DiDi Starts Driverless Passenger Trials With R2 Robotaxi
Passengers in parts of Beijing and Guangzhou can now summon DiDi’s purpose-built R2, equipped with 33 sensors and redundant controls.
A purpose-built vehicle enters passenger service
DiDi Autonomous Driving has begun fully driverless passenger trials with its next-generation R2 robotaxi in designated areas of Beijing and Guangzhou. Eligible users can request the vehicle through DiDi’s service, moving the R2 from road qualification and closed testing into public-facing operation without a safety driver inside.
Developed jointly with GAC Aion, the R2 is a dedicated robotaxi rather than a consumer car retrofitted after production. It uses DiDi’s full-stack Level 4 driving system and carries 33 sensors, including lidar, cameras, 4D millimetre-wave radar, infrared cameras and acoustic sensors. Their inputs are combined through sensor fusion to provide coverage around the vehicle and additional visibility in difficult lighting or weather.
A three-domain integrated computing platform coordinates autonomous-driving, cockpit and vehicle-control functions. DiDi says the vehicle incorporates redundant braking, steering, communications and power systems, while the underlying vehicle platform was designed to meet five-star Chinese and European safety standards. The passenger compartment has been optimized for rear-seat use, including seats with a wider reclining range.
The service remains geographically constrained and is still described as a trial, so the announcement does not amount to broad commercial deployment. DiDi has not disclosed the fleet size, daily ride capacity, intervention rate or timetable for expansion. Those figures will be necessary to compare the R2’s operating maturity with established robotaxi fleets.
Why it matters
DiDi controls one of the world’s largest ride-hailing networks, giving it a potentially powerful route from autonomous-vehicle testing to customer demand, dispatch data and fleet operations. Introducing a purpose-built vehicle in two major cities tests whether that network can become an advantage in driverless transport rather than simply a source of app traffic.
The important threshold is operational replication. If DiDi can add vehicles and service zones without proportionally increasing remote assistance, maintenance and safety staffing, the R2 could strengthen China’s already crowded robotaxi sector. Until fleet and intervention data are published, however, public availability demonstrates product integration more clearly than commercial scalability.