Two government departments have announced a plan targeting the deployment of humanoid robots in key operational scenarios by the end of 2026.
According to information from the Ministry of Industry and Information Technology (MIIT), the MIIT and the State-owned Assets Supervision and Administration Commission (SASAC) have jointly issued a notice to launch a special action plan for the real-world training and application of humanoid robots and embodied intelligence in 2026.
The notice outlines the goal that by the end of 2026, key products like humanoid robots will have completed application verification and routine deployment in a group of representative scenarios, initiating an "operational mode." It aims to distill over a hundred high-value application scenarios, further enrich the application spectrum of embodied intelligence, and foster a capacity for deployment on a scale of tens of thousands of units.
Key Tasks Outlined
To achieve this, the notice details six main tasks.
The first is to create real-world training spaces. This involves focusing on application demands for humanoid and quadruped robots in key sectors like industry, services, and specialized fields, including production, inspection, maintenance, warehousing, retail, healthcare, safety, emergency response, and disaster prevention. Relevant enterprises and institutions will be organized to select real scenario units—such as workstations, service points, or emergency stations—that have clear objectives, defined conditions, high standardization, and economic feasibility to serve as training spaces. These units will be supported to adapt their environments with minimal intervention and reuse of existing resources, defining technical requirements for training to create verifiable and testable spaces.
The second task is to form innovation and application consortia. Support will be given to establish consortia for each scenario, led by the user unit and the primary equipment manufacturer or service provider, and including supply chain partners for components, model algorithms, and research institutes.
The third is to tackle practical operational skills. The consortia will be guided to develop operational skill packages through real-world training that meet actual job requirements, forming replicable and deployable integrated solutions.
The fourth task is to strengthen real-world application verification and routine deployment. User units will be guided to establish verification procedures and success criteria based on scenario characteristics. Key metrics such as real-world task success rate, efficiency improvement, safety, reliability, and economic viability will be scientifically evaluated, resulting in an application verification report. For solutions that pass verification, provincial regions and central state-owned enterprises will promote their routine deployment in the user units and similar or related scenarios. Mature foundational products like operating systems, simulation platforms, and key components will be promoted for broader adoption across equipment manufacturers.
The fifth is to enhance key resource support. This includes supporting consortia to participate in standardization work, strengthening the implementation of key standards, improving information management for robot lifecycle tracking, and continuously refining human-robot safety collaboration mechanisms. It also involves cultivating a multi-disciplinary talent pool with expertise in both core technology and industry applications, and providing comprehensive financial services through equity, debt, and insurance instruments.
The sixth and final task is to distill mature experience. Successful training programs with advanced technology, significant results, and mature models will have their full-process operational guidelines—covering scenario adaptation, environment modification, deployment verification, and daily maintenance—extracted and shared to promote cross-regional and cross-industry exchange.