Qianxun CEO: BeiDou Shifts to Quality, Spatiotemporal Intelligence as Physical AI's Third Network
The 5th BeiDou Scale Application International Summit was held in Zhuzhou on September 15-16. Chen Jinpei, CEO of Qianxun Position, said in an interview that BeiDou scale application previously focused on quantity—how many terminals and customers were served—but the next stage will require quality improvements to drive further scale growth.
He explained that quality improvement means applying the valuable spatiotemporal data generated from BeiDou spatiotemporal intelligence to various scenarios more effectively, and truly integrating this data with AI technologies such as large models. In this process, machines will evolve from manual control and simple intelligence to deep intelligence capable of acting in the physical world—what is now called physical AI.
Chen noted that robots, as silicon-based life forms, must rely on infrastructure to give them spatial and temporal perception, whereas natural humans have innate spatiotemporal perception. His proposed solution is to build a complete infrastructure based on BeiDou satellites, with centimeter-level positioning, millimeter-level sensing, and nanosecond-level timing. He calls this infrastructure the third network after the computing network and the communication network—the third network for physical AI.
The computing network handles computing resource scheduling, the communication network handles data transmission, and the spatiotemporal network handles positioning, navigation, and collaboration for robots in the physical world. Without all three, robots cannot operate stably in real environments.
At the summit's global premiere of new products, technologies, and application scenarios, Qianxun Position officially released the Embodied Spatiotemporal Brain, SpatiXBot. Through indoor and outdoor autonomous walking, environmental perception, and swarm collaboration capabilities, it aims to accelerate the large-scale deployment of robots in inspection, security, emergency rescue, and other scenarios. Qianxun Position operates a global spatiotemporal intelligent service platform with more than 10,000 GNSS satellite-based and ground-based augmentation stations worldwide, serving about 70 million smartphones, over 5 million intelligent connected vehicles, nearly 700,000 low-altitude aircraft, and more than 80,000 embodied intelligent terminals. Its monthly service calls exceed 1.2 trillion.
The figure of 80,000 embodied intelligent terminals indicates that spatiotemporal services have achieved actual deployment scale in the robotics category. The 1.2 trillion monthly calls correspond to high-frequency positioning and timing needs; robots constantly ask where they are and what time it is. Chen places the focus of quality improvement on combining spatiotemporal data with large models. He points out that large models excel at understanding and reasoning but do not know specific locations and times in the physical world. Spatiotemporal data provides coordinate anchors and timestamps, enabling model decisions to land on specific places and moments.
Qianxun Position is also accelerating the overseas expansion of BeiDou spatiotemporal intelligence capabilities. Leveraging the technology, services, and operational experience gained from domestic infrastructure construction, it helps leading Chinese companies in intelligent driving and intelligent manufacturing go global, while promoting mature products and application models overseas. Chen believes whether the "third network" concept holds depends on the penetration speed and depth of spatiotemporal services in the robotics category. The 80,000 terminals are a starting point; as robot shipments grow, spatiotemporal service calls will amplify accordingly.