Show summary Hide summary
- Large-scale driving data: what Waymo analyzed
- How Waymo’s crash rates compare to human drivers
- Safety for pedestrians, cyclists and motorcyclists
- Why the robotaxi advantage may be understated
- Real-world near-misses captured on video
- Context: traffic fatalities and daily risk
- Takeaways about robotaxi reporting and transparency
Waymo says its fleet of self-driving taxis is proving safer than human drivers, and it has released a large dataset to back that claim. After more than a quarter-billion miles without a human behind the wheel, the company is pointing to sharp reductions in serious crashes and incidents involving vulnerable road users.
Large-scale driving data: what Waymo analyzed
Waymo reviewed performance across its robotaxi operations in several U.S. cities. The data covers routine service through the end of June.
Jet-powered Harley rockets to 60 mph on pure thrust: twin diesel jets, no drivetrain
Expensive gasoline pushing drivers to electric vehicles: will you switch?
- Cities included: Atlanta, Austin, Phoenix, Los Angeles, and San Francisco.
- Total autonomous distance logged: 271 million driverless miles.
- Types of incidents tracked: collisions causing serious injury, minor injuries, and interactions with pedestrians, cyclists, and motorcyclists.
How Waymo’s crash rates compare to human drivers
Waymo says the robotaxis are far less likely to be in harmful collisions than the average motorist.
- 95% fewer crashes that result in serious injury or worse, compared with typical human-driven cars.
- 82% fewer collisions that cause minor injuries.
- Overall, the company reports its fleet avoided 841 injuries that human drivers would likely have suffered over the same distance.
Safety for pedestrians, cyclists and motorcyclists
The Waymo dataset highlights protection for people outside vehicles.
- 93% fewer crashes involving pedestrians.
- 86% fewer collisions with cyclists.
- 82% fewer incidents involving motorcyclists.
Why the robotaxi advantage may be understated
Independent reporting practices can make human crash totals appear higher or lower. Waymo argues the gap could be even wider than the numbers show.
- Many human-caused property-damage crashes go unreported. Estimates say about 60% are not filed with police.
- Roughly 32% of injury crashes involving human drivers may not be reported.
- Autonomous fleet operators, by comparison, must log and disclose even minor events.
Real-world near-misses captured on video
Waymo released footage showing close calls its software handled. The clips illustrate how sensors and software respond to sudden hazards.
- One clip shows a pedestrian stepping out from behind a parked car. The robotaxi stopped in time.
- Another scene shows a cyclist running a red light at night. Waymo’s vehicle avoided a collision.
These examples highlight detection and braking in complex, split-second scenarios.
Context: traffic fatalities and daily risk
Traffic deaths remain a major public safety issue. On average, about 100 people die each day on U.S. roads, a number policymakers aim to lower.
Waymo frames its results as evidence that autonomous driving technology can reduce that toll, particularly where robotics manage risk better than humans.
Takeaways about robotaxi reporting and transparency
Waymo’s disclosure aims to show measurable benefits from autonomous mobility at scale. The company pairs raw counts with context about reporting differences.
- Large fleets generate detailed telemetry for every event.
- Regulatory rules require firms to log incidents other operators might not disclose.
- Public video and datasets give researchers and regulators material to evaluate safety claims.
