On August 31, the General Office of the Ministry of Industry and Information Technology released the "Notice on Carrying Out a Special Initiative for the Cultivation of Artificial Intelligence Application Service Providers" (hereinafter referred to as the "Notice"). The Notice clarifies the initiative's goals: using this special program as a driving force to establish a resource pool for AI application service providers, precisely enhancing their capabilities in technological innovation, system integration and delivery, and security compliance, while building a workforce of providers that understand industry pain points, master technical mechanisms, are aware of security risks, and excel in delivery and operations.
By the end of 2026, the number of service providers within the national resource pool is expected to surpass 2,000, forming a well-structured, clearly divided, and collaboratively innovative multi-tiered echelon with significantly improved delivery capabilities for complex scenarios. By the end of 2027, the pool is expected to include no fewer than 3,000 providers, supporting the formation of an AI application service ecosystem characterized by full-factor collaboration, full-chain integration, and full-scenario coverage.
Four Key Tasks Outlined
The Notice specifies that AI application service providers refer to enterprises or institutions that offer services such as consulting and planning, delivery and implementation, operations management, and security governance for AI solutions, based on the intelligent needs of user organizations. Pan Helin, a member of the Expert Committee on Information and Communications Economics under the Ministry of Industry and Information Technology and a well-known economist, noted in a WeChat interview that AI application service providers are essentially service enterprises that empower real businesses by providing AI solution offerings. The "Opinions of the State Council on Deepening the Implementation of the 'AI+' Action," released in August 2025, proposed fostering AI application service providers, developing new models such as "model-as-a-service" and "agent-as-a-service," and building an AI application service chain.
Pan Helin stated that the purpose of this special initiative is to construct a multi-tiered service provider echelon that is structurally sound, orderly in division of labor, and innovative in collaboration, adapting to the expanding landscape of AI applications and achieving large-scale implementation. It also aims to elevate service levels and optimize the supply of AI services. According to the report, this special initiative outlines four key tasks: establishing a service provider resource pool, improving service supply capabilities, driving large-scale application, and strengthening support safeguards for providers. Zhao Gang, president of the Sait Research Institute of Industry, commented via WeChat that the initiative aims to facilitate the implementation of "AI+" actions across various industries, cultivate a group of frontline deployment engineers (FDEs) who understand both AI technology and industry operations, build AI application service providers with FDEs as core talent, and nurture the growth of the AI application industry.
Lowering Trial-and-Error Costs in AI Procurement
Regarding driving large-scale application, the Notice states that service providers should be guided to package modular, standardized solution products around high-frequency, essential, and reusable business needs, creating a batch of "small, fast, light, and precise" AI products and services. Utilizing platforms such as industry alliances, trade associations, and open-source communities, targeted supply-demand matchmaking activities should be organized by industry, with promotional efforts for service providers. A range of real-world scenarios should be opened, actively introducing providers for implementation practice, advancing their technology R&D and commercialization, creating benchmark solutions, and generating demonstration effects.
At the same time, mechanisms such as first-purchase-first-use and risk compensation should be explored to increase procurement of services like large models, agents, and tokens, thereby enhancing the quality and efficiency of industries adopting AI. In regions where conditions permit, comprehensive "going global" service systems should be established, leveraging cooperative mechanisms and platforms such as the Belt and Road Initiative, BRICS, and China-ASEAN to facilitate the overseas deployment of high-quality AI application projects. Pan Helin explained that first-purchase-first-use refers to a system where government agencies and public institutions use fiscal funds to give priority to purchasing innovative products that are entering the market for the first time and have yet to establish commercial track records, thereby incentivizing accelerated innovation in the AI industry.
"Risk compensation, through mechanisms like fiscal subsidies and insurance schemes, shares the risks buyers may face when trialing new technologies, with the primary goal of lowering 'trial-and-error costs' so enterprises dare to experiment with new technology," Zhao Gang told reporters. He added that large model and agent technologies are relatively new and may be immature, potentially leading to hallucinations or security issues. To encourage companies to be the first to "eat the crab," the first-purchase-first-use policy encourages government and major enterprise users to be the first buyers and users, offering service providers market opportunities, application scenarios, and validation cases. "First purchases may entail risk, but if problems arise during trials, government risk compensation encourages experimentation," Zhao Gang noted. Through such policy measures, application can be encouraged, setting examples and facilitating widespread adoption.