From Singapore: Waymo's robotaxis have now travelled over 170 million miles on public roads without a human driver behind the wheel. With 92% fewer bodily injury claims and 88% fewer property damage claims over 25 million miles, Swiss Re (one of the world's leading reinsurers) concluded that Waymo is significantly safer than human-driven vehicles.
The company updated its online safety hub this month with data spanning commercial robotaxi services available to the public in Phoenix, the San Francisco Bay Area, Los Angeles, Atlanta, Austin, and Miami. The figures represent a watershed moment for autonomous vehicle technology: at scale, across genuine urban environments, with paying passengers aboard.
Compared to an average human driver covering the same distance in Waymo's operating cities, the Waymo Driver achieved: 90% fewer serious injury or worse crashes (27 fewer crashes) 82% fewer airbag deployment crashes in any vehicle (173 fewer crashes) 81% fewer injury-causing crashes (411 fewer crashes).
These numbers matter because they represent something the autonomous vehicle industry has struggled to demonstrate: quantifiable safety gains under real-world conditions. Yet the data also reveals genuine tensions within how we evaluate autonomous systems.
The Methodology Question
Safety advocates have correctly noted that Waymo publishes its own research about its own vehicles. Not all human crashes are reported. NHTSA estimates that 60% of property damage crashes and 32% of injury crashes aren't reported to police. In contrast, AV companies report even the most minor crashes in order to demonstrate the trustworthiness of autonomous driving on public roads. This creates an asymmetry: Waymo is measured against an incomplete human baseline.
Yet independent researchers have examined the methodology closely. David Zuby, chief research officer of The Insurance Institute for Highway Safety (IIHS) after reviewing the papers, said: "These reports represent a good-faith effort by Waymo to evaluate how the safety of its autonomous driving system compares with the safety of human driving. The results are encouraging and represent one step in our evolving understanding of autonomous driving safety."
The peer-review process itself has strengthened the analysis. Over 3.8 million miles, the Waymo Driver reduced the frequency of property damage insurance claims by 76% and completely eliminated bodily injury claims compared to human drivers. This comparison used insurance claims data rather than Waymo-generated crash reports, providing an external validation point.
What the Crashes Reveal
Waymo's reported incidents tell a story worth examining. Of serious crashes analysed in detail, of the 41 crashes that involved driving mistakes, 37 seemed to be mostly or completely the fault of other drivers. 24 crashes happened when another vehicle collided with a stationary Waymo, including 19 rear-end collisions.
Some failures revealed mechanical rather than software issues. A Waymo crashed after one of its wheels literally fell off. Others exposed design gaps: Of the 45 most serious crashes, three were dooring incidents, but they accounted for two of the seven crashes that caused significant injuries. The chime the car plays, however, may not always be loud enough.
The Adoption Puzzle
Despite these safety figures, public acceptance remains fragile. After years of being known as overly polite, Waymo robotaxis are now reportedly acting more "confidently assertive" in city environments like San Francisco. The Wall Street Journal reported on a series of incidents, including cars performing simultaneous lane changes in a tunnel and an illegal U-turn in San Bruno, California, that led to a police stop. This shift toward more human-like driving efficiency creates a perception problem even as safety metrics improve.
The international expansion adds complexity. Expansion plans include 20 more cities in 2026, including Tokyo and London. Each market brings different regulatory frameworks, traffic cultures, and public expectations about what autonomous vehicles should prioritise. Safety data alone will not determine success.
For Australian businesses watching this space, the implications are significant. Waymo's success in proving autonomous vehicles can be safer than human drivers removes a key regulatory barrier. But it also reveals that safety alone is not enough. Trust, transparency about limitations, and responsiveness to emerging issues around passenger safety features like door alerts matter equally. The next phase of autonomous vehicle adoption will likely depend less on engineers and more on whether companies can maintain the social licence to operate in congested urban environments where most people live.