Soochow Securities has released a research report indicating that the competitive focus in embodied intelligence is shifting from "whether movement can be achieved" to "whether tasks can be stably completed and scaled manufacturing can be realized." The firm projects the global humanoid robot market size will grow from 23 billion yuan in 2026 to 277.6 billion yuan by 2030, representing a compound annual growth rate (CAGR) of 86.4% from 2026 to 2030. With continuous improvements in core robot performance, the localization of key components, and cost reductions driven by scaled manufacturing, the commercialization of the humanoid robot supply chain is expected to accelerate.
Key Perspectives of Soochow Securities:
Humanoid robots are transitioning from prototype demonstrations to scenario validation and small-batch delivery. Embodied intelligence embeds AI capabilities into a physical robot body, creating a closed loop of "perception, decision-making, execution, and feedback." Since 2025, global leading manufacturers have successively entered the phases of factory testing, small-batch delivery, and commercial deployment. The industry's competitive focus has shifted from "whether movement can be achieved" to "whether tasks can be stably completed and scaled manufacturing can be realized."
Structured scenarios are expected to achieve commercial deployment first. Areas like manufacturing, logistics, and inspection feature high process standardization and repetitive tasks, with clear deployment boundaries, making them the first to enter the scaled validation phase. In the short term, manufacturing, logistics, and inspection remain core application directions. In the medium to long term, as embodied models and robot capabilities improve, application scenarios are expected to gradually expand to open environments like healthcare and home services.
The synergy of perception, decision-making, and execution builds the robot's core capabilities. On the perception side, the fusion of multimodal sensors and visual language models enhances the robot's ability to understand the environment, objects, and task states. On the decision-making side, large models are driving robots from rule-based control towards task understanding, reasoning planning, and autonomous decision-making, with Vision-Language-Action (VLA) models becoming the mainstream path to connect "perception, decision-making, and execution." On the execution side, technologies like joint modules, dexterous hands, and force feedback are continuously improving complex manipulation capabilities.
Cost reduction and operational efficiency improvements together determine the economics of commercialization. The cost reduction path for humanoid robots primarily involves three stages: "localization of core components, cost reduction through scaled manufacturing, and optimization of software algorithms." According to the firm's estimates, when the price of a single robot falls below 400,000 yuan and operates at over 75% work efficiency, replacing human labor with robots will lower overall labor costs, laying the foundation for large-scale commercial use.
The market space for humanoid robots will gradually open up as sales volume grows and costs decline. The firm projects the global humanoid robot market size will grow from 23 billion yuan in 2026 to 277.6 billion yuan by 2030, a CAGR of 86.4%. Based on Wright's Law, the firm estimates that when cumulative global robot sales reach 2.91 million units, the manufacturing cost per unit will fall from approximately 350,000 yuan in 2025 to 93,000 yuan.
Domestic and international core manufacturers are diverging in their approaches, with domestic manufacturers accelerating orders and capitalization. Overseas manufacturers place greater emphasis on Physical AI, foundational models, and software capabilities. Domestic manufacturers, leveraging their manufacturing base and supply chain system, are strengthening the synergy between hardware, algorithms, and application scenarios. A diverse landscape has formed in China, involving "technology companies, automakers, core robot startups, and university-affiliated teams." Products are transitioning from factory training to small-batch delivery, with industrial orders primarily driven by domestic manufacturers such as UBTECH Robotics, Zhiyuan Robot, and Unitree Robotics.
Risk Factors: Risks related to the development of humanoid robots falling short of expectations; risks related to policy falling short of expectations; risks of intensified competition within the industry; risks related to the cost reduction process for components falling short of expectations; and risks related to the application scenarios for humanoid robots falling short of expectations.