Vendors of AI design software, machine-vision inspection and predictive maintenance in Shanghai gain a defined grant schedule, and the plan drops revenue and track-record thresholds for suppliers.
Shanghai’s Municipal Commission of Economy and Informatization has set a target for more than 85% of the city’s industrial services, the engineering, supply chain, inspection and maintenance work sold to manufacturers, to use AI by 2028. The goal sits in a new Shanghai Action Plan to Accelerate AI-Enabled High-Quality Development of Industrial Services (2027-2028) (上海市加快推动人工智能赋能工业服务业高质量发展行动方案（2027-2028年）), which also aims to produce 10 "AI-native" unicorns. The notice carries an October 16 signature date and was made public on October 23, when The Paper reported it after learning of it from the commission.
The plan is a funding schedule as much as a target sheet. The full text, posted on The Paper’s government-account channel, lists 18 articles under six headings. Ten carry a grant ceiling between RMB5 million and RMB20 million, and nearly all name the commission as the responsible body. The notice gives no total budget, no document number and no separate effective date beyond the 2027-2028 period in its title.
The new plan adds a services layer to the factory-focused program Shanghai launched at the China International Industry Fair this month. China News Service reported the two together and kept their targets apart. That earlier action calls for 500 advanced-level smart factories, 10 vertical large models and 300 AI agents by 2028.
The plan funds five kinds of service, from design to maintenance
The commission names five areas for development: research and design, supply chain services, flexible manufacturing services, quality inspection and equipment maintenance. For design, it will back model-driven research platforms, virtual pilot testing and generative design with up to RMB10 million per project. It also wants AI-assisted chip design software and coding tools used more widely in integrated circuits. By 2028 the city wants five industry-level collaborative design platforms.
The largest single grants go to the hardest technical work. Automatic parsing of engineering drawings, intelligent process planning and production-scheduling systems can receive up to RMB20 million, as can 3D vision and image-sensor development for inline inspection and fault-prediction algorithms and digital-twin engines for maintenance. The plan pushes embedded quality inspection in integrated circuits and new energy vehicles. Sing Tao’s STNN, like The Paper, also highlighted a goal of getting 500 companies to build flexible manufacturing service systems and become advanced-level smart factories.
Cheaper compute and data are written into the subsidy list
One article, on lowering the cost of using AI inputs (降低智能要素使用成本), pays companies directly to use AI. A company that calls a third-party large model through an API or deploys one privately for an industry-specific application can receive up to RMB5 million. Shanghai will subsidize up to 30% of what companies pay to rent computing power from unaffiliated providers, and up to RMB5 million for buying high-quality training data. Cloud platforms for industrial computing are encouraged to offer low-code tools for building AI agents.
The plan also funds technology underneath the applications. Industrial AI agents, physical AI and the core components of industrial software, such as geometry kernels and solvers, are eligible for up to RMB20 million. The same ceiling applies to small models compressed to run at the edge and adapted to domestic AI chips. A separate program will turn veteran engineers’ know-how into structured training data drawn from drawings, defect records and maintenance logs.
For buyers and suppliers, the plan removes revenue, company-age and past-performance thresholds for industrial service providers. It invites insurers to cover trial use of new flexible production lines and inspection equipment. It also asks for at least 100 trainees a year in a specialist track for industrial services that the notice calls FDE, without defining the term.
Funds and district clusters will show how far the money goes
The money side of the plan is still open. The commission says it will shift the focus of its special fund for high-quality industrial development and raise the share of the city’s AI lead investment fund that goes to industrial services. It gives no percentage for either. It will also keep a rolling list of high-potential AI-native companies for tailored support, and it asks districts to build their own AI and industrial-services clusters.
National policy is moving the same way. This week the Ministry of Industry and Information Technology led a five-year plan calling for AI-native basic products across China’s industrial base. Shanghai’s version is more specific about who gets paid for what. The first call for project applications under these articles, with its own deadlines and selection criteria, will show how quickly the grant ceilings turn into contracts.