WeRide launches WITT physical-world foundation model
WeRide introduced WITT, a physical-AI cognitive foundation model for extracting trusted physical facts from autonomous-driving data.
Evidence notes
- WITT is designed to turn real-world video and driving experience into structured physical-world knowledge.
- The model targets long-tail scenario discovery and autonomous-driving development workflows.
- WeRide positions the system as a cognition layer built from atomic physical facts rather than a vehicle-control release.
Company context
WeRide develops autonomous-driving technology and commercial autonomous mobility products across robotaxi, robobus, robovan, robosweeper and ADAS use cases.