Baidu launched Apollo World on November 3, an AI system that simulates real-world road environments for training robotaxis and predicts how those environments evolve2,3. The system can generate up to 100,000 km of simulated driving daily, which Baidu said is equivalent to roughly three years of real-world testing by a safety-driver-equipped autonomous vehicle.
Apollo World produces real-time multimodal data streams, including camera and lidar observations, enabling self-driving systems to directly sense and navigate simulated environments.
Baidu said Apollo World is already being used to train the fleet of Apollo Go, the company's robotaxi service and China's largest robotaxi operator, for global operation. Apollo Go competes with WeRide, Pony.ai, and Momenta Global.
The launch comes as automakers race to cut the cost of developing autonomous vehicles. Developing autonomous-driving technology remains expensive, and Chinese robotaxi firms have largely pursued asset-light expansion strategies, supplying autonomous-driving technology and partnering with automakers, ride-hailing platforms, and local fleet operators overseas rather than bearing the full cost of owning and operating fleets.
Volkswagen's Chinese partner XPeng plans to launch robotaxi operations in 2027. Morgan Stanley has estimated that China's robotaxi fleet could grow to one million vehicles by 2035.
ANALYSIS By compressing years of physical road testing into a single day of simulation, Apollo World targets one of the steepest cost barriers in autonomous-vehicle development: the expense and time required to accumulate diverse driving scenarios with safety-driver-equipped cars. The system's multimodal output, spanning both camera and lidar feeds, means Baidu's self-driving stack can train on synthetic sensor data without deploying additional hardware on public roads.