Tesla reveals new details about robotaxi crashes and the humans involved

At a glance:

• Tesla disclosed 17 robotaxi incidents from July 2025 to March 2026, including two crashes where remote safety monitors took control.
• Both crashes occurred in Austin at under 10 mph, involving safety monitors seated in the passenger seat.
• The incidents highlight Tesla’s reliance on human teleoperators, a practice that differs from most U.S. self‑driving fleets.

What the database shows

Tesla’s latest filing to the National Highway Traffic Safety Administration (NHTSA) lists 17 robotaxi incidents over an eight‑month span. The filing provides the most granular public view of two crashes that involved a human safety monitor actively steering the vehicle. Both events took place in Austin, Texas, and were recorded at speeds below 10 mph.

• July 2025 incident – A safety monitor asked Tesla’s remote driving team for assistance after the car stalled on the curb. A remote operator took control, drove the vehicle up the curb and into a metal fence at 8 mph. The monitor suffered minor injuries but was not hospitalized.
• January 2026 incident – After a safety monitor again requested navigation help, the remote driver steered the robotaxi into a temporary construction barricade at 9 mph. The front‑left fender and tire were scraped; no injuries were reported.

The database also notes a third, unrelated event from September 2025 where a robotaxi nudged a dog into the path of an oncoming van; the animal later ran away unharmed. These entries are the first detailed public disclosures of Tesla’s robotaxi mishaps since the company began filing crash reports in 2024.

Role of remote safety monitors

Tesla’s approach to human backstops differs markedly from other U.S. autonomous‑vehicle operators. In Tesla’s fleet, a “safety monitor” sits in the passenger seat of many robotaxis, ready to intervene. When the monitor signals trouble, a remote operator can assume full control and drive the vehicle, as seen in the two Austin crashes.

Industry peers typically limit remote input to advisory cues; Waymo, for example, allows its teleoperators to steer cars only up to 2 mph and primarily for training scenarios. Tesla’s practice of granting remote drivers direct control more frequently raises questions about the quality of video feed, latency, and the operator’s situational awareness, especially in areas with spotty cellular coverage. Independent researcher Noah Goodall warned that the crashes “raise questions about what the teleoperator can see in both coverage and resolution, and what kind of latency they are experiencing while driving.”

Impact on Tesla’s robotaxi service

Tesla’s robotaxi rollout remains modest: fewer than 100 vehicles operate across Austin, Dallas, and Houston, compared with Waymo’s nearly 4,000 units. Service availability is uneven; Reuters reported wait times of 35 minutes in Houston and Dallas, while Austin riders sometimes find no cars at all. The limited fleet size and the reliance on in‑vehicle safety monitors suggest Tesla is still refining its autonomous stack.

Elon Musk has repeatedly emphasized that autonomous driving and robotics are the company’s strategic focus, tying his compensation to vehicle and robot deliveries and future self‑driving subscriptions. The newly disclosed crashes could pressure regulators and investors to scrutinize Tesla’s remote‑driving protocols, especially as NHTSA continues to require detailed crash reporting. How Tesla adapts its safety‑monitor model may shape the competitive landscape for robotaxi services in the United States.