Two departments: By the end of 2026, a "data-model-scenario application" virtuous cycle will be basically formed to promote the high-level empowerment of artificial intelligence in the new industrialization.

On April 28, the General Office of the Ministry of Industry and Information Technology and the National Bureau of Statistics Comprehensive Department issued a notice on the joint implementation of the "Modular Resonance" action plan for 2026. The notice aims to focus on industries such as steel, petrochemicals, non-ferrous metals, building materials, industrial machinery, automobiles, medical equipment, power equipment, ships, aerospace, home furnishing, pharmaceuticals, biomanufacturing, history classics, electronic components, consumer electronics, new displays, software, information and communication, network security, etc. (referred to as "industries"), to promote the development of a batch of high-value and technically feasible artificial intelligence application scenarios, develop industry models containing industrial and information technology mechanisms, establish high-quality common and specialized industry datasets, foster joint efforts, optimize industry ecological support systems for talent and standards. By the end of 2026, they aim to establish a virtuous cycle of "data-model-scenario application" to promote high-level AI empowering new industrialization. The original text: Notice from the General Office of the Ministry of Industry and Information Technology and the Comprehensive Department of the National Bureau of Statistics on the Joint Implementation of the "Modular Resonance" Action for 2026 Industrial and Information Technology Department Joint Letter [2026] No. 193 Provincial, autonomous region, and municipality industrial and information technology authorities, telecommunications management bureaus, data management departments, relevant central enterprises: In order to thoroughly implement the decision-making arrangements of the Party Central Committee and the State Council on the development of artificial intelligence, implement the Opinions on Deepening the Implementation of the "AI+" Action of the State Council and the Implementation Opinions of the "AI+ Manufacturing" Special Action, and promote the synergy and resonance between AI models and data resources, the Ministry of Industry and Information Technology and the National Bureau of Statistics are jointly launching the "Modular Resonance" action for 2026. The relevant matters are notified as follows: 1. Work Objectives Targeting industries such as steel, petrochemicals, non-ferrous metals, building materials, industrial machinery, automobiles, medical equipment, power equipment, ships, aerospace, home furnishing, pharmaceuticals, biomanufacturing, history classics, electronic components, consumer electronics, new displays, software, information communication, network security, etc. (referred to as "industries"), through the action, they aim to promote the development of a batch of high-value and technically feasible artificial intelligence application scenarios, develop industry models containing industrial and information technology mechanisms, build high-quality common and specialized industry datasets, foster joint efforts, optimize industry talent, standards, and other industrial ecological support. By the end of 2026, a virtuous cycle of "data-model-scenario application" will be basically formed to promote high-level AI empowering new industrialization. 2. Key Tasks (A) Construct common industry datasets, build industry models Based on the unique strengths of each provincial region/central enterprise, within the industries listed in the "work objectives," select several key industries to determine the scope of action (each provincial region should select at least 3, each central enterprise should select at least 1). Analyze the data resources within each industry, clarify the production subjects, data types, resource scales, etc., and refine and form high-quality common industry datasets (no less than 5 for each industry) through data annotation, knowledge engineering, and other methods, forming a list of "Key Industry High-Quality Common Datasets." Based on the common industry datasets, focus on the pain points in AI applications within the industry, develop industry models that understand industry technology mechanisms and serve common industry application scenarios (develop at least 1 for each industry), and, under the premise of complying with laws and policies and ensuring security, carry out applications in a stable and orderly manner (with no less than 5 application cases for each industry model). At the end of the action, summarize and compile a list of "Key Industry Models" to provide a common foundation for industry intelligence transformation. (B) Identify high-value scenarios, build specialized industry datasets, develop feature intelligent bodies Around the selected key industries, identify and cultivate a number of AI application scenarios with great potential, wide application scope, strong replicability, suitable for transformation through AI technology, and elaborate and compile the "Key Industry AI High-Value Scenario Files" (no less than 30 for each industry). For the needs of scaling up various high-value scenarios, construct specialized industry datasets containing specialized knowledge related to the scenarios (develop at least 1 dataset for each scenario), and build specialized models or feature intelligent bodies that meet the specific application needs of the scenarios (if the scenario application requires self-planning and execution capabilities, develop intelligent bodies; in other cases, develop specialized models, with at least 1 for each scenario). Under the premise of complying with laws and policies and ensuring security, carry out the practical implementation of specialized models/feature intelligent bodies in a stable and orderly manner (with no less than 3 landing cases for each specialized model/feature intelligent body). At the end of the action, summarize and compile a list of "Key Industry Specialized High-Quality Data Sets" and "Specialized Model/Feature Intelligent Body Lists." (C) Establish a sound evaluation dataset, improve model evaluation mechanisms Relying on qualified professional institutions, establish a customized evaluation dataset for the evaluation needs of industry models, feature intelligent bodies, etc., play the benchmark role of the evaluation dataset in model capability diagnosis, construct a model capabilities evaluation system oriented towards industry applications and specific scenarios, and use model evaluation results as important basis for the construction and optimization of high-quality datasets in the industry, forming a virtuous cycle of "evaluation diagnosis-data set directional optimization-model capability enhancement." (D) Create a "Modular Resonance" space, explore collaborative mechanisms Select third-party neutral institutions or leading enterprises in various regions as the main body for building and operating the "Modular Resonance" space, including developing a set of software and hardware infrastructure capable of aggregating cross-entity data and training models, as well as establishing a set of management mechanisms capable of achieving cross-entity data collaboration, model co-construction, responsibility division, and security guarantee, with the ability to ensure the reliable and efficient flow of cross-entity data, collaborative training of models, and safe and compliant application (each provincial region develops no less than 3, each central enterprise develops no less than 1). Encourage the "Modular Resonance" space to be interconnected with the national data infrastructure, enabling efficient and trustworthy circulation of multi-entity data, empowering model training, intelligent body development, and application, gradually evolving into an "intelligent body factory." (E) Build "Modular Resonance" innovation consortia, develop full-stack solutions Around the selected key industries, guide computing companies, model companies, data companies, and application development companies to form "Modular Resonance" innovation consortia (at least 1 for each selected key industry), support consortia in collaborating with the "Modular Resonance" space, jointly engage in model research and development, software and hardware adaptation, data processing, application scenario design, and development. Organize consortia to strengthen industry-level, full-stack solution research and development and application, creating a "model space" that empowers industries with AI. (F) Enhance ecological support, strengthen key element guarantees Through activities such as organizing the "Deepening Campaign" for AI empowering new industrialization in key industries, establishing training bases, etc., promote exchanges, project cooperation, and outcome dissemination among various entities. Systematically train multidisciplinary talents proficient in industry applications, data science, and model mechanisms. Strengthen AI, data technology innovation, and tool research and development through methods such as "Revealing Leaders" and others. By conducting activities such as the "Standard Campaign," increase the promotion of key standards in AI data engineering, encourage enterprises to participate in the development and promotion of AI-related standards. (G) Identify "Key Cities" as benchmarks Select cities with a strong foundation in the AI industry and abundant data resources (including districts under municipalities) as key cities for the implementation of the "Modular Resonance" action. Among them, the provincial government for the National Data Element Comprehensive Experimental Zone, the subordinate National AI Innovation Application Pilot Zone (referred to as "Pilot Zone"), or the National AI Application Pilot Base in the manufacturing industry (referred to as "Pilot Base") can select up to 3 (priority given to cities in the Pilot Zone and cities with Pilot Bases), while other provincial governments can select up to 2. Key cities should select several key industries within the scope of industry focus in the province, combine local strengths, determine work strategies, and develop implementation plans to advance the tasks mentioned above. 3. Work Requirements (A) Develop implementation plans. Provincial industrial and information technology authorities, together with data management departments (for regions involving the "information communication" and "network security" industries, coordination with local telecommunications management bureaus is also necessary) and relevant central enterprises, should organize relevant forces to develop implementation plans for the "Modular Resonance" action (see template in Appendix 1) around the seven key tasks (task seven to be executed by provincial government departments only), and send them to the Ministry of Industry and Information Technology (Science and Technology Department), and the National Bureau of Statistics (Digital Science and Infrastructure Construction Department) by May 30, 2026. (B) Track and conduct midterm evaluations. By August 30, 2026, each action implementation region and entity should review the progress of the action and produce a phase summary report (template in Appendix 2), sending it to the Ministry of Industry and Information Technology (Science and Technology Department) and the National Bureau of Statistics (Digital Science and Infrastructure Construction Department). The two departments will organize relevant experts, form service teams, and evaluate the progress through technical evaluations, on-site guidance, and other methods to optimize and adjust the implementation process based on the evaluation findings. (C) Conduct effectiveness summaries. By November 30, 2026, each action implementation region and entity should summarize the completion status of each task and submit a comprehensive summary report (also template in Appendix 2). The Ministry of Industry and Information Technology, together with the National Bureau of Statistics, will evaluate the overall implementation situation, publish the list of completed actions and cities, and review the series of achievements in high-value scenarios, industry models, specialized intelligent bodies, high-quality datasets, "Modular Resonance" space, "Modular Resonance" innovation consortia, etc. 4. Organizational Support The Ministry of Industry and Information Technology and the National Bureau of Statistics will play a coordinating and guiding role, providing policy support to various implementation regions and relevant central enterprises, creating a unified platform for carrying and displaying various achievements formed during the action, promoting and spreading advanced experiences and successful practices, and strengthening demonstrative and leading effects. For regions and enterprises with good implementation results, supportive policies and projects will be favored. Local industrial and information technology authorities, telecommunications management bureaus, and data management departments should work with action cities to strengthen organizational support, mobilize the enthusiasm of various entities, such as AI application companies, data service companies, model technology companies, AI application service providers, and provide policy and financial support to the action. Central enterprises should provide organizational support, resource input, enhance scenario openness, and intensify technological research efforts. General Office of the Ministry of Industry and Information Technology Comprehensive Department of the National Bureau of Statistics April 24, 2026 This article is compiled from the Ministry of Industry and Information Technology, edited by GMTEight: Chen Wenfang.