On 31 August 2026, the General Office of the Ministry of Industry and Information Technology issued the Notice on Carrying Out the Special Action to Cultivate AI Application Service Providers (MIIT General Office Science and Technology Letter [2026] No. 414, dated 27 August). The Notice implements the deployment in the State Council’s Opinions on Deeply Implementing the “AI+” Action to “cultivate AI application service providers” and “build an AI application service chain”. For the first time at the level of a ministry-level document, it defines “AI application service provider” and sets quantitative targets over three years.
Who is an “AI application service provider”
The Notice defines an AI application service provider as an enterprise or institution that, around the intelligent needs of user organisations, provides services such as consulting and planning, delivery and implementation, operations management and security governance for AI solutions.
This definition is worth noting. It is not limited to model or algorithm providers but brings the whole chain of “consulting and planning, delivery and implementation, operations management and security governance” into scope. For the many companies doing AI deployment delivery, industry solutions and system integration, this is a clear confirmation of identity — they are no longer merely “technology service providers” but are recognised by policy as part of the AI application service ecosystem.
Quantitative targets and timetable
The targets are staged: by the end of 2026, the number of providers in the national service-provider resource pool should exceed 2,000, forming a well-structured, well-divided, collaboratively innovative multi-tier echelon; by the end of 2027, no fewer than 3,000.
For localities there are more specific requirements: by the end of 2027, provinces with national AI industry innovation and application pilot zones should have no fewer than 100 enterprises in their regional resource pool. The resource pool must be reported by 1 December 2026, aggregated by each region to the Ministry of Industry and Information Technology (Department of Science and Technology), and updated periodically.
Four key tasks
Building the service-provider resource pool. Each region organises a survey of its service providers (including central enterprises registered locally), builds provider records, and the Ministry aggregates them into a national resource pool to be released in due course. This is the foundational work of the special action and the entry point for companies into the policy view.
Improving the level of service supply. Service providers are organised to lead the formation of “AI application service teams” of types such as “model-data resonance” and “computing-power-electricity coordination”, working with upstream and downstream units to build integrated solutions; innovation carriers such as national AI application pilot-testing bases and technology-enterprise incubators are used in a coordinated way; and providers are supported in tackling key technical challenges such as hardware-software adaptation and multi-model collaboration.
Promoting large-scale application. Providers are guided to encapsulate and build modular, standardised solution product packages around high-frequency, essential and reusable business, forming “small, fast, light, precise” products and services; industry-by-industry supply-demand matching is carried out; and models such as first-purchase and first-use and risk compensation are explored to increase procurement of large-model, AI-agent and Token services.
Strengthening support. Providers are supported in connecting to national computing-network hub nodes and using “computing vouchers” to lower computing costs; drawing on the results of actions such as “model-data resonance” and “industrial data foundation”, interfaces or trial channels are opened to providers; and the building of forward-deployed engineer (FDE) teams rooted at user sites is encouraged.
Compliance capability written into the task list
A detail that is easy to overlook but important: in the “improving the level of service supply” task, the Notice expressly requires “guiding service providers to embed network security, data security, ethics governance and business compliance management requirements into the whole R&D, deployment and application process, turning passive response into proactive prevention”, and requires timely provision to providers of safety-risk information and risk events concerning AI hardware and software products.
In other words, compliance capability is no longer only a regulator’s demand but one of the dimensions on which service providers are cultivated and selected. The Notice repeatedly stresses forging a team that “understands industry pain points, knows the technical mechanism, is aware of security risks and is good at delivery and operations”, listing “awareness of security risks” as one of four capabilities.
For companies doing AI delivery, this means that in bidding, shortlisting and resource-pool filing, security and compliance capability will become a substantive plus — or even a threshold. For related capability building, see our Algorithm Filing and Data Compliance service pages.
Going global and coordination
The Notice also brings going global into the scope of service providers’ tasks: regions with the conditions should build and improve a comprehensive “going global” service system and, drawing on cooperation mechanisms such as the Belt and Road Initiative, BRICS and China-ASEAN, promote the overseas landing of quality AI application projects. This connects with the overseas compliance service system in the Action Plan for the High-Quality Development of Cyberspace Enterprises; for related discussion of going-global pathways, see our Multi-Jurisdiction Compliance service.
On work arrangements, the Notice requires each region in principle to report no fewer than 10 service teams, each led by one service provider with no fewer than two upstream and downstream units participating, reported together with the resource-pool information.
Three things for AI companies
First, judge whether you are a “service provider”. If the business provides consulting and planning, delivery and implementation, operations management or security governance around clients’ intelligent needs, you should connect proactively with the local MIIT department during the resource-pool survey to avoid missing the window.
Second, make compliance capability into presentable materials. The Notice requires security and compliance to be embedded in the whole process; companies need to be able to produce institutional documents, assessment reports and filing records as evidence, not mere verbal commitments.
Third, watch the opportunity to form service teams. Service teams are the main vehicle for undertaking large-scale scenario deployment; leading or joining a team is both a business opportunity and a path into the national resource pool. For support on sorting out service-provider qualifications, building a security and compliance system and arranging delivery contracts, see our Data Compliance and Commercial Contracts services.
This article is based on the Notice on Carrying Out the Special Action to Cultivate AI Application Service Providers (MIIT General Office Science and Technology Letter [2026] No. 414) published on 31 August 2026 by the General Office of the Ministry of Industry and Information Technology. It is for reference only and does not constitute legal advice. Specific application should be based on the official document.