By Chengdu | Edited by Yang Xuran
In 2026, the wave of physical AI is accelerating from conceptual technological exploration to tangible industrial implementation. In the mobility sector, which serves as a pioneer track, Caocao Mobility has launched a model project in Xiangcheng District, Suzhou, exploring the fulfillment path for the large-scale commercial deployment of physical AI.
On October 27, Gong Xin, CEO of Caocao Mobility, was invited to attend the 8th Intelligent Driving Conference and the 2026 Huanxiu Lake Physical AI Industry Development Promotion Conference. He jointly initiated the "Suzhou Xiangcheng · Caocao Mobility Physical AI Full-Scenario Intelligent Mobility Model Project" with leaders of Suzhou Municipality. As the most critical commercialization carrier of Geely Holding Group's RoboX strategy, Caocao Mobility selected Suzhou as the first city nationwide to simultaneously deploy Robotaxi and Robovan services and operate Green Intelligent Mobility Islands. This marks a significant milestone in Caocao Mobility's RoboX strategy and a key node in the commercialization of physical AI.
Full-Factor Collaborative Implementation
Opening a New Chapter in Autonomous Driving Pilots
Over the past few years, domestic autonomous driving pilots have generally followed a route of single vehicle models and limited-area technical verification, focusing on the vehicles' autonomous mobility capabilities. Caocao Mobility's model project in Suzhou breaks this framework by exploring a full-factor collaborative mode for physical AI mobility centered on "intelligent transport capacity + infrastructure," verifying an entire operational and replicable industrial system.
The dual-track layout on the intelligent transport capacity side serves as the fulfillment foundation for this ecosystem. Caocao Mobility's Robotaxi service targets passenger mobility, covering six cities including Suzhou, Hangzhou, Shanghai, and Zhuhai, and has already initiated tests without safety drivers in the main seat in Hangzhou. The Robovan service has commenced commercial operations, covering scenarios such as express delivery, retail supermarkets, and fresh food catering. The simultaneous advancement of passenger and cargo scenarios significantly enhances the reusability of mobility capabilities and scenario extensibility.
The synchronous layout of infrastructure is a key feature distinguishing Caocao Mobility's solution from many industry pilots. Suzhou's first Green Intelligent Mobility Island has been built and put into operation, integrating functions such as automatic battery swapping, automatic cleaning, and vehicle maintenance. In the future, autonomous driving vehicles can autonomously enter stations to complete automated operations and maintenance. Meanwhile, the RAS Vehicle-Road-Cloud Integration Center has constructed a security system integrating "vehicle-cloud-road-person," providing 7×24-hour cloud-based security monitoring and remote intervention guarantees to ensure the safe and stable operation of autonomous fleets.
Thus, Caocao Mobility's full-scenario and full-factor layout in Suzhou no longer just answers the technical question of "whether autonomous vehicles can go on the road," but further verifies whether a "physical AI mobility ecosystem can operate stably and sustainably and achieve large-scale commercialization."
Trinity Closed Loop
Breaking Long-Term Bottlenecks in Physical AI Commercialization
Competition in the mobility ecosystem in the AI era has evolved from fighting for entry points to winning through integrated supply. Caocao Mobility has built a trinity closed-loop ecosystem comprising "intelligent customized vehicles + intelligent driving technology + intelligent operations."
Intelligent customized vehicles are the source of cost optimization. The design logic of traditional passenger vehicles revolves around private use scenarios, creating a natural mismatch with the operational needs of Robotaxis. Whether in terms of hardware architecture, space utilization, or maintenance convenience, it is difficult to adapt to high-frequency operational requirements. Caocao Mobility is jointly advancing the third-generation natively developed Robotaxi model with partners such as Geely Auto. Designed from the product definition stage around autonomous driving service needs, the preparation time, maintenance frequency, and depreciation curve over the full lifecycle will also be optimized accordingly. Caocao Mobility's customized version is scheduled for mass production in 2027.
Unlike many technology companies that focus on algorithmic data-driven iteration paths, Caocao Mobility possesses massive real-world mobility scenarios and operational data. It can directly feed back road condition characteristics, user demands, and dispatching pain points encountered in actual operations to the technology side, driving intelligent driving technology to iterate towards better adaptation for commercial operations.
Intelligent operations represent Caocao Mobility's core differentiating capability. Leveraging 11 years of shared mobility operational experience, Caocao Mobility has established a system covering transport capacity scheduling, vehicle asset management, user service, and compliant operations across 215 cities nationwide. This market-validated operational capability can be migrated to the era of unmanned transport capacity. Taking the independently developed CurbX Urban Curb Intelligent Engine as an example, it integrates the superior operational experience of manned ride-hailing services, supporting the large-scale screening of pick-up and drop-off points suitable for Robotaxi operations, thereby improving service landing efficiency in new cities and rationality of station networks.
According to industry research data from Goldman Sachs and China International Capital Corporation (CICC), the Chinese Robotaxi market size is expected to reach USD 47 billion by 2035, with a compound annual growth rate (CAGR) of 96%, and long-term gross margins could reach 45%. Caocao Mobility's trinity full-factor closed-loop ecosystem allows it to exert force simultaneously across three dimensions: vehicle full lifecycle costs, technology iteration efficiency, and operational fulfillment efficiency, building differentiated barriers for the large-scale commercial deployment of physical AI.
Using Suzhou as a Pivot
Laying the Path for National Replication and Internationalization
In the future, Caocao Mobility plans to advance the construction and replication of the Suzhou model in three steps: completing demonstration verification in 2026 to close the loop for both Robotaxi and Robovan businesses; initiating large-scale replication in 2027 by increasing vehicle density and expanding service scope to achieve standardization of products and operations; and entering the ecosystem leadership phase in 2030 by building an open service platform, participating in the co-construction of industry standards, and promoting the mature physical AI mobility model polished in Suzhou to the rest of the country.
Gong Xin proposed that Caocao Mobility aims to encapsulate all mobility tools—such as taxis, buses, freight trucks, unmanned vans, and low-altitude eVTOLs—as mobile service capabilities callable by various AI Agents. By connecting with global mainstream large model agents such as ChatGPT and Gemini, these agents can directly call upon Caocao Mobility's transport capacity to execute tasks in the real world. This model possesses strong general adaptability and can explore implementation in the future by combining different regional industrial environments and policy rules, providing ideas for the overseas expansion of Chinese physical AI mobility solutions.
Physical AI is regarded as the next industrial wave following mobile internet and large models. Mobility, characterized by high frequency, standardization, and strong fulfillment, has become the strategic pioneer track for the commercial implementation of physical AI. However, for physical AI to validate its commercial narrative, it must ultimately pass one test: Can an autonomous fleet continuously accept orders, replenish energy, handle exceptions, and generate profitable operational data in real cities?
This question is also a subject that Caocao Mobility must continuously verify while advancing its RoboX strategy. Now, Caocao Mobility has placed this exploration in Suzhou, attempting to gradually verify the feasible path from technical capability to scaled services for physical AI mobility through collaborative operations in real urban scenarios.