2026 NEXT New Realm! JIDOU Technology, Zhipu, and China UnionPay jointly announced the official release of the in-vehicle native payment Agent.

January 28, 2026, a joint press conference themed "2026 NEXT New Realm | In - car Payment Agent, a New Paradigm for Travel" was held at the Modli Community in Zhangjiang, Pudong, Shanghai. Jidou Technology, Zhipu, and China UnionPay jointly launched the In - car Native Payment Agent. This product systematically reconstructs the payment capabilities around the intelligent cockpit scenario, providing new technical paths and application paradigms for the intelligent vehicle service ecosystem.

As the intelligent cockpit accelerates its evolution towards the "third living space in the car", payment has gradually become an important basic capability connecting navigation, energy, content, and services. However, current in - car payment mostly exists in the form of external app calls or function superposition, making it difficult to form a unified closed - loop with cockpit interaction and service flow, which has become a key factor restricting the further upgrade of the cockpit service experience.

The In - car Native Payment Agent launched this time is not a simple integration of traditional payment functions. Instead, led by Jidou Technology, based on its self - developed cockpit intelligent agent architecture, the payment capability is designed as a native part of the cockpit service link. Through comprehensive understanding of user intentions, vehicle status, and travel scenarios, the product realizes the coordinated operation from service triggering, decision - making and execution to payment completion within the cockpit system, allowing payment behavior to be naturally integrated into the travel experience.

In terms of the technical system, this in - car native payment agent adopts a hierarchical architecture of "model base + scenario intelligent agent". Among them, Zhipu provides the GLM large - scale model as the model base, providing the system with stable, general, and continuously evolving cognitive and reasoning capabilities. On this basis, Jidou Technology constructs an intelligent agent capability system for the in - vehicle scenario, realizing controllable execution of cockpit services and payment processes.

Zhipu, as one of the leading large - scale model R & D institutions in China, has long been deeply involved in general artificial intelligence and multi - modal large - scale model technologies. Its GLM (General Language Model) series of models have mature capabilities in language understanding, context reasoning, multi - modal semantic alignment, etc., and have been applied on a large scale in multiple industry scenarios. In this cooperation, the GLM large - scale model provides a unified semantic understanding and reasoning foundation for the in - car payment agent, enabling the system to understand users' natural language instructions, vague intentions, and context relationships in complex and continuous travel scenarios.

In the in - vehicle environment, users' expressions often have characteristics such as incomplete instructions, strong context dependence, and rapid scenario changes. The GLM's ability in long - context understanding and multi - round reasoning enables the cockpit system to maintain semantic consistency in continuous conversations and dynamic scenarios, providing stable input for the decision - making and execution of upper - layer intelligent agents, thus supporting the continuous operation of the in - car payment agent.

On top of the GLM model base, Jidou Technology, combining its more than 10 years of engineering experience in the intelligent cockpit field, constructs an intelligent agent capability system for the in - vehicle environment, including key modules such as intention analysis, task decomposition, service orchestration, and execution control. This enables the large - scale model capabilities to be transformed into implementable and mass - producible system capabilities under the conditions of limited in - vehicle computing power, high real - time requirements, and strict security levels. Relying on the practical foundation of serving more than 30 domestic and overseas automobile enterprises and covering more than 40 million mass - produced vehicle models, this in - car native payment agent can be flexibly adapted to different cockpit platforms and system architectures, reducing the complexity of automobile enterprises in terms of function integration and system deployment.

In terms of payment capabilities, China UnionPay provides the China UnionPay MCP (Model Context Protocol) payment service based on large - scale model capability calls for this product. Through this service, the in - car native payment agent can connect to multiple types of payment scenarios and services under the premise of complying with financial supervision and security regulations. Through standardized docking with the cockpit service process, UnionPay's payment capabilities can be smoothly called in typical travel scenarios such as refueling, parking, and in - car entertainment, providing reliable basic support for in - car payment scenarios and ensuring that the payment process complies with financial security and regulatory requirements.

With the collaboration of the three parties, the in - car native payment agent can already operate stably in multiple high - frequency travel scenarios. The system can combine itinerary information and user operations to achieve continuous connection between service triggering and payment execution. On the premise of ensuring safety and controllability, it significantly reduces the user's operational burden during in - car payment and improves the coherence of the overall travel experience.

Wang Yifei, CEO of Jidou Technology, said that the in - car native payment agent is one of the important implementation results under the theme of "2026 NEXT New Realm". The realization of this result benefits from the ecological empowerment of the Modli Community and the collaborative support of multiple ecological partners at different technical levels. Among them, the computing resources provided by PPIO also provide strong support for the stable operation and continuous iteration of the cockpit intelligent agent capabilities. Wen Jianhui, Vice President of Zhipu, said that Zhipu relies on its self - developed GLM series of large - scale model technologies to build the core multi - modal interaction and scenario task execution capabilities for the in - car native payment agent. This in - depth collaboration with Jidou Technology and China UnionPay is a typical practice of the integration of large - scale model technology, in - vehicle scenarios, and the payment ecosystem. The industrial agglomeration advantage of the Zhangjiang Modli Community enables the efficient linkage of the technologies and resources of all parties. Looking to the future, Jidou Technology will continue to improve the cockpit intelligent agent capability system, and jointly with Zhipu, China UnionPay, and more ecological partners, promote the large - scale application and popularization of the product in more vehicle models, further expand diversified scenarios such as in - car entertainment, health, and life consumption. At the same time, it will also actively explore the space for collaborative cooperation with excellent ecological enterprises such as Dingdong Maicai in the Modli Community in users' high - frequency life scenarios, accelerating the evolution of the intelligent cockpit towards an "integrated travel and life service platform".

Industry insiders believe that the launch of this product reflects the collaborative exploration among large - scale model technology, cockpit system engineering, and payment infrastructure. It provides a new practical path for building a service closed - loop for intelligent connected vehicles and also provides valuable references for the industry to explore the development direction of intelligent cockpits.