In Wuhan, more than 100 Baidu robotaxis became immobilized simultaneously in a significant fleet malfunction.

      On Tuesday evening in Wuhan, over 100 of Baidu’s Apollo Go robotaxis ceased operation. They did not pull to the side, nor did they initiate an emergency protocol. They simply stopped, scattered throughout the city's streets and elevated highways, some even blocking the middle lane of ring roads while traffic flowed past them. Passengers trapped inside called the police for assistance. Videos shared on Weibo displayed Apollo Go vehicles stuck at intersections with their hazard lights flashing, unable to move. One video seemed to show the incident leading to a highway collision, though Wuhan police reported no injuries and confirmed that all passengers exited their vehicles without harm.

      The Wuhan traffic police stated on Weibo that initial investigations indicated a "system malfunction" was responsible for the vehicles coming to a halt. They noted that the cause was still under investigation. Baidu did not respond immediately to inquiries from various news organizations.

      Wuhan is not a minor test location; it is the largest deployment of Apollo Go’s fleet in China, featuring over 1,000 driverless vehicles operating across the city. According to Baidu’s latest earnings report in February, Apollo Go has been introduced in 26 cities worldwide, amassing over 20 million cumulative orders. In the fourth quarter of 2025, the service provided 3.4 million fully driverless rides, with weekly figures exceeding 300,000 during peak periods. This is not merely an experimental initiative. It constitutes a commercial operation at a real scale, which makes the simultaneous failure of more than 100 vehicles in one city distinctly more serious than an isolated event.

      Jack Stilgoe, a professor of science and technology policy at University College London, told BBC News that while driverless technology might, on average, be safer than human drivers, this incident illustrated that it could “still go wrong in entirely new ways.” This distinction is crucial. A human driver experiencing a medical emergency creates one hazard, while a system-wide failure could lead to dozens or hundreds of incidents simultaneously. “To make informed decisions about this technology, we must comprehend entirely new types of risks,” Stilgoe expressed.

      The Wuhan incident is the most significant in a series of failures that have accompanied the rapid growth of the robotaxi sector. In December 2025, a significant power outage at a Pacific Gas and Electric substation in San Francisco caused electricity loss to about a third of the city, disrupting traffic signals in multiple neighborhoods. Waymo’s robotaxis, which treat non-functioning traffic lights as four-way stops, started requesting confirmation checks from the company’s fleet response team. The sudden influx of requests overwhelmed the system, resulting in many vehicles halting and blocking streets, leading to severe traffic jams. Waymo subsequently suspended its services and later released a software update to enhance the vehicles’ navigation of disabled traffic signals.

      Three months earlier, in August 2025, an Apollo Go robotaxi in Chongqing, with a passenger aboard, drove into a construction pit after ignoring barriers and warning signs, ultimately dropping into a trench. The passenger, a woman, was safely rescued by local residents with the aid of a ladder. This incident underscored a recognized vulnerability within autonomous driving systems: their struggle to detect large, irregular road obstacles that are not included in their training datasets.

      The timing of the Wuhan incident is particularly troubling for Baidu. In December 2025, Uber and Lyft both declared partnerships with the Chinese tech firm to introduce Apollo Go vehicles to UK roads, with initial trials focused on London. Uber's pilot was set to commence in the first half of 2026 using Apollo Go RT6 vehicles, which are specially designed electric cars with removable steering wheels. Lyft's testing was planned to start with a fleet of 50 vehicles, part of its operational rollout following its $197 million acquisition of the FreeNow mobility platform. Both companies still need regulatory approval before commencing trials, and the footage from Wuhan depicting frozen robotaxis blocking traffic will likely hinder those discussions.

      Apollo Go has also made expansions into the Middle East, launching fully autonomous ride-hailing services in Abu Dhabi under the name AutoGo and obtaining the first fully driverless testing permit in Dubai. Recently, it started offering rides through the Uber app in Dubai. These regions have cultivated a regulatory environment that is intentionally accommodating, establishing Gulf cities as global centers for autonomous mobility. However, a widespread fleet failure in a flagship deployment in China raises concerns that mere hospitality cannot resolve.

      The more profound issue is systemic. When autonomous vehicles function as isolated units, a software error or sensor malfunction results in a singular event. Conversely, when operating as centrally managed fleets connected to shared backend systems and dependent on common infrastructure, they introduce a unique risk category—correlated failure. Every vehicle may fail similarly, at once, for the same reason. The consequence is not merely a car accident but a significant traffic system event.

      Regulators are keeping a close watch. Tesla has faced scrutiny from US safety authorities regarding erratic robotaxi behavior, and the broader industry