite curb, while Waymo says its vehicle did not leave its lane or make contact. In another, a pedestrian passed in front of a stopped Waymo, then later approached the vehicle and may have made contact with it before claiming a minor injury.
Where Waymo appears responsible
The clearest Waymo-at-fault crash in the reviewed period happened in December in Los Angeles. A Waymo hit a plastic crate and pushed it into the path of a scooter in the next lane. The scooterist hit the crate and fell. Waymo did not know whether the rider was injured.
The three harder-to-judge crashes all involved another vehicle making an unprotected left turn across the Waymo’s lane of travel. In two cases, Waymo says its vehicle braked hard but could not avoid the crash. In the third, the other vehicle hit the Waymo from the side. The summaries made the other car sound responsible, but the source article does not treat those cases as certain.
Even under the more conservative assumption that all three uncertain crashes were Waymo’s fault, the majority of the 38 serious crashes reviewed would still not be Waymo’s fault.
How Waymo compares with human drivers
Waymo has also published broader comparisons between its driverless vehicles and human drivers. Its most recent safety data hub update covered crashes through the end of 2024.
The company’s analysis focused on 44 million miles driven in Phoenix and San Francisco through December, leaving aside smaller operations in Los Angeles and Austin. Waymo estimated that human drivers on the same roads would have had 78 crashes serious enough to trigger an airbag. Waymo’s driverless vehicles had 13 airbag crashes, which Waymo described as an 83 percent reduction compared with typical human drivers.
For injury-causing crashes, Waymo estimated that human drivers would have had 190 over the same 44 million miles. Waymo reported 36 injury-causing crashes across San Francisco or Phoenix, an 81 percent reduction. That was an improvement from last September, when Waymo estimated a 73 percent reduction over its first 21 million driverless miles.
The source article also notes a December study co-authored by Waymo and Swiss Re that looked at successful insurance claims against Waymo. That lens is important because third parties decide whether a claim is filed, and claims adjusters decide whether Waymo is held responsible. However, the source also notes that insurance claims take months to appear, so the December report covered crashes only through July 2024.
The overall picture is not that driverless cars have removed road risk. It is that, in the data described here, Waymo’s driverless vehicles are involved in fewer serious crashes per mile than human drivers, and many of the worst Waymo-involved crashes appear to start with human road users breaking traffic rules or losing control.