Safety Data

Waymo vs Tesla: the robotaxi safety record

Last updated October 12, 2026 · refreshed weekly · every number sourced

Neither Waymo nor Tesla's robotaxi service has caused a fatality. Beyond that headline, the records diverge sharply in scale and transparency: Waymo has logged 220M+ driverless miles with independently studied crash rates, while Tesla does not disclose robotaxi mileage, so its per-mile safety cannot be independently verified.
MetricWaymoTesla Robotaxi
Deaths caused00
Involved in fatal crashes3 (struck by human drivers, not at fault)0
Injury crashes (NHTSA reports)1172
All reported crashes1,42923
Crashes per million miles1.28n/a
Human-driver benchmark~1.9 crashes per million miles (≈6 million police-reported crashes a year ÷ ~3.2 trillion miles driven)

How to read these numbers

IIHS matched study: 1.28 police-reportable crashes per million miles vs 4.06 for human drivers in the same areas and years (~50M driverless miles, 2021–2024)

68% fewer police-reportable crashes and 81% fewer injury crashes than matched human drivers (IIHS, Jul 2026). Waymo's own safety dashboard, updated Sep 24, now covers 270M driverless miles through Jun 2026 and reports 82% fewer injury-causing crashes and 95% fewer serious-injury crashes than matched human benchmarks (~841 injury crashes avoided, by Waymo's own methodology). Our former “6.5” figure divided a Nov 2025 crash count by Mar 2026 mileage — mismatched cutoffs — and has been retired. Raw SGO totals count every contact, including not-at-fault, and can't be read as a matched rate.

Tesla: Not calculable from public data — Tesla claimed 1M cumulative unsupervised miles on Sep 3 but does not break out crashes by driving mode or publish audited mileage. 23 unique SGO incident reports from launch (Jun 2025) through mid-Aug 2026 filings, incl. a bollard/chain-fence strike in Austin. Most were other drivers' fault (rear-endings while stopped); at least 2 were caused by Tesla's remote operators. Mileage undisclosed, so no per-mile rate can be computed.

source: IIHS · Jul 23, 2026 ↗source: Waymo Safety Impact · through Jun 2026 ↗source: Austin American-Statesman via GovTech · Sep 16, 2026 ↗source: NHTSA Standing General Order (raw data) ↗

FAQ

Has a Waymo ever killed anyone?
No fatal crash has ever been attributed to Waymo's driving. Struck by human drivers in SF (Jan 2025) and Tempe (Sep 2025); in Dallas (Aug 7, 2026) an SUV threw a pedestrian into a Waymo's path — Texas' first fatal AV-involved collision. Not at fault in any of the three..
Has a Tesla robotaxi ever killed anyone?
No deaths in any Robotaxi-service crash since the June 2025 launch. Consumer Autopilot / FSD cars are a separate program and are not counted here.
Is Waymo safer than a human driver?
68% fewer police-reportable crashes and 81% fewer injury crashes than matched human drivers (IIHS, Jul 2026). Waymo's own safety dashboard, updated Sep 24, now covers 270M driverless miles through Jun 2026 and reports 82% fewer injury-causing crashes and 95% fewer serious-injury crashes than matched human benchmarks (~841 injury crashes avoided, by Waymo's own methodology). Our former “6.5” figure divided a Nov 2025 crash count by Mar 2026 mileage — mismatched cutoffs — and has been retired. Raw SGO totals count every contact, including not-at-fault, and can't be read as a matched rate.
Why is there no crash rate for Tesla?
Not calculable from public data — Tesla claimed 1M cumulative unsupervised miles on Sep 3 but does not break out crashes by driving mode or publish audited mileage. Historical context only: an early-Austin estimate (Feb 2026) put the rate around 9–18 per million miles — 14 crashes ÷ ~0.8M estimated miles, most driven with safety monitors aboard — roughly 4× a human benchmark. That estimate covers a different window, fleet, and driving mode than Waymo's figures and is not comparable to them. Robotaxis must also report every minor contact while humans only generate police reports.

Figures come from the sourced dataset behind the Waymo vs Tesla scoreboard, updated every Monday. Crash counts are NHTSA Standing General Order reports, which record involvement, not fault. This page is informational and is not investment advice.