Tesla’s 29 Austin Robotaxis have been involved in eight crashes since they launched in June, Electrek reports, citing National Highway Traffic Safety Administration data you can download here. Those crashes for the most part involved property damage, and only in one case led to a minor injury. Notably, the crashes occurred with a safety monitor in the front seat
As Electrek notes, that data suggests Tesla Robotaxis are crashing once every 40,000 miles, whereas the average human driver in the US crashes about once every 500,000 miles. On Tesla’s Full Self-Driving page, the company claims vehicles with the technology engaged have 7x fewer major and minor collisions — a claim that experts like Carnegie Mellon’s Phil Koopman have said doesn’t stand up to scrutiny.
Over the weekend, Tesla began testing two more Robotaxis without those safety monitors, in hopes of achieving CEO Elon Musk’s latest goal of removing them from the Austin fleet by year’s end.
Ginger sent that link to me, knowing that I would be interested. And I am, though what has really grabbed my attention is the claim that “the average human driver in the US crashes about once every 500,000 miles”. That just seems too infrequent to me. Like, how many total miles does one drive in a lifetime and how many crashes does the average driver experience?
First, this more recent Electrek article provides more context about the Tesla situation.
As we have been tracking in our previous coverage of the Robotaxi pilot in Austin, Tesla is required to report crashes involving its automated driving systems (ADS) to the NHTSA under a Standing General Order.
For months, we’ve seen these reports trickle in from Tesla’s small pilot fleet in Texas. In November, we reported that the fleet had reached 7 total crashes as of September.
Now, a new report filed by Tesla reveals an 8th crash occurred in October 2025.
According to the filing, the incident took place on October [Day Redacted], 2025, in Austin. The valid report (Report ID: 13781-11986) lists the “Highest Injury Severity Alleged” as “No Injured Reported,” but details are scarce because, as is typical for Tesla, the narrative description of the crash has been redacted to hide proprietary information.
We have been highlighting how Tesla often abuses NHTSA’s capability to redact much of the information in the crash reports, especially the ‘Narrative’ section, which explains precisely what happened in the incident.
It’s possible that Tesla’s Robotaxis are not responsible for some of these crashes, but we wouldn’t know because Tesla redacts most information.
In this new filing for the accident that happened in October, Tesla went even further as it even refrains from answering some of the sections. Instead, it says “see the narrative,” which again is redacted.
[…]
Last month, Tesla confirmed the fleet had traveled roughly 250,000 miles. With 7 reported crashes at the time, Tesla’s Robotaxi was crashing roughly once every 40,000 miles (extrapolating from the previously disclosed Robotaxi mileage).
For comparison, the average human driver in the US crashes about once every 500,000 miles.
This means Tesla’s “autonomous” vehicle, which is supposed to be the future of safety, is crashing 10x more often than a human driver.
While Tesla’s Robotaxi fleet reportedly increased in November, with the number of cars spotted going up to 29, there’s no evidence that the Robotaxi mileage increased. In fact, the utilization rate indicates Tesla is running only a few vehicles at a time – meaning that mileage might have actually gone down.
So that’s how the rate of Tesla robotaxi accidents was calculated. What’s the rate for humans? Googling that question mostly led me to the pages of law firms that do crash-related litigation, and they cited a range of numbers. Often the stats are cited as your odds of being in an accident and what the fatality rate was, neither of which were helpful for me. I did find a few relevant citations. These guys claim that people drive 1.5 million miles in their lives, which sounds high to me – you’d have to drive over 20K miles a year for 75 years to get to that total. They also say the average person gets in three or four crashes in their driving lifetime, which puts the rate in the “one in 500,000 miles” range. These guys also cited “3 to 4 crashes in a lifetime”, so let’s go with that.
That leaves the total miles one drives, from which we can derive an average. This page cited Federal Highway Administration stats that say the average driver covers about 13,500 miles per year. I’m thinking if you start driving as a teenager and live to be in your 80s and drive the whole time, that puts you in the 900K lifetime miles range. At three or four crashes a year, that’s one every 225K to 300K miles. Less than that “one every 500K” number that I kept seeing, but still a lot higher than Tesla’s one-every-40K performance in Austin.
So yeah, that’s a real concern. And as those Electrek stories note, this is with a human safety supervisor in the car. Which are supposed to be going away soon. Maybe we need to be paying a little more attention to this. If better road safety is the big advantage of autonomous vehicles – and to be clear, Waymo’s record on this is vastly better – then we need to be seeing that from all of the competitors in this space. Because otherwise, what are we doing here?
UPDATE: Ready or not…
Tesla has launched its first fully autonomous robotaxi tests in Austin, Texas, sending Model Y vehicles into city streets with no driver behind the wheel. The move represents a significant milestone in Tesla’s push to compete in driverless ride-hailing.
The fleet’s early testing involved invited passengers, initially with Tesla employees at the wheel. Now, sightings confirm that cars are navigating more of Austin’s streets independently, though the fleet remains small, with around 25 to 30 vehicles.
On Sunday, Elon Musk wrote on X confirming that testing is still ongoing.
Tesla has not announced when rides will open to paying customers but has emphasized a gradual rollout.
Exercise caution when you see one of these, that’s my advice.