Shanghai issues a new-generation intelligent manufacturing action plan, aiming to bring 8,000 humanoid robots into factories by 2028

Shanghai has just issued a document on accelerating the development of next-generation intelligent manufacturing.

On October 27, the Shanghai Municipal Commission of Economy and Informatization issued the "Shanghai Action Plan for Accelerating the Development of Next-Generation Intelligent Manufacturing" (hereinafter referred to as the "Action Plan"). It mentions cultivating single-point implementation service providers such as corpus services, digital twins, and AI Agents; supporting the deep integration of AI technology and equipment to upgrade intelligent manufacturing equipment into equipment agents; and supporting innovative breakthroughs in robots such as humanoid robots, collaborative robots, and high-precision heavy-duty robots.

The Action Plan states that by 2028, the city will achieve full coverage of smart factories among key manufacturing enterprises, promote the full popularization and application of AI in smart factories, cultivate 500 advanced-level smart factories, achieve more than 2,000 AI application scenarios in manufacturing, develop 10 advanced and applicable vertical large models, launch more than 300 high-level agents, build new infrastructure that meets the needs of next-generation intelligent manufacturing, reach an industrial robot density of 600 units per 10,000 people, promote the entry of 8,000 humanoid robots into factories, drive the scale of the AI-empowered new intelligent manufacturing equipment industry to exceed 200 billion yuan, and reach a scale of 100 billion yuan for the intelligent manufacturing system integration industry, completing the "54321" goals.

The Action Plan mentions supporting the deep integration of AI technology and equipment and promoting the upgrade of intelligent manufacturing equipment into equipment agents. It calls for developing and promoting physical AI equipment, developing intelligent manufacturing equipment such as robots, industrial machine tools, and testing equipment with autonomous perception and decision-making capabilities, and using the demand for smart factory construction to drive equipment upgrades and large-scale application.

The Action Plan states that it supports innovative breakthroughs in robots such as humanoid robots, collaborative robots, and high-precision heavy-duty robots; accelerates the iterative upgrade of AI-empowered industrial machine tools such as ultra-precision grinding machines, large five-axis machine tools, and high-speed dual five-axis mirror milling machines; and promotes the establishment of independent capabilities in testing equipment such as X-ray photon spectrometers, residual gas mass spectrometers, and industrial CT.

In addition, the Action Plan mentions creating a new generation of intelligent manufacturing service providers. It strengthens the cultivation of next-generation intelligent manufacturing service providers through scenario opening. It promotes the intelligent upgrade of producer services such as R&D and design, quality testing, and supply chain services. It cultivates single-point implementation service providers such as corpus services, digital twins, and AI Agents; cultivates system integration service providers with full-domain capabilities spanning processes, equipment, and software and capable of cross-link innovation; strengthens original leadership in industrial AI and cultivates intelligent leading service providers that understand processes, equipment, software, and AI. It supports manufacturing enterprises in enhancing their comprehensive "product + service + solution" capabilities and developing new models and new forms of service-oriented manufacturing.

Consolidating industrial network and computing infrastructure

The Action Plan states that it supports manufacturing enterprises in carrying out new industrial network transformation and promotes the integrated evolution of control systems, industrial networks, and computing facilities. It advances pilot projects for industrial 5G/EUHT independent private networks and supports the deployment of high-quality networks such as 5G-A and 10-gigabit optical networks. It accelerates breakthroughs in high-performance AI chips. It accelerates the construction of a "hub-region-edge" computing facility system. It strengthens the coordinated scheduling of intelligent computing resources, accelerates the construction of national computing interconnection nodes, and promotes standardized access to and on-demand use of computing resources. It advances actions for inclusive supply of computing services and promotes the upgrading and efficiency improvement of industrial intelligent computing cloud platforms.

In terms of creating a new engine for industrial model empowerment, the Action Plan proposes building more than 10 vertical large models for industries such as high-end equipment, ship and marine engineering, and industrial software, and improving large models' understanding of physical laws, industrial drawings, time-series data, code, and more; launching more than 300 agents or specialized models for high-value scenarios such as aided design, adaptive production adjustment, and predictive maintenance; building a multi-level industrial model system and improving the collaboration efficiency of large and small models. It guides smart factories to take the lead in carrying out pilot applications of industrial models and promotes the application of models across the entire chain of R&D and design, pilot testing and validation, production and manufacturing, marketing services, and operations management.

Full text:

Shanghai Action Plan for Accelerating the Development of Next-Generation Intelligent Manufacturing

Next-generation intelligent manufacturing is an advanced production mode formed by the deep integration of artificial intelligence and manufacturing. To thoroughly implement the national strategic deployment for developing next-generation intelligent manufacturing and promote the intelligent transformation and upgrading of all factors, all links, and the entire process of manufacturing, this action plan is formulated.

I. Overall Goals

By 2028, achieve full coverage of smart factories among key manufacturing enterprises in the city, promote the full popularization and application of AI in smart factories, cultivate 500 advanced-level smart factories, achieve more than 2,000 AI application scenarios in manufacturing, develop 10 advanced and applicable vertical large models, launch more than 300 high-level agents, build new infrastructure that meets the needs of next-generation intelligent manufacturing, reach an industrial robot density of 600 units per 10,000 people, promote the entry of 8,000 humanoid robots into factories, drive the scale of the AI-empowered new intelligent manufacturing equipment industry to exceed 200 billion yuan, and reach a scale of 100 billion yuan for the intelligent manufacturing system integration industry, completing the "54321" goals.

The "54321" goals for next-generation intelligent manufacturing

Core carriers: cumulatively cultivate 500 advanced-level smart factories, 50 excellence-level smart factories, and 5 navigator-level smart factories.

Tiered cultivation: form a four-level smart factory cultivation system with clearly defined tiers: basic level, advanced level, excellence level, and navigator level.

Supply system: build a three-tier ladder of comprehensive service providers for next-generation intelligent manufacturing according to "single-point implementation type, system integration type, and intelligent leading type."

Public platforms: build two major service platforms with powerful functions and global reach: intelligent manufacturing supply-demand matching and intelligent manufacturing going global empowerment.

Pilot validation: Establish 1 intelligent manufacturing pilot validation platform equipped with functions such as intelligent manufacturing equipment validation, industrial large model testing, and standards compliance testing.

II. Implement the New-Generation Intelligent Manufacturing Supply Leadership Project

(1) Make breakthroughs in core technologies and basic components. Focus on common fields in the key links of intelligent manufacturing—"perception-control-execution-interconnection"—and tackle core technologies such as distributed sensing, multimodal fusion governance, high-dynamic motion planning and control, ultra-precision machining process improvement, industrial wireless real-time communication, and edge intelligence integration. Focus on making breakthroughs in basic components such as high-end sensors, high-torque-density servo motors, high-precision heavy-load joint modules, high-performance servo drives, and high-reliability reducers.

(2) Develop and promote new intelligent manufacturing equipment. Support the deep integration of AI technology and equipment, and promote the upgrade of intelligent manufacturing equipment into equipment agents. Develop and promote physical AI equipment, develop intelligent manufacturing equipment such as robots, industrial machine tools, and testing equipment with autonomous perception and decision-making capabilities, and use the demand for smart factory construction to drive equipment upgrading and large-scale application. Support innovative breakthroughs in robots such as humanoid robots, collaborative robots, and high-precision heavy-load robots; accelerate the iterative upgrading of AI-empowered industrial machine tools such as ultra-precision grinders, large five-axis machine tools, and high-speed dual-five-axis mirror milling machines; and promote the building of independent capabilities in testing equipment such as X-ray photon energy spectrometers, residual gas mass spectrometers, and industrial CT.

(3) Accelerate the creation of industrial intelligent control systems. Strengthen technical breakthroughs in high-performance fault-tolerant runtime engines, large-scale node concurrent processing, and other technologies, and improve supporting capabilities such as protocol stacks, data middleware, standardized interfaces, and dynamic tools. Accelerate the creation of high-level intelligent control products such as programmable logic controllers (PLCs), distributed control systems (DCSs), supervisory control and data acquisition (SCADA) systems, and safety instrumented systems (SISs). Focusing on key industries such as automobiles, high-end equipment, and new materials, guide enterprises to advance industrial control system upgrades in stages and create typical applications.

(4) Build a new generation of intelligent manufacturing service providers. Strengthen the cultivation of new-generation intelligent manufacturing service providers through scenario opening. Promote the intelligent upgrading of producer services such as R&D and design, quality testing, and supply chain services. Cultivate single-point implementation service providers such as corpus services, digital twins, and AI Agents; cultivate system integration service providers with full-domain capabilities spanning processes, equipment, and software and capable of cross-link innovation; strengthen original leadership in industrial AI, and cultivate intelligent leading service providers that understand processes, equipment, software, and AI. Support manufacturing enterprises in enhancing their comprehensive "product + service + solution" capabilities and develop new models and new forms of service-oriented manufacturing.

III. Implement the "AI + Manufacturing" Foundation-Building Innovation Project

(5) Consolidate industrial network and computing power infrastructure. Support manufacturing enterprises in carrying out new industrial network transformation and promote the integrated evolution of control systems, industrial networks, and computing power facilities. Advance pilot projects for industrial 5G/EUHT independent private networks, and support the deployment of high-quality networks such as 5G-A and 10-gigabit optical networks. Accelerate breakthroughs in high-performance AI chips. Accelerate the construction of a "hub-region-edge" computing power facility system. Strengthen the coordinated scheduling of intelligent computing resources, accelerate the construction of national computing power interconnection nodes, and promote standardized access to and on-demand use of computing power resources. Advance the inclusive supply action for computing power services and promote the upgrading and efficiency enhancement of industrial intelligent computing cloud platforms.

(6) Build a new engine empowered by industrial models. Build more than 10 vertical-domain large models for industries such as high-end equipment, shipbuilding and marine engineering, and industrial software, and improve the large models' understanding of physical laws, industrial drawings, time-series data, code, and the like; launch more than 300 agents or specialized models for high-value scenarios such as aided design, adaptive production adjustment, and predictive maintenance; build a multi-level industrial model system and improve the collaboration efficiency of large and small models. Guide smart factories to take the lead in carrying out pilot applications of industrial models and promote the application of models across the entire chain of R&D and design, pilot validation, production and manufacturing, marketing services, and operations management.

(7) Build high-value industrial datasets and professional service platforms. Support manufacturing enterprises in collaborating with service providers, universities, and research institutes to transform process mechanisms and expert experience into industrial knowledge bases and knowledge graphs with unified heterogeneous process semantics. Strengthen technical breakthroughs in data synthesis, semantic alignment, multimodal fusion, and other technologies, and establish 10 municipal-level high-quality manufacturing data construction consortia focusing on high-end equipment, electronic information, nuclear power, and other fields. Build high-quality industrial datasets with complete structures, accurate annotations, and model adaptability, and, focusing on experience-intensive scenarios such as design review, defect determination, and equipment diagnosis, accumulate 10 categories of "master craftsman" experience datasets. Promote the construction of a general-purpose and specialized integrated industrial corpus service platform.

(8) Promote integrated innovation of artificial intelligence and industrial software. Promote the evolution of traditional industrial software into software agents, and promote the deep application of software agents in the core links of production and manufacturing. Support manufacturing enterprises in opening scenarios, encourage AI and software enterprises to collaborate with universities and research institutes, and create chained application scenarios around R&D and design, production control, equipment operation and maintenance, and quality testing. Implement the "Hundred Enterprises, Hundred Projects" initiative for AI + industrial software, focus on directions such as AI-native industrial software development and AI-empowered intelligent upgrading of industrial software, and carry out collaborative technical and product breakthroughs.

(9) Strengthen industrial cybersecurity protection capabilities. Conduct cybersecurity attack-defense drills and emergency response drills to enhance the ability to respond to cybersecurity incidents. Promote the implementation of AI security products such as security large models and security agents in manufacturing, create 10 benchmark scenarios including intelligent penetration testing and AI security operations centers, launch 20 cybersecurity products that deeply apply large models, and build a new protection paradigm of "using models to govern models." Promote more than 10% of high-quality industrial enterprises to achieve classified and graded management of industrial internet security. Strengthen classified and graded data management and security protection, and promote more than 50 high-quality enterprises to carry out important data identification, graded protection, and risk assessment.

IV. Implement the tiered cultivation project for smart factories

(10) Build basic-level smart factories with broad coverage. Guide manufacturing enterprises to reasonably deploy necessary intelligent manufacturing equipment, software, and systems, and achieve real-time collection of core data, automation of key production processes, and informatization of production and operations management. Each district shall promote manufacturing enterprises in its region to carry out "self-assessment and self-construction" of basic-level smart factories. By 2028, the cumulative number of basic-level smart factories shall reach 1,600.

(11) Systematically build advanced-level smart factories. Promote basic-level smart factories to widely deploy intelligent manufacturing equipment, software, and systems, achieve interoperability and sharing of production and operation data, precise control of key production processes, and coordinated management and control of production and operations, accelerate AI application in typical scenarios, form regional demonstrations, and accelerate upgrading to advanced-level smart factories. Each district shall establish a "potential pool" of advanced-level smart factories. By 2028, the cumulative number of advanced-level smart factories shall reach 500.

(12) Build excellence-level smart factories through integrated optimization. Promote advanced-level smart factories to carry out comprehensive optimization across the entire product lifecycle and all links of the supply chain, promote system-level applications in multiple scenarios, achieve integrated and connected design, production, and operation data, intelligent management and control of manufacturing equipment, online optimization of production processes, deepen collaborative application of AI in multiple scenarios, and steadily leap to excellence-level smart factories. By 2028, the cumulative number of excellence-level smart factories shall reach 50.

(13) Build leading-level smart factories as benchmarks. Promote excellence-level smart factories to accelerate the deep integrated application of vertical domain models and industrial agents, explore future manufacturing models, drive changes in industrial models and enterprise forms, and strive to advance toward leading-level smart factories. Guide leading-level smart factories to export new technologies, new processes, new equipment, and new models externally, and lead changes in R&D paradigms, production methods, service systems, and organizational structures. By 2028, cultivate a cumulative total of 5 leading-level smart factories.

(14) Expand the scale of innovative scenario applications. Focus on all links including R&D and design, production and manufacturing, operation and maintenance services, marketing services, and supply chain management, and promote the AI penetration rate of basic-level smart factories to 100% according to the path of "single-scenario pilots, multi-scenario integration, and factory-level reconstruction"; the proportion of AI application scenarios in advanced-level, excellence-level, and leading-level smart factories shall reach 10%, 30%, and 70% respectively; smart factory AI application scenarios shall exceed 2,000, effectively improving enterprise production efficiency and steadily shortening R&D cycles.

(15) Create "AI + Manufacturing" model enterprises. Build 10 "AI + Manufacturing" model enterprises characterized by high-density distribution of agents and multi-scenario integration, launch landmark intelligent production lines and landmark intelligent products, promote the entry of humanoid robots into factories, explore paths to change the underlying logic of manufacturing and achieve leaps in quality and efficiency, and form full-process digital and intelligent transformation demonstrations that can be replicated and promoted within the industry.

V. Implement the project to foster a new-generation intelligent manufacturing ecosystem

(16) Promote collaborative transformation of large, medium, and small enterprises. Support chain-leading enterprises in driving upstream and downstream enterprises to carry out "chain-style" transformation. In response to the transformation needs of small and medium-sized enterprises, develop and promote specialized "small, fast, light, and precise" solutions to lower the threshold for transformation. Support enterprises in different industries and with different forms of ownership to accelerate the formation of intelligent transformation paths with their own characteristics based on existing foundations.

(17) Build a highland for intelligent manufacturing standards innovation. Improve the construction of the intelligent manufacturing standards system, and prioritize the development and revision of standards focusing on industrial agents, digital and intelligent upgrading of equipment, and the integration of intelligent scenarios with new technologies. Deepen international cooperation, promote the transformation of advanced international standards and the going-global of local standards.

(18) Build a new-generation intelligent manufacturing public platform. Build an intelligent manufacturing supply-demand matching platform, integrate the city's scenario construction needs and service providers' product capabilities, and establish a closed-loop ecosystem of "technical breakthroughs—scenario implementation—model replication." Build an intelligent manufacturing going-global empowerment platform to accelerate the overseas expansion of high-quality products and high-level solutions from intelligent manufacturing service providers. Support high-quality enterprises in jointly building intelligent manufacturing pilot-testing platforms with upstream and downstream partners in the industrial chain, and carry out pilot-testing services such as data simulation, process improvement, sample trial production, test verification, and demonstration applications.

(19) Jointly build intelligent manufacturing leading zones. Deepen city-district coordination, and support each district in formulating "one district, one policy" plans for new-generation intelligent manufacturing focused on leading industries. Support the construction of a batch of intelligent manufacturing leading zones, strengthen precise matching with high-quality enterprises and projects, promote the accelerated clustering of advantageous industries, guide each district to lead and demonstrate in production organization paradigms, new manufacturing models, and industrial chain coordination, and form a new-generation intelligent manufacturing development pattern featuring distinctive characteristics, multi-point support, and citywide linkage.

VI. Safeguard measures

Strengthen talent cultivation. Build a tiered cultivation system for interdisciplinary intelligent manufacturing talent, and implement cultivation programs for intelligent manufacturing frontier deployment engineers and smart factory continuous operation engineers. Deepen industry-education integration, improve a training course system covering multi-level intelligent manufacturing talent, and increase training for industrial workers and skilled talent. Strengthen overall coordination. Strengthen departmental coordination; the city, districts, and high-quality enterprises shall establish special task force mechanisms, establish a work system with clear responsibilities, orderly coordination, and effective supervision, strengthen task coordination, and carry out dynamic assessment. Strengthen policy support. Accelerate the planning of major projects, actively undertake major national special projects such as the "Two Major" initiatives, and make coordinated use of policies such as those promoting high-quality industrial development. Encourage each district to provide policy support in light of actual conditions.

Source: Shanghai Municipal People's Government website

Proofreader: Lü Jiubiao