Tesla’s Self-Driving Cars in Austin Facing High Crash Rates
Data submitted to federal safety regulators reveals that Tesla’s self-driving vehicles in Austin, Texas, experience crashes at a rate four times higher than those driven by humans. This information was shared in a report from Tesla to the National Highway Traffic Safety Administration (NHTSA).
The fleet, consisting of 43 autonomous cars, has logged about 800,000 miles over the past eight months, resulting in 14 reported crashes. This averages out to one crash for every 57,000 miles driven by the robotaxi fleet. In contrast, Tesla’s own studies indicate that human drivers usually have a minor accident every 229,000 miles and a major incident every 699,000 miles. Additionally, NHTSA data shows that, on average, drivers face a road accident necessitating police involvement once every 500,000 miles.
Of the 14 accidents with Tesla’s self-driving cars, there were collisions involving five other vehicles, five stationary objects, one bicycle, one animal, and two categorized as “other.” Notably, one collision where an SUV was impacted at just 2 miles per hour resulted in someone needing hospitalization. This particular incident raises concerns about the risks associated with low-speed crashes in self-driving technology.
It’s worth mentioning that most of the miles driven by the Austin robotaxi fleet include safety monitors that can trigger emergency stops to avoid crashes. However, it’s uncertain how many potential collisions were avoided thanks to these human interventions.
Interestingly, crash rates appear to be worsening rather than improving. In just December and January, Tesla recorded five additional crashes, which constitute nearly 36% of all reported incidents, despite being a small portion of the total operational timeframe. This trend suggests that, contrary to expectations, autonomous systems aren’t demonstrating the usual advancements that come with more operational experience.
There are also worries about the transparency surrounding Tesla’s reporting of these incidents. For example, the hospitalization incident was initially classified in July 2025 as “property damage only,” but wasn’t changed to include hospitalization until December, five months later. This delay in updating raises important questions about the accuracy and promptness of Tesla’s accident reporting, which critics claim is often heavily redacted.
Tesla’s robotaxi fleet outcomes contrast sharply with Elon Musk’s earlier assertions regarding the reliability and expansion of the self-driving car initiative. Previous estimates predicted that, by now, the fleet in Austin would be ten times its current size and operating in many regions across the country. Instead, the program is limited to parts of just two cities, operating only 42 vehicles.
Moreover, it has been reported that fleet availability is under 20% of total operating time, illustrating substantial constraints on the program’s deployment capabilities. This low utilization rate, paired with the rising crash frequency, highlights the significant difficulties facing the advancement of self-driving vehicles.
