Waymo Reports 82% Fewer Injury Crashes Across 270 Million Autonomous Miles

A self-driving car completing over a quarter of a billion miles without a human at the wheel used to sound like science fiction. Now Waymo, the autonomous-driving company owned by Google’s parent Alphabet, is publishing the numbers to show what that actually looks like in safety terms — and the figures are striking by any measure.

Waymo published its updated safety data on 24 September 2026, covering more than 270 million fully autonomous miles driven through the end of June 2026. These are rider-only journeys: no human driver, no safety operator ready to take the wheel. The analysis spans five US metropolitan service areas — Atlanta, Austin, Los Angeles, Phoenix and San Francisco.

What the Numbers Actually Say

The headline figure is an 82% reduction in injury-causing crashes compared with Waymo’s human-driver benchmark across those same five cities. The company also reports a 95% reduction in serious-injury-or-worse crashes — described in the data as roughly 20 times fewer than would be expected of human drivers covering the same distance.

Translated into a single number, Waymo says its vehicles were involved in 841 fewer injury-causing crashes than would have been expected had human drivers made those same journeys.

That last figure deserves a moment’s attention. The 841 is a modelled estimate based on a statistical comparison with human-driver crash rates, not a separately verified count of 841 named individuals who walked away unharmed because a robot was driving. The methodology is Waymo’s own, and the results have not been independently verified by a road-safety regulator or peer-reviewed research body. That doesn’t make the data meaningless — but it does mean it should be read as a company-reported benchmark comparison rather than a definitive public-health finding.

How Waymo Measures Its Safety Performance

The comparison works like this: Waymo takes its crash data from fully autonomous operations and sets it against what safety researchers would expect from human drivers travelling equivalent distances in the same cities. The measure covers crashes regardless of fault, so it includes incidents where another driver or road user was responsible.

Waymo’s chief external affairs officer Tekedra Mawakana has previously described the company’s mission in straightforward terms, saying the goal is “to be the world’s most trusted driver.” The safety data is central to that argument.

What the figures don’t cover is equally worth knowing. Aggregate crash rates across selected urban US service areas won’t necessarily tell you how an autonomous vehicle performs on rural roads, in heavy rain, or in the dense stop-start traffic of a market town on a Saturday morning. Pedestrians, cyclists and motorcyclists — who interact with autonomous vehicles differently from other car drivers — may require separate safety assessment. Performance in one set of cities doesn’t automatically transfer to every road type or operating condition.

The UK Regulatory Picture

Britain isn’t standing still on this. The Automated Vehicles Act 2024 creates a legal framework for deploying self-driving vehicles on UK roads, setting out safety responsibilities for what the legislation calls “authorised self-driving entities.” It’s a serious piece of legislation, built on the assumption that autonomous vehicles will eventually operate here at scale.

But Waymo’s US figures don’t constitute evidence of safety performance on UK roads. The driving environment is genuinely different — road design, weather, signage conventions, and the behaviour of other road users all vary considerably from San Francisco or Phoenix. Any company seeking to operate autonomous vehicles in Britain would need to satisfy UK regulators on UK terms.

Waymo does not currently operate a public passenger service in the United Kingdom. No UK government body has independently verified the headline percentages published in the company’s September 2026 safety data.

What Comes Next

The autonomous-vehicle industry knows that public trust is built slowly and lost quickly. Publishing detailed safety data — even company-reported data — is part of how firms like Waymo try to build that trust ahead of wider deployment. The question of independent verification will become more pressing as autonomous vehicles move beyond a handful of US cities and into new markets.

Regulators, road-safety researchers and transport authorities in multiple countries are watching these figures closely. So are insurers, who will eventually need to price the risk of vehicles that have no human driver to blame when something goes wrong.

And the technology is moving fast. Waymo has gone from early trials to more than 270 million fully autonomous miles in a relatively short period. Whether that trajectory continues, and where it leads geographically, is uncertain.

What This Means for Kent Residents

Waymo’s data covers US cities only, and there’s no autonomous ride-hailing service operating in Kent or anywhere else in the UK right now. For residents here, the more immediate relevance is the Automated Vehicles Act 2024, which sets the rules under which any future self-driving service — from any company — would need to operate on British roads before Kent passengers could ever climb into a driverless car. If and when trials do come to the UK, safety claims based on American roads would need to be assessed against independently verified performance in UK conditions — different roads, different weather, and very different traffic — before local authorities such as Kent County Council or Medway Council could reasonably consider any deployment.