Suppliers of components, materials and industrial software now have national backing to build AI models and agents into their products, with defense-industry and state-asset regulators among the plan's eight issuers.
China’s Ministry of Industry and Information Technology and seven other agencies have published a five-year plan to strengthen the country’s industrial foundations, the basic components, materials, software and processes that finished products depend on, and have written AI into its core tasks. The MIIT notice for the 15th Five-Year Plan for Improving Industrial Foundation Capabilities (《产业基础能力提升"十五五"规划》) was posted on October 22, although the document is dated September 1.
The plan sets a goal for 2030: weaknesses in China’s industrial base should be markedly eased and its foundation capabilities should reach an internationally advanced level overall, according to Xinhua. It lays out 22 key tasks in seven areas, from self-reliance and innovation to quality, product adoption and the industry ecosystem.
The co-issuers show how broad the plan is meant to be. Alongside MIIT they are the National Development and Reform Commission, the Ministry of Education, the Ministry of Finance, the Ministry of Human Resources and Social Security, the State-owned Assets Supervision and Administration Commission, the State Administration for Market Regulation and the State Administration of Science, Technology and Industry for National Defense, according to an MIIT summary. The 21st Century Business Herald described it as the first industrial-foundation plan of its kind released publicly.
AI moves from a tool in factories to a feature of basic products
The plan calls for general-purpose and industry large models for foundation sectors, and for fusing industrial process knowledge with AI-agent decision models. It also asks companies to develop "AI-native" basic products and to grow business models in which models and agents are sold as services, according to a report carried by the 21st Century Business Herald. AI is to be applied at every stage, from product research and design simulation to pilot testing and production operations.
Li Zhiqi, a special researcher at the Beijing municipal government’s counselors’ office, told the Business Herald that adding a layer of intelligence to an existing product differs from building agent decision-making into a basic product’s core. The plan asks for the second, which places the requirement on component and materials suppliers, a group usually far from the AI developers.
The plan’s frontier section also names embodied AI, alongside quantum technology, biomanufacturing, hydrogen and fusion energy, brain-computer interfaces, 6G and deep-space and deep-sea exploration, as areas where key basic products should be developed in advance.
The targets name company types but give no numbers
The 2030 goals are qualitative. The plan aims to produce a group of key basic products that lead innovation in complete machines and several internationally known brands, and to cultivate new manufacturing "single champion" companies and specialized "little giant" firms in foundation sectors, Xinhua reported. It also calls for a stronger long-term funding mechanism. No source reported numerical targets for any of these.
The main vehicle is the Industrial Foundation Reengineering Project (产业基础再造工程), and the plan pairs breakthroughs with early use, so that new domestic components get customers, China News Service reported. The Business Herald also reported that the plan improves support policies for first units, first batches and first versions of new products, which are mechanisms that reduce a buyer’s risk in adopting untested domestic equipment and software.
The plan as reported does not single out chips. In Xinhua’s English coverage, academician Zhou Ji of the Chinese Academy of Engineering said, "The most critical issue is that China’s industrial foundation is large but not strong," and named integrated circuits and industrial software among the weak points.
Implementation rests on ministries and provinces
MIIT’s notice is addressed to provincial industry authorities, which are asked to carry out the plan in light of local conditions. The plan relies on coordination among the eight agencies and between central and local governments, with monitoring and evaluation of progress, MIIT said. MIIT did not publish a written interpretation or hold a press briefing alongside the release.
The first test will be the provincial implementation plans and the first lists of supported products under the reengineering project. Those documents will show whether the AI-native goal translates into named products and funding, and which component and software makers are chosen to lead.