Two Months In: NHTSA Reports 8 Cybercab Crashes as Tesla's Robotaxi Fleet Hits Growing Pains

Source: Date:

Just two months after Tesla officially launched its robotaxi service in Austin, Texas, federal safety regulators are beginning to paint a picture of how the Cybercab fleet is performing in the real world. According to newly released data from the National Highway Traffic Safety Administration (NHTSA), Tesla's driverless vehicles were involved in 8 crashes during that window. However, when all robotaxis running Level 4 Automated Driving Systems are counted together, the number of reported incidents climbs to 43 in a single month. The figures reveal a divided story about autonomous driving safety.

NHTSA organizes these incidents into two distinct reporting categories. The first covers crashes involving Tesla's Level 4 Automated Driving System, meaning fully driverless operation with no human behind the wheel. The second tracks incidents tied to Tesla's Level 2 supervised ADAS, such as Autopilot and Full Self-Driving (Supervised), which require an attentive driver at all times.

At first glance, the driverless data looks encouraging. Most of the reported Cybercab crashes were minor in nature, and two of them weren't the fault of Tesla's vehicles at all. In many cases, the robotaxis were traveling at speeds no greater than 5 mph when the incidents occurred — typically rolling over large potholes or making low-speed contact with parked cars. Still, not every event was trivial. One September report detailed a troubling episode in which a Cybercab with a passenger on board rolled over the raised concrete edge of a parking lot, crossed a sidewalk, and descended down a curb, all at very low speed.

But experts caution against reading these numbers as proof that Tesla has cracked autonomous safety. The dataset is inherently skewed. The robotaxi service operates exclusively within dense urban zones that are carefully geofenced, and the vehicles almost never travel on highways. Under those conditions, fatal crashes are statistically unlikely regardless of how well the software performs.

That limitation is exactly why the second dataset matters. When Tesla's software is unleashed at highway speeds under Level 2 supervision, the consequences can be far more severe. NHTSA's records show 69 fatal crashes tied to Level 2 ADAS use, with 52 of them occurring at a median speed of 53 mph. Of those 69 fatalities, 29 took place at speeds exceeding 50 mph.

There is an important caveat here: these reports don't explicitly establish that Tesla's software caused the crashes. The criteria only require that the ADAS was active within 30 seconds before the collision occurred.

Taken together, the two datasets point to a clear pattern — Tesla's automated driving technology appears to have a speed problem. In controlled, low-speed urban environments, the system mostly causes minor property damage. But once it engages in high-speed highway traffic, the same software is repeatedly linked to serious and sometimes deadly outcomes.

Via

Scroll to Top