The Zhuzhou Robotaxi Crash Was China's First, and the Blame Had Three Owners
What happened
In December 2025, a Level 4 autonomous vehicle operated by Hellobike struck two pedestrians in Zhuzhou, China, shortly after passing a marked crosswalk. One man was pinned beneath the vehicle. A woman was thrown nearby. Bystanders rushed in to lift the car and free the trapped victim, who was visibly bleeding. Both were taken to intensive care at the Hunan Provincial Hospital of Traditional Chinese Medicine. Local authorities suspended all robotaxi operations in Zhuzhou indefinitely and triggered a national review of the industry.
The vehicle involved was a Baidu Apollo RT6, deployed under Hellobike's "Hello Autonomous Driving" brand. The distinction matters. The hardware and core AI came from Baidu. The operational decisions, deployment parameters, and service launch came from Hellobike, a company affiliated with Alibaba. When the collision happened, the question of where the fault actually sat had no clean answer: the cause could lie in Baidu's perception stack, in Hellobike's operating procedures, or in road conditions in Zhuzhou that neither company had fully modeled. None of the early reporting resolved it, and the architecture of the product made it straightforward for each party to point toward the other.
The timing added pressure to every angle of the story. Just days before the crash, Hellobike had announced plans to deploy 50,000 robotaxis across China by 2027. That scale would place driverless vehicles in conditions far more varied than controlled test corridors, in dense urban environments where pedestrians move unpredictably. The collision happened near a crosswalk, precisely the kind of high-traffic zone where a sensor failure or a decision-making gap carries the highest possible cost. Observers noted quickly that the pace of commercial expansion had outrun the safety validation the technology required.
China's regulators responded by ordering every robotaxi operator in the country to submit detailed emergency response reports. The Zhuzhou incident was widely described as the first major accident for China's autonomous driving industry, which also made it the first test of how the country handles the distance between an AI system's error and a legal outcome. For the two pedestrians involved, that distance is immediately practical: without a human driver, liability flows through corporate product law, which is slower and harder to navigate than a standard road accident claim.
The deeper problem the incident surfaces is one of attribution. A vehicle operated by one company, running software built by another, made a decision that put two people in intensive care, and no single entity held a complete, auditable record of what the system detected, what it decided, and who had signed off on the configuration under which it was operating. A provable record of what a system did at the moment of impact, tied clearly to the entity that owned that decision, would not prevent all future collisions. It would at least stop accountability from dissolving into the gaps between vendor contracts.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.