On November 2, Fuling Economic Development Zone, Chongqing Bus Operation Company, and Pony.ai signed a tripartite agreement. The day before the signing, the "Chongqing Regulations on Promoting the Development of Intelligent Connected Vehicles" officially came into effect, explicitly supporting districts and counties in opening up scenarios for pilot demonstrations of unmanned transportation. Policy windows and industrial actions are closely aligned, and a collaborative framework characterized by "enterprise as the main body, government providing the platform, and state-owned enterprises empowering" is taking shape in Fuling.
Manufacturing Side: Annual Capacity of 54,000 Units and Sales Target of 3,000 Units
Pony.ai was established as a joint venture between Xinyuan Auto and Rhino.ai (Bai Xi Niu) in April 2026, positioning itself as an OEM for RoboVans.
The Fuling production base covers 1,704 mu. On October 20, the first RX full vehicle-grade unmanned box truck rolled off the line. The annual capacity is 54,000 units, with the vehicle delivery cycle compressed to 30 days.
According to Xie Zhipeng, General Manager of Pony.ai, the RX is a full vehicle-grade unmanned logistics vehicle developed over one and a half years by Rhino.ai and Xinyuan Auto. It has a standard lifespan of 8 years and a mileage rating of 300,000 km, having completed 200,000 km of testing and verification in accordance with national IATF 16949 quality control requirements. After safety upgrades, it features 15 radars and 13 cameras, employing dual-chip redundancy backup. Cumulative product deployment has exceeded 1,000 units, with a sales target of 3,000 units for 2026 and 13,000 units for 2027.
The designed annual capacity of 54,000 units against a sales target of 3,000 units indicates a low capacity utilization rate. The RoboVan sector is currently transitioning from scenario validation to small-scale commercialization. While pre-building production lines according to planned capacity is common, whether capacity utilization can ramp up in sync with scenario opening largely depends on the actual absorption capability of downstream operational networks, while also being constrained by the pace of road right-of-way opening and commercial order density. This is also the key reason Pony.ai brought in Chongqing Bus—after vehicles roll off the line, they need management, parking spaces, and work.
Management Side: Bus Depot Network Fills the Implementation Gap
Chongqing Bus Operation Company possesses over 790 bus routes and more than 10,000 operating vehicles, with a daily passenger flow exceeding 3.44 million person-trips. In the context of traditional passenger transport, these figures represent scale; in the context of unmanned vehicles, they constitute an existing physical network: depots serve as parking nodes, charging piles as refueling facilities, and maintenance teams as operations and maintenance foundations.
Rao Minghua, Director of the Data Platform Center, stated at the signing ceremony that Chongqing Bus holds depot networks covering the central urban area, charging infrastructure, maintenance teams, and digital-intelligent operational capabilities, which are "precisely the scarce resources needed for the large-scale implementation of unmanned vehicles."
In the past, an unmanned delivery company wanting to deploy vehicles in Chongqing had to build its own depots, charging infrastructure, and O&M teams, making every dollar a heavy asset investment. By opening up this network, Chongqing Bus directly lowers the barrier to entry for unmanned vehicles. However, opening bus depots and charging piles involves practical issues such as operating hours, site safety, and liability allocation. Specific details of the jointly built asset operation platform have not yet been disclosed, and the actual degree of openness and scheduling efficiency of this network still require validation through subsequent operational data.
Fuling's Role: Closed Loop Between Manufacturing and Application Within the Same District
Fuling Economic Development Zone's role extends beyond merely providing land and policies. Fuling has constructed an industrial chain encompassing "complete vehicles + power batteries + lightweight components," aiming to establish itself as the "third pole" of Chongqing's automotive industry. Pony.ai's production line is located in Fuling; after the RX rolls off the line, it prioritizes scenario validation in Fuling before replicating across Chongqing and the Southwest region.
Wang Linlin, Director of the Fuling Economic Development Zone Management Committee, summarized the cooperation content as: manufacturing production, equity investment via industrial funds, a jointly built asset operation platform, and comprehensive scenario opening. The division of labor logic of "enterprise as the main body, government providing the platform, and state-owned enterprises empowering" is clear—each party focuses on what it does best.
The core value of this "manufacturing + scenario + operation" closed loop within the same district lies in shortening the physical distance and decision-making chain from the production line to scenario validation. Chongqing Bus's depots, charging networks, and maintenance teams are existing infrastructure. Opening them to unmanned vehicle operations is more economical than building dedicated facilities from scratch. However, the degree of resource usage rights opening, usage costs, and revenue distribution rules directly determine whether unmanned vehicle operating costs can truly decrease. While government support can resolve scenario access and initial coordination issues, whether the asset operation platform can form market-based pricing and scheduling mechanisms determines if this model remains at the level of physical space opening or transforms into sustainable transportation service capabilities.
RoboVan Transition from "Asset" to "Service": Who Determines the Ceiling?
Xie Zhipeng, General Manager of Pony.ai, stated at the signing ceremony that the goal is to transform unmanned vehicles from "equipment" into marketized public service resources. This statement points to a shift in role: previously, unmanned vehicles were several units purchased by a specific company, running proprietary routes to serve proprietary customers; in the future, unmanned vehicles will resemble buses, becoming dispatchable and allocatable public transport capacity. Express delivery companies will call upon the platform for capacity, postal services will arrange schedules via the platform for coverage, and community group buying platforms will assign orders via the platform for replenishment. The vehicles remain the same, but their identity shifts from "asset" to "service."
In this platform, Chongqing Bus assumes a trinity of roles: "property management + dispatch + maintenance"—it does not manufacture vehicles but manages them; it does not deliver goods but dispatches the vehicles that do. Pony.ai extends from a vehicle supplier to a smart transportation service provider, while Fuling transforms from a "district with an unmanned vehicle production line" to a "node with an unmanned transport capacity network."
The development potential of this model lies in the fact that when unmanned vehicles become urban public service resources, their market ceiling is no longer determined by the orders of a single enterprise, but by how much unmanned transport capacity the entire city is willing to allocate. However, building a public transport capacity platform involves a series of issues including road right-of-way allocation, pricing mechanisms, safety responsibilities, and multi-entity scheduling standards. Pony.ai's capacity corresponds to "can be manufactured," Chongqing Bus's depots and maintenance correspond to "can be managed," and Fuling's industrial chain and policy window correspond to "can be implemented quickly." The tripartite signing establishes a cooperation framework, but the answer to the commercial closed loop is not on the signing table, but in subsequent order density and the pace of road right-of-way opening.
Source: Autonomous Driving Vehicles, Publicly Available Online Data