Gasgoo Munich- At the 2026 China Auto Forum, Cai Ning, vice president of Horizon Robotics' intelligent vehicle division, laid out a bold vision: the company aims to become a "super platform" for the Physical AI era. He detailed a strategic roadmap that extends beyond smart vehicles into general robotics and broader applications.
Image source: 2026 China Auto Forum
Physical AI is moving from concept to reality, Cai noted. Digital agents now possess the capacity for autonomous planning and execution, while embodied intelligence is rapidly moving out of the lab and into the real world.
Looking back at the PC and mobile internet eras, the winners were platform companies that combined deep software-hardware integration with massive ecosystems. Horizon Robotics believes that succeeding in Physical AI requires clearing four hurdles: underlying software-hardware synergy, true openness, cross-scenario optimization, and the ability to extend technology from smart vehicles into other industrial intelligence fields.
Smart cars are the most mature and large-scale "four-wheeled robots" available today — and the toughest testing ground for technology, Cai emphasized. Horizon Robotics has validated the reliability and leadership of its tech through mass production deliveries exceeding 10 million units and scaled data loops involving more than 40 partners.
To keep pace with evolving vehicle architectures, Horizon Robotics unveiled its "Starry Series" central computing platform. It achieves a three-domain fusion of intelligent driving, the cockpit, and body control, while also showcasing an in-vehicle AI assistant capable of memory, skill acquisition, and personalized evolution. Cai reiterated the company's commitment to openness: it provides core chips and software platforms without encroaching on partners' businesses, mandating data feedback, or locking users into an ecosystem. Looking ahead, the fourth-generation BPU will be natively built for general robots, helping the industry maximize innovation at the lowest compute cost — truly extending Physical AI capabilities from cars to everything.
