The Shanghai Municipal Commission of Economy and Informatization recently issued the "Shanghai Action Plan for Accelerating the Development of a New Generation of Smart Manufacturing," setting out a number of hard targets, including: by 2028, industrial robot density reaching 600 units per 10,000 workers, pushing 8,000 humanoid robots into factories, driving the scale of the AI-empowered new smart manufacturing equipment industry past 200 billion yuan, and the smart manufacturing system integration industry reaching 100 billion yuan.
Today, humanoid robots are steadily moving onto real production lines. The implementation of the plan is not an isolated industrial stimulus measure, but one move within the nationwide push for a new generation of smart manufacturing, reflecting the path and logic of intelligent upgrading in China's manufacturing sector.
Full-chain integration of AI and manufacturing
Unlike earlier smart manufacturing policies that focused on single-point technology demonstrations, the plan takes "scenario-driven, model-led, equipment upgrading, capability spillover" as its overall approach, treating the full-chain integration of AI and manufacturing as the main thread, using smart factories as the key vehicle, consolidating the smart technology foundation and strengthening the supply of smart products and services.
The most closely watched item in the plan is the "8,000 humanoid robots into factories" initiative, which is not an advance blueprint detached from industrial reality. A report previously released by International Data Corporation (IDC) showed that in the first half of 2026, global humanoid robot shipments approached 25,000 units, up 432.1% year on year. Of that, China's market shipments exceeded 19,000 units, accounting for about 77.9% of the global total.
Turning to specific application scenarios, Xiaomi's car factory has already delivered mass-production results: in less than half a year after its self-developed humanoid robot entered the Beijing Yizhuang production line, it stably overcame the challenge of long-duration continuous work on flexible workpieces such as center console covers, with an assembly success rate approaching that of skilled workers. At the end of June, Beijing Galbot Robotics Co., Ltd. reached a deep strategic partnership with CATL, with its heavy-duty humanoid robots achieving regular autonomous handling operations in power battery production workshops.
Wu Qi, a senior researcher at the Pangu Think Tank, told reporters that humanoid robots can adapt to workstation heights, aisle widths and operating tools designed around human ergonomics without large-scale modification of existing production lines, matching the transformation needs of flexible manufacturing. The quantitative goal of 8,000 humanoid robots entering factories provides a concrete, actionable anchor for smart factory construction and opens a growth curve for large-scale implementation of smart manufacturing in China.
Pushing humanoid robots into factories is only one goal, and one link in the tiered cultivation of smart factories. The plan calls for building "AI + manufacturing" model enterprises. It aims to build 10 "AI + manufacturing" model enterprises characterized by high-density distribution of intelligent agents and multi-scenario integration, launch landmark smart production lines and landmark smart products, promote humanoid robots entering factories, explore feasible paths for improving quality and efficiency, and form replicable and promotable full-process digital-intelligent transformation demonstrations within the industry.
Wu Qi said: "This means the 8,000 humanoid robots will not be spread evenly across all factories in the city, but will first run through high-value scenarios at leading-level, excellent-level and model enterprises, and then be output in reverse as replicable industry solutions. This path of first proving high-value scenarios and then replicating them also greatly reduces the uncertainty of industrial investment, allowing funds to flow precisely to technology links that have already been verified in scenarios, accelerating cost declines in core components and the maturing of the supply chain, and further feeding back into the intelligent transformation of more manufacturing scenarios, forming a positive cycle of coordination among policy, industry and capital."
Localities break through with differentiated paths
As humanoid robot technology gradually improves and is applied on the ground, localities have rolled out support policies one after another, with both common features and differences in their approaches. For example, in August 2025, the "Several Measures of the Beijing Economic-Technological Development Area on Promoting the Innovative Development of Embodied Intelligent Robots" was released, using real money to support robot product purchases and sales and the opening of application scenarios.
In the view of industry experts, Beijing and Shanghai have formed a highly coordinated development consensus on encouraging innovative applications and industrial implementation, both anchoring the integration of embodied intelligence and smart manufacturing and treating humanoid robots as an important vehicle for implementation. But at the specific execution level they have their own differences. The Beijing Economic-Technological Development Area has built a full-chain closed-loop system of "policy guidance - scenario traction - technology verification - application implementation - industrial upgrading" to solve common problems in enterprise development and accelerate technology iteration and product maturation. Shanghai, based on its strong local industrial manufacturing foundation, actively promotes embodied intelligent equipment into key industries such as automobiles, high-end equipment and new materials, and actively pushes the large-scale replication of a new generation of smart manufacturing in industrial scenarios.
Shenzhen's policy, by contrast, focuses on cluster effects. The "Shenzhen Action Plan for Technological Innovation and Industrial Development of Embodied Intelligent Robots (2025-2027) (Revised Edition)" released in June this year set out three major tasks: leading breakthroughs in core technology, building a matrix of public service platforms, and creating the best ecosystem for technological innovation. The document clearly aims for more than 1,200 companies related to the embodied intelligent robot industry cluster by 2027.
In practice, localities are turning their own core strengths into unique competitiveness in the embodied intelligence track. Zhang Xiaorong, dean of the Deep Technology Research Institute, told the Securities Daily that local competition in China's embodied intelligence industry has long moved past the initial stage of homogenized subsidy battles. Localities, relying on their own industrial endowments, have embarked on differentiated breakthrough paths. This diversified exploration not only avoids the fragmented internal friction of industrial resources, but also lays the foundation for forming a superposable and complementary industrial ecosystem at the national level, giving the large-scale implementation of the embodied intelligence industry a multi-layered support system.
Comments