Based on the latest NHTSA report, Tesla’s ‘Robotaxis’ keep crashing in Austin, Texas, despite ‘safety monitors’ preventing an unknown number of crashes.
Under an NHTSA Standing General Order SGO, automakers are required to report crashes involving their autonomous driving (ADS) and advanced driver assistance systems (ADAS) within five days of being notified of them.
For years, Tesla was only reporting ADAS crashes, since, despite the names of its Autopilot and Full Self-Driving systems, they are only considered level 2 driver assistance systems.
Since the launch of the Robotaxi service in Austin, Texas, where Tesla moved the supervisor from the driver’s seat to the passenger seat, it has now reported its first few crashes under the ADS reporting.
In the first month of operation in July, Tesla reported three crashes with its ‘Robotaxi’ service in Austin.
This week, NHTSA has updated its crash report and revealed a 4th crash that happened in September:
| Report ID | Incident Date | Incident Time (24:00) | Make | Model | Model Year | Automation System Engaged? | Highest Injury Severity Alleged | Crash With | Roadway Type | Weather |
| 13781-11687 | SEP-2025 | 01:25 | TESLA | Model Y | 2026 | ADS | Property Damage. No Injured Reported | Other Fixed Object | Parking Lot | Partly Cloudy |
As we previously highlighted, when it comes to both ADS and ADAS crash reporting, Tesla abuses the redacting capacity and hides most information about its crashes, unlike most of its competitors.
Therefore, we don’t have much information about this new crash, but it reportedly occurred in a parking lot and involved a Tesla Robotaxi crashing into a “fixed object,” resulting in property damage.
What’s most interesting about this crash is that it comes as Tesla released the first bit of data about its Robotaxi program in Austin.
During its earnings call last week, Tesla confirmed that the Robotaxi fleet has traveled 250,000 miles since its launch in late June.
Therefore, Tesla Robotaxi currently crashes at a rate of about once every 62,500 miles. That’s with a safety monitor with a finger on a kill switch, ready to stop the vehicle at all times.
We have no data on how often Tesla’s safety monitors prevent crashes in its robotaxis.
For comparison, the NHTSA report lists 1,267 crashes involving Waymo vehicles. However, Waymo’s robotaxis have covered over 125 million fully driverless miles since inception. That’s a crash every 98,600 miles and without any onboard safety monitor.
Electrek’s Take
That’s the problem with comparing Tesla and Waymo.
At least we can now clearly see that Waymo’s incident rate is much lower than Tesla’s, but that’s with a safety monitor in Tesla robotaxis that prevents an untold number of crashes.
The actual difference could be 10x higher. We simply don’t know. Tesla has consistently refused to share any data regarding disengagement or intervention rates.
Top comment by taverngeek
Tesla is also using some number of remote drivers in Austin. I'd guess more for tricky situations such as exactly where to pick up or drop off riders than trying to react quickly to avoid crashes. I think the best indicator of FSD's true capacity is that Tesla has yet to apply for the California permit to begin driverless testing. That would require disclosing critical disengagements which Tesla is clearly unwilling to do at this time.
It’s also fair to note that Waymo doesn’t redact information about crashes as much as Tesla does. Therefore, we know more about the circumstances around the crashes, and many of them, in fact, the majority, point to the Waymo AI driver not being responsible. Over 125 million miles, crashes are inevitable, and we can see many examples of other drivers rear-ending and crashing into Waymo vehicles.
It could also be the case with Tesla’s robotaxis, but we don’t know as Tesla redacts the data.
One thing is clear: Tesla is way behind Waymo in autonomous driving safety.
FTC: We use income earning auto affiliate links. More.
Comments