At the World Robot Conference (WRC) held on August 22, Zhongqing Robotics’ CTO, Li Liyun, laid out the company’s full-stack embodied intelligence technology roadmap and its novel AI model architecture, officially debuting the brain-inspired embodied intelligence engine, EngineAI Awaken.
According to the team, this engine fuses WAM (World Action Model) with VLA (Vision-Language-Action) models, allowing robots to hit over 98% success rates on long-horizon tasks after just two hours of real-machine deployment and fine-tuning. The result is highly efficient generalization even when sample data is scarce.
Currently, this architecture has been fully integrated with Zhongqing’s self-developed force-controlled joints, dexterous hands, and the robot body itself.