Tesla has revealed new details about two robotaxi crashes in which remote teleoperators were actively involved in controlling or assisting vehicles during incidents, shedding further light on the challenges facing the company’s autonomous driving ambitions. The disclosures emerged through newly unredacted crash reports filed with the U.S. National Highway Traffic Safety Administration (NHTSA). According to reports, the crashes occurred in Austin, Texas, between late 2025 and early 2026 while Tesla was testing or operating its robotaxi service. In one incident, a remotely assisted vehicle reportedly struck a metal fence after a teleoperator attempted to maneuver the car when the autonomous system stalled. The crash caused minor injuries to a human safety monitor inside the vehicle. In another case, a robotaxi collided with a construction barricade during remote intervention, resulting in vehicle damage but no reported injuries. The newly disclosed reports are significant because Tesla had previously requested regulators redact portions of its robotaxi crash data, arguing the information contained confidential business details. The latest filings provide one of the clearest indications yet that Tesla relies on remote human operators — sometimes referred to as teleoperators — to support its autonomous vehicles in difficult or ambiguous driving situations. Industry experts note that remote assistance is not uncommon in the autonomous vehicle sector, but Tesla’s approach appears more hands-on than many competitors. Companies such as Waymo generally use remote personnel to provide guidance or suggestions rather than directly controlling vehicles during live road operations. Reuters previously reported that Waymo stated its remote assistance teams do not directly steer or drive robotaxis in normal operations. The incidents also highlight broader technical and operational hurdles facing Tesla’s robotaxi rollout. Reports indicate the company’s service expansion into Dallas and Houston has faced issues including long wait times, inconsistent vehicle availability, navigation errors, and operational limitations.