Baidu has launched Apollo World, an AI system designed to simulate real-world driving environments to train robotaxis and predict how those environments evolve, the company said on Tuesday, as reported by Reuters.
What Was Announced?
Apollo World can generate “up to 100,000 km” (62,000 miles) of simulated driving daily, which Baidu says is equivalent to about three years of real-world testing by an autonomous vehicle equipped with a safety driver. The system produces real-time multimodal data streams, such as camera and lidar observations, so self-driving systems can directly sense and navigate the simulated environments.
Baidu said Apollo World is already being used to train the robotaxi fleet of Apollo Go, the company’s ride-hailing arm and China’s largest robotaxi service operator, for operation globally.
Why It Matters
Training autonomous vehicles is expensive, and simulation is becoming the industry’s shortcut to scale. Automakers and tech companies are racing to cut the cost of developing self-driving technology, and virtual mileage that replicates camera and lidar feeds lets them test scenarios that are rare or dangerous on public roads.
“They still have the capex of training models … and vehicles, a challenge that every autonomous vehicle company is going to face,” Frank McCleary, an expert in autonomous mobility and urban transportation at management consulting firm Arthur D. Little, told Reuters. McCleary added that launch speed comes down to a technology’s ability to adapt to and learn from new environments, an area where AI has helped accelerate advances.
Apollo World lands in a crowded field. Apollo Go competes with rivals such as WeRide, Pony.ai and Momenta Global, while Volkswagen’s Chinese partner XPeng plans to launch robotaxi operations in 2027. Morgan Stanley has estimated China’s robotaxi fleet could grow to 1 million vehicles by 2035. Chinese firms have largely pursued asset-light expansion, supplying autonomous-driving technology and partnering with automakers, ride-hailing platforms and local fleet operators overseas rather than owning fleets outright.
The move also invites comparison with Tesla, which is seeking to expand its Cybercab robotaxi fleet using an “end-to-end neural network” trained on video of human driving, a different training philosophy from Baidu’s simulation-heavy approach.
