Implementation Plan for the Development of the Embodied Intelligence Industry in Shanghai
This plan is formulated to thoroughly implement the spirit of the Central Economic Work Conference, carry out the national action plan for accelerating the development of “AI+,” seize opportunities in the development of embodied intelligence, accelerate the formation of an innovation cluster ecosystem for future industries, and make every effort to build a global hub for innovation in the embodied intelligence industry, thereby realizing the “Model Power Gathering in Shanghai” initiative.
I. Main Objectives
Guided by the overarching principles of “model-driven, application demonstration, cluster-chain synergy, and open-source ecosystem,” and through the deployment of major technological breakthroughs, the construction of exemplary application demonstrations, the formation of distinctive industrial clusters, and the development of high-quality open-source projects, by 2027, we aim to achieve no fewer than 20 breakthroughs in core algorithms and technologies related to embodied models and embodied corpora;establish no fewer than 4 high-quality embodied intelligence incubators; achieve the clustering of 100 leading industry enterprises, the implementation of 100 major innovative application scenarios, and the promotion of 100 internationally leading products; and expand the scale of our city’s core embodied intelligence industry to exceed 50 billion yuan.
II. Model-Driven Innovation
Adhering to a model-driven approach and strengthening AI-powered capabilities, we will accelerate the construction of a fully independent innovation system covering the entire embodied intelligence technology chain. Priority support will be given to key technological breakthroughs in perception and decision-making, motion control, embodied corpora, and operating systems. Funding of up to 30% of the total approved project investment, not exceeding 50 million yuan, will be provided.
(1) Research and Development of Perception and Decision-Making Models. Support universities and research institutes in building integrated capabilities for large-scale embodied intelligence models. Encourage research into technologies such as active perception, planning and decision-making, and voice interaction to enhance core capabilities like real-time interaction and autonomous learning, thereby enabling large-scale applications.
(2) Research and Development of Motion Control Models. Support local robotic entities in collaborating with foundational model enterprises to accelerate the development of motion control models, and to build a universal motion control system and a library of manipulation skills. Breakthroughs in underlying algorithms will enhance dexterous manipulation and adaptive capabilities, fostering new paradigms for the emergence of motion intelligence and the generalization of capabilities in robotic entities.
(3) Breakthroughs in Embodied Collaboration Technologies. Accelerate the exploration of embodied intelligence technology architectures, such as dual-system and end-to-end approaches, to improve coordination between upper and lower limbs, hand-eye coordination, and brain-body coordination. Encourage the R&D of autonomous embodied models, platforms, and frameworks; achieve breakthroughs in autonomous operation and heterogeneous collaboration technologies; and enhance information sharing and collaboration levels.
(4) Embodied Data Technology. Leveraging municipal-level data platforms, collaborate with multiple parties to tackle common data technology challenges and build large-scale databases; develop data toolchains; and construct high-quality embodied datasets and a comprehensive technical system. Provide data vouchers worth up to 5 million yuan per year to enterprises purchasing data service products.
(5) Autonomous Operating Systems. Support enterprises, universities, and research institutes in jointly developing multimodal data fusion and processing technologies to create autonomous operating systems. Collaborate with robot manufacturers and research institutions to promote pilot applications of embodied intelligence operating systems in industrial manufacturing, logistics and distribution, and commercial services. Drive iterative optimization of system versions to gradually build a mature, reliable, and widely adopted ecosystem for embodied intelligence operating systems.
III. Building Public Platforms
Addressing the common needs for the development of embodied intelligence, we will integrate city-wide resources to build five major platforms—computing power, training facilities, pilot testing, investment, and leasing—to solidify the foundational infrastructure for the embodied intelligence industry. For industrial innovation service platform construction projects, support of up to 50% of the approved total project investment, not exceeding 20 million yuan, will be provided.
(6) Computing Power Platform. Optimize the municipal-level public intelligent computing power service platform and the municipal-level intelligent computing power resource coordination and dispatch platform. Coordinate the city’s existing intelligent computing resources, strengthen the dispatch and management of computing power resources, and provide computing power vouchers worth up to 40 million yuan per year to embodied intelligence enterprises that reach a certain scale of computing power rental.
(7) Digital Twin Training Facilities. Innovate methods, scenarios, tasks, and motor skills for data collection; establish digital twin training facilities; and achieve a closed-loop digital twin training cycle of “simulation training—real-world validation—iterative learning” to empower the training of embodied intelligence large models.
(8) Pilot Testing Platforms. Support multi-stakeholder efforts to build embodied intelligence pilot testing platforms. Encourage the provision of integrated pilot testing services prior to large-scale mass production to facilitate the commercialization of technological achievements, cultivate multidisciplinary teams of pilot testing engineers, and enhance the industry’s pilot testing service capabilities and supply capacity.
(9) Investment Platform. Strengthen the strategic guiding role of the municipal-level AI Pioneer Industry Mother Fund, encourage high-quality enterprises to establish corporate venture capital (CVC) funds and industrial ecosystem funds, improve the “early-stage incubation + early-stage investment” mechanism, and provide targeted support to enterprises in the embodied intelligence sector. Explore models such as contributing computing power and corpora as equity, support the development of new financial instruments tailored to embodied intelligence, and enhance financing accessibility for small and medium-sized enterprises.
(10) Leasing Platform. Leverage Shanghai’s advantages in financial resources and application scenarios to support the exploration and construction of a municipal-level leasing platform driven by a dual-engine model of “financial tool chains + industrial service chains.” Encourage financial institutions to collaborate with enterprises producing embodied intelligence terminals and core components to accelerate the entry of technological achievements into the application market. Cultivate a group of operation and maintenance organizations specializing in leasing services to enhance professional service capabilities in the embodied intelligence sector.
IV. Application Demonstration Benchmarks
Building on the city’s industry strengths and focusing on sectors such as logistics and assembly, industrial manufacturing, commercial retail, healthcare and wellness, and home services, we will launch scenario-based calls for proposals and open challenges to explore new business models for embodied intelligence applications. For demonstration projects that integrate industrial innovation, support will be provided at a rate of up to 20% of the approved total project investment, not exceeding 10 million yuan.
(11) Embodied Intelligence + Logistics and Assembly. Support the development of embodied intelligence applications in express sorting, warehouse handling, and last-mile delivery. Encourage collaboration between logistics enterprises and hub warehousing bases to create demonstration models with complete operational workflows and diverse scenario elements, thereby establishing a set of replicable and scalable new smart logistics models.
(12) Embodied Intelligence + Industrial Manufacturing. Support industrial enterprises in promoting the practical application of embodied intelligence in manufacturing processes such as assembly, inspection, emergency response, and material handling. Support leading enterprises in industries such as automotive, steel, and shipbuilding to conduct integrated innovation based on actual production needs, create human-machine collaborative smart manufacturing units, enhance the automation level of complex processes, and accelerate the construction of industry-leading embodied workshops.
(13) Embodied Intelligence + Commercial Retail. Support commercial venues in deploying embodied intelligence across multiple scenarios in spaces such as supermarkets, exhibition halls, airports, and industrial parks. Support commercial institutions in integrating services such as guided tours, customer reception, and security patrols to enhance multimodal interaction and human-machine collaboration capabilities, thereby creating new forms of intelligent, high-quality commercial services.
(14) Embodied Intelligence + Healthcare and Elderly Care. Support medical institutions and elderly care facilities in exploring demonstration applications of embodied intelligence in healthcare and elderly care services, and develop medical and nursing robots equipped with multimodal perception, compliant motion, and scenario-adaptive decision-making capabilities. Build simulation training platforms to overcome bottlenecks in multi-task adaptability under complex scenarios, develop safe and appropriate auxiliary medical and rehabilitation solutions, and drive the industry’s transition from standalone intelligent devices to group-based service models.
(15) Embodied Intelligence + Home Services.Support local smart home service providers, public housing service agencies, and other entities in collaborating with humanoid robot manufacturers and large-scale model enterprises to implement embodied intelligence in home cleaning, meal assistance, emotional companionship, and safety monitoring. Support the development of products that are deeply integrated into living spaces and enhance the sense of well-being at home. Build a motion control algorithm library driven by both data and mechanistic principles to deliver safe, human-like domestic services and create new models of smart living.
V. Coordinated Development of Clusters and Chains
Strengthen citywide coordinated planning, leverage the agglomeration effects of industrial clusters and supply chains, accelerate the development of embodied intelligence terminal products, promote the R&D and production of core components, optimize and strengthen high-quality enterprises, cultivate innovative industrial chains, build distinctive industrial clusters, and drive high-quality coordinated development across the entire industrial chain.
(16) Develop Core Components. Leverage Shanghai’s strengths in the concentration of enterprises specializing in autonomous intelligent computing chips, robotic platforms, and large-scale models to accelerate the development of core components such as dedicated embodied intelligence chips and core motherboards. Enhance the ability to coordinate and adapt software and hardware, and develop key supporting products that meet the needs of embodied robotic platforms. Promote the establishment of R&D and production bases in Shanghai for high-value component enterprises, including joint modules, smart sensors, and vision cameras.
(17) Develop embodied intelligence terminal products. Support the R&D of terminals, particularly reference models, and accelerate mass production; encourage the development of operational and intelligent embodied intelligence terminal products to create “best-selling products” and recommend their inclusion in the city’s innovative product catalog. Cultivate the industrialization capabilities of the embodied intelligence sector in phases, driving large-scale commercial applications from mature scenarios to pilot scenarios.
(18) Strengthen and optimize high-quality enterprises. Attract and cultivate a group of high-quality enterprises and industry ecosystem leaders. For enterprises that sell or lease embodied intelligent robots, provide incentives of up to 5% of the contract value, not exceeding 5 million yuan, in accordance with regulations. Deepen strategic cooperation with central and state-owned enterprises, promote the establishment of regional and functional headquarters, and support leading industry enterprises, as well as large technology and internet companies, in setting up embodied intelligent entities in Shanghai.
(19) Build distinctive industrial clusters. Establish a coordinated cultivation mechanism between the municipal and district levels. With the Zhangjiang area in Pudong as the core hub, accelerate the layout of the entire embodied intelligence industry chain. Collaborate with key regions related to embodied intelligence to form distinctive industrial clusters characterized by complementary functions and differentiated development. Introduce special policies to increase support for the first-of-a-kind (set) products in the embodied intelligence sector.Strengthen coordinated industrial planning, support the integrated development of innovation enterprise clusters, and build an embodied robotics industrial cluster that integrates R&D, manufacturing, and services. Enhance industrial synergy within the Yangtze River Delta region, promote cross-sectoral corporate collaboration, and improve the ability to integrate resources across the industrial chain.
VI. Building a World-Class Ecosystem
Focus on open-source innovation, standards development, safety and ethics, talent recruitment and cultivation, and open collaboration to further improve the supporting infrastructure and innovation ecosystem for the embodied intelligence industry.
(20) Strengthen open-source innovation. Establish a comprehensive open-source ecosystem encompassing models, data, algorithms, operating systems, and toolchains. Support the integration of open-source models and tools to provide standardized, low-threshold public services to research institutions, small and medium-sized enterprises, developers, and other stakeholders. Award open-source communities and products up to 5 million yuan in incentives in accordance with regulations.
(21) Improve the Standards System. Explore the development of an embodied intelligence standards system, with a focus on accelerating the development and provision of standards for data corpora, application scenarios, and pilot testing and evaluation. Support local enterprises in participating in the development and revision of various standards to continuously expand industry influence; provide up to 1 million yuan in support to enterprises leading standard development and revision projects.
(22) Security and Ethical Safeguards. Focusing on critical application scenarios, we will establish a security risk assessment system to ensure compliance with safety constraints and human values throughout complex interactions. We will build a multi-stakeholder collaborative governance structure, encourage public participation and continuous evaluation, and establish a closed-loop governance system to achieve the trustworthy, secure, and sustainable development of embodied intelligence.
(23) Accelerating Talent Recruitment and Development. We will accelerate the recruitment and cultivation of a group of internationally leading talents and high-level innovation teams, establishing a national pilot zone for training top-tier embodied intelligence talent. We will promote the establishment of a “industry-academia-research” collaborative, demand-driven training mechanism and build a multi-tiered talent development system. We will launch the “Embodied Intelligence 100-School Tour” initiative to provide high-quality learning and practical resources for teachers and students in our city’s schools.
(24) Deepen Open Cooperation. Leveraging platforms such as the World Artificial Intelligence Conference and the Global Developers Pioneer Conference, we will build global cooperation platforms and networks for exchange and collaboration. We will support local embodied intelligence enterprises in undertaking project collaborations with top international organizations. We will accelerate the promotion of commercial application services, broaden application scenarios, and continuously enhance the value of commercial partnerships. We will host the International Humanoid Robot Skills Competition to establish an international brand and expand into international markets.
This document shall take effect on July 28, 2025, and remain valid until July 27, 2028.
