Gasgoo Munich- On June 3, Hong Kong University of Science and Technology (HKUST) and CalmCar, an autonomous driving perception and decision-making firm, have forged a strategic partnership to launch the "HKUST Physical AI Innovation Center." The initiative aims to resolve critical bottlenecks—such as the lack of physical understanding and decision reliability in AI for autonomous driving and robotics—and lay the groundwork for next-generation artificial intelligence.
Headed by Guo Song, a fellow of the Canadian Academy of Engineering and an authority on distributed computing and edge AI, the center boasts a strategic committee packed with top scholars, including Zheng Weimin of the Chinese Academy of Engineering and HKUST's Provost Guo Yike. Moving beyond traditional reliance on data fitting, the center will focus on "World Model" technology. By employing "physical alignment," it aims to equip AI with the ability to conduct high-fidelity simulations and interactions within virtual environments—bridging the gap from "perception" to "cognition and decision-making."
Image Source: CalmCar
On the industrial front, CalmCar will serve as the initial "test bed," becoming the first to apply physical AI to advanced intelligent driving systems. This means future systems won't merely "see" the road; they will understand physical causality and anticipate traffic flow evolution, significantly enhancing safety and generalization in complex urban environments.
Notably, the unveiling ceremony marked the launch of both an "Industry-Research Alliance" and a "Capital Alliance." The event drew a wide range of participants, including academic institutions like Tsinghua University and Microsoft Research Asia, domestic chipmakers such as MetaX and Huixi, and investors including HSBC and Lion X Ventures. This full-stack ecosystem signals Hong Kong's bid to leverage the Greater Bay Area's supply chain advantages, establishing a national strategic hub for physical AI and accelerating the conversion of research into real-world productivity.
