Tesla says its robotaxis have covered 3,112,541 driverless miles with zero injuries, and it published those figures in a new Tesla robotaxi safety report on Wednesday. The page went live hours before the company’s third-quarter earnings call, and it tracks driverless miles across Texas and Florida.
It is the first company-published safety record for a robotaxi fleet that is expanding into new states, so the numbers deserve a close read. They are also Tesla’s own figures, and they have not been independently audited. Driveteslacanada and Teslanorth both covered the report.
The timing adds pressure. NHTSA is auditing the Cybercab, and Tesla has reported its first Cybercab crash to federal safety regulators.
What the Tesla robotaxi safety report claims
The report counts 3,112,541 autonomous miles between January 22 and October 18, 2026. Austin accounts for more than 2.1 million of them and Dallas for about 608,000. Tesla also breaks out harder conditions, including 862,000 after-dark miles, at least 113,000 rain miles and 575,000 rush-hour miles.
Tesla lists zero people injured, zero robotaxi airbag deployments and zero pedestrians or cyclists hit over that period. The service began in Austin in June 2025 with safety monitors on board, according to Driveteslacanada. The current count excludes miles with a monitor inside the car, so it reflects the no-monitor service rather than those early supervised rides.
Tesla frames the results with a line of its own: “The Safety Results stand on their own, but Texas police reports offer useful context.” That context is the comparison Tesla built next, and it is where the numbers get more complicated.
How the robotaxi record compares with human drivers
Tesla benchmarked its record against an estimate of human driving. Using Texas police crash data, the company says human drivers would have been involved in about 9.1 injury collisions, 4.1 airbag-deployment collisions and 0.52 pedestrian or cyclist collisions over 2.95 million miles in Austin, Dallas and Houston.
| Incident type | Human drivers (Tesla estimate) | Robotaxi (Tesla report) |
|---|---|---|
| Injury collisions | 9.1 | 0 |
| Airbag-deployment collisions | 4.1 | 0 |
| Pedestrian or cyclist collisions | 0.52 | 0 |
The gap is widest for injury collisions. The pedestrian figure is harder to read. An expected 0.52 events is a small number, so zero observed says little on its own, and a fleet can record zero over a few million miles even when its underlying rate is not zero.
The two sides also do not count the same way. Tesla’s injury metric counts only collisions where a robotaxi made first contact, while the human benchmark includes collisions regardless of which vehicle made contact. The benchmark covers only Tesla’s Texas markets, even though the report also tracks miles in Florida. The benchmark uses 2.95 million miles while the robotaxi total is 3.1 million, so the two sides do not share the same base.
What the report leaves out
Tesla’s own caveats matter as much as the headline numbers. The figures exclude miles driven with onboard safety monitors. They include trips where a remote operator was assisting or had taken over, so a count with no one inside the car is not the same as a measure of fully autonomous driving.
The mileage itself depends on the source. Tesla’s official count is 3.1 million miles through October 18, while a live counter showed about 3.3 million at the time of writing, according to NotATeslaApp and TAO Media. Neither figure has been independently verified, and no outside audit of the zero-injury claim has been published.
The Cybercab crash the safety report does not count
The new report also leaves out the Cybercab. According to Teslanorth, the crash happened in August, when the Cybercab hit a water-filled pothole at about 5 mph. KRLD reports that it involved property damage only, with no injuries. Teslanorth also notes that the crash falls outside the new report, which counts only rides with nobody from Tesla inside the car.
The Cybercab began public rides in Austin on September 3, and it has no steering wheel or pedals. That design is why regulators are watching the crash closely. NHTSA had opened an audit query into Tesla’s self-certification that the Cybercab complies with federal standards before the crash report became public, according to Gadget Review.
NHTSA’s press release and audit document state that the Cybercab complies with applicable standards, but the audit is still open. It is unclear whether its findings will change the timeline. Yahoo covers the audit in detail.
The sensor setup is part of the debate. Tesla’s robotaxis rely on cameras alone, while competitors often pair cameras with lidar. Teslanorth reports the robotaxi lacks lidar and calls the pothole crash a possible edge case for camera-based perception, though no source has established that the sensors caused the crash.
Broader crash data adds context. According to Yahoo’s coverage of the first Cybercab crash, 24 robotaxi crashes in Texas were reported to federal regulators in the four weeks ending September 15, nine of them in Austin. Tesla’s Austin total since June 2025 stood at 26.
Which cities come next, and what remains unannounced
Tesla now operates in six cities: Austin, Dallas, Houston, Miami, Orlando and Tampa. The report lists Las Vegas, Jacksonville, Phoenix and San Antonio as coming soon, and it gives no launch dates for any of them.
Some coverage has run ahead of the facts. A Tesla Oracle report on October 23 described the four cities as a launch that would bring the total to 10 U.S. cities. Tesla has named those cities, but it has not launched service in them.
On the earnings call, Tesla said it is launching unsupervised Cybercab rides in Dallas next month, according to two blogs that reported the remarks: NotATeslaApp and Basenor. USA 24 has not located the primary transcript, so treat the timing as company-stated but unconfirmed. Our coverage of the Cybercab fleet in Austin covers the Dallas plans in more detail.
Rivals are moving too. Reuters reported on October 23 that Baidu is testing a robotaxi on London streets, according to US News.
How to check robotaxi safety claims yourself
- Confirm the city before you plan a ride. Open Tesla’s Robotaxi app and check whether your city shows as live, because a city named in a report is not yet a service area.
- Read the denominator. Ask whether a safety figure includes rides with a monitor, remote assistance or a remote takeover, and which benchmark it uses.
- Check the crash filings. Tesla’s filings appear in NHTSA’s Standing General Order reports, so compare crash counts across periods rather than relying on one company summary.
- Note the date on every mileage figure. The 3.1 million figure is an official count through October 18, while the 3.3 million figure came from a live counter at the time of writing.
What to watch next
The next checkpoint is Dallas. If Tesla starts unsupervised Cybercab rides there next month, as its earnings-call remarks suggested, the city will add another set of miles and crash reports to the record. Tesla is also putting its Optimus robot into production at Fremont, a separate bet that investors will weigh alongside the robotaxi timeline. Our Optimus coverage has the details.
Regulators have three open questions: the outcome of NHTSA’s audit of the Cybercab, whether the agency will act on robotaxi crash data, and whether Tesla will break out its figures more clearly or add independent verification. The first answer is likely to come from the audit itself.
The next report will show whether Tesla’s safety record holds up under outside scrutiny. Until the data is verified and the four named cities get dates, the zero-injury figure remains a company claim with a useful but limited benchmark.
