Latest data from the National Development and Reform Commission shows that in the first half of 2026, in the field of intelligent robots, 8 out of every 10 humanoid and quadruped intelligent robots sold globally were produced in China. This impressive track record is encouraging, but it also raises a crucial question: Is China's robotics industry completely secure? How far do we still have to go from "leading in scale" to "achieving comprehensive leadership"?
0180% Global Market Share Releases Three Industrial Signals
Data shows that currently, there are over 140 complete machine humanoid robot enterprises in China, with an annual shipment volume of 14,400 units, accounting for 84.7% of the global share. This means 8 out of every 10 humanoid robots come from China. Zhu Keli, founding dean of the Guoyan New Economy Research Institute and chief expert on intelligent economy, believes this data releases three industrial signals.
First, China's robotics industry has completed the critical leap from "prototype R&D" to "scaled mass production," thoroughly breaking the old pattern where domestic companies focused solely on R&D while overseas enterprises monopolized the complete machine market. Currently, relying on mature scaled manufacturing capabilities for complete machines, domestic manufacturers have formed extremely strong competitiveness at the shipment level, making hardware mass production capability the most solid moat of the industry.
Second, the upstream and downstream supporting industrial chain system for domestic robots has achieved full-stack integration. From servo motors and sensors to complete machine assembly and debugging, industrial clusters now possess the capability for stable batch delivery, effectively avoiding the risk of supply disruption for single components. Meanwhile, the logic of global market demand has shifted. The industry no longer purely pursues ultimate bipedal walking performance, but values product cost-effectiveness, delivery cycles, and post-maintenance services, which perfectly aligns with the advantages of China's industrial development.
Third, this is a phased achievement resulting from the resonance of industrial policies, capital investment, and downstream manufacturing demand, marking that humanoid and quadruped robots have transformed from laboratory exhibits into commercial products recognized by the global market. It must be clarified that this data focuses on the complete machine shipment end, not indicating comprehensive leadership in technology across the entire industrial chain. It also sounds an alarm for the industry to subsequently make up for software and hardware shortcomings and stride towards high-quality development.
Zhu Keli added that the dividends of a complete whole-industry-chain cluster are the core underlying support for China to seize the global market. After years of deep cultivation, the Yangtze River Delta, the Pearl River Delta, and the Chengdu-Chongqing region have formed highly dense robotics component industrial belts, covering all categories such as basic structural parts, passive components, and power components. The short supporting distance between upstream and downstream manufacturers enables complete machine manufacturers to quickly complete component selection, prototype iteration, and batch production, significantly compressing R&D and delivery cycles. Relying on scaled production to continuously dilute manufacturing costs, they have created a cost-effectiveness advantage that fragmented overseas supply chains and single enterprises cannot replicate.
Meanwhile, technology spillover from the New Energy Vehicle (NEV) industry provides crucial impetus. The two industries are highly homologous in fields such as electronic control systems, power batteries, multi-sensor fusion, and motion algorithms. The mature three-electric system and precise complete vehicle manufacturing processes of domestic carmakers can be quickly transferred to robot R&D and production, significantly reducing hardware R&D trial-and-error costs. Coupled with the massive local iteration scenarios brought by the labor shortage in domestic manufacturing and the demand for intelligent upgrading in the service industry, as well as the continuous empowerment from the policy and capital markets, the superposition of multiple advantages ultimately forged the global scale advantage of domestic complete machines.
02Hidden Worries Beneath the Scale: Structural Imbalance and Multiple Challenges
Behind the scale leadership, the problem of China's robotics industry being "large but not strong" remains prominent. Zhu Keli stated that the industry still faces a structural imbalance of "strong hardware complete machines, but weak core components and software brains."
At the hardware level, core components such as high-end harmonic reducers, high-precision roller screws, and long-life servo motors still highly rely on overseas supply. Domestic components show obvious gaps in long-term operation stability, service life, and torque accuracy, only meeting the use of basic mass-produced models and struggling to adapt to high-end industrial scenarios. At the software level, the embodied intelligence algorithm system is not yet mature. Robots have weak environmental generalization capabilities, low operation success rates in non-standardized scenarios, and cannot autonomously handle complex and changeable work tasks, being only able to complete preset simple actions.
The industry ecosystem and development norms also face numerous issues. First, there is a lack of a standardization system. The hardware architectures and control systems of various enterprises are incompatible, making components difficult to be universally interchangeable, which drives up the operation, maintenance, and replacement costs across the industry. Second, severe involution at the low end exists, with a large number of enterprises crowding into the manufacturing of basic wheeled models, relying on low prices to seize the market, while R&D investment in high-end models for industrial precision manufacturing and special operations is insufficient. Third, industry testing norms are imperfect, lacking unified standards for product safety, lifespan, and performance, leading to uneven product quality in the market.
In addition, the cross-border entry of carmakers brings a dual impact. On the one hand, carmakers bring in mature quality control systems, supply chain resources, and global sales networks, accelerating the cost reduction and overseas expansion of complete machines, and promoting the scaled upgrading of the industry. On the other hand, their massive capital volume squeezes the survival space of small and medium-sized native robotics enterprises, forcing them to abandon the complete machine track and turn to specialized and sophisticated tracks in segmented components and vertical scenario solutions, promoting the refined division of labor in the industry.
03"Year One of Mass Production": Barriers Remain for Commercial Implementation
2026 is regarded as the "Year One of Mass Production" for humanoid robots. IDC predicts that the global shipment volume for the whole year will approach 60,000 units, with Chinese manufacturers contributing over 80%. However, there is an obvious temperature difference between mass production and commercial implementation in the industry, which is essentially caused by the structural contradiction of "mature hardware mass production technology, but immature software intelligence system."
Zhu Keli stated that current enterprise mass production solves the manufacturing challenges of hardware consistency and scaled delivery. However, industrial scenarios have stringent requirements for equipment stability and practicality, needing to meet multiple standards such as long-term continuous operation, handling sudden working conditions, rapid fault self-inspection, and equipment safety protection. Currently, mass-produced models can only complete standardized preset actions in exhibition halls. After being deployed in complex industrial production lines, the failure and error rates surge significantly. Out of safety production considerations, enterprises are reluctant to replace existing equipment on a large scale. Meanwhile, industrial enterprises value the long-term return on investment when purchasing equipment. At this stage, the high procurement cost of humanoid robot complete machines, limited battery life, and high post-maintenance costs make commercial implementation lack economic viability, which is the core reason why mass production capacity is difficult to transform into actual productivity.
"In 2026, the industry is still in the stage of commercial exploration, with shipments mainly concentrated in fields such as exhibition, scientific research, and data collection. To achieve long-term stable operation on production lines, it still needs to break through three key issues: task generalization, reliability, and data closed-loop," said Li Junlan, Research Manager at IDC China. In the future, competition in the robotics industry will be comprehensively upgraded: shifting from hardware competition to embodied intelligence capability competition; transforming from selling robots to providing Robot + AI + industry solutions, and establishing a data flywheel for the robotics industry. In the next 2 to 5 years, truly leading enterprises will not only need supply chain and manufacturing capabilities but also simultaneously possess AI model, data accumulation, engineering delivery, and ecosystem construction capabilities.
Combining the pace of industrial development, Zhu Keli outlined a clear timeline for commercial implementation: within the next 3 years, wheeled humanoid robots will take the lead in achieving scaled implementation in scenarios with high standardization and simple working conditions, such as warehousing and logistics, workshop inspection, and basic loading and unloading. Relying on existing hardware and basic algorithms, they can achieve commercial profitability and are the main models to complete implementation first. In contrast, bipedal humanoid robots with full-scenario adaptability will require at least 5 to 8 years to enter various complex industrial production lines on a large scale. This requires waiting for breakthroughs in the lifespan of high-end components, iteration of long-endurance battery technology, maturation of general embodied intelligence algorithms, and comprehensive perfection of the complete machine operation and maintenance supporting system.
04Achieving Comprehensive Leadership: China Must Play Four Key Moves
During the "15th Five-Year Plan" period, how can embodied intelligence truly achieve the ability to "think and act"? And how can the industry move from "leading in scale" to "achieving comprehensive leadership"? Zhu Keli believes that the industry must play four key moves.
First, use the scaled market to reversely drive the domestic substitution of core components. Relying on massive mass production orders, unite enterprises to carry out joint research. Aiming at "choke point" links such as reducers and servo motors, use continuous orders to guarantee R&D investment and gradually make up for the underlying hardware shortcomings.
Second, concentrate efforts to jointly build an open-source foundation for general embodied intelligence to solve the software "brain" shortcomings. Industry leaders should join hands with research institutes and AI (Artificial Intelligence) enterprises to build an open-source algorithm platform, break down data barriers, and use massive real-scenario data to train general embodied intelligence large models. This will enable robots to possess autonomous perception and complex working condition response capabilities, solving the fundamental problem of strong hardware but weak intelligence, and achieving coordinated leadership in software and hardware.
Third, promote the construction of an industry standardization system to achieve hardware modularization, interface universalization, and standardized operation and maintenance systems. Unify the external interfaces of complete machine hardware and control system communication protocols to achieve universal interchangeability of components, reducing the production and post-maintenance costs of complete machines across the industry. Introduce national mandatory industry standards covering product factory performance, safety, and service life to eliminate low-end homogenized and inferior production capacity, guide enterprises towards differentiated innovation in high-end industrial special models and segmented service models, and bid farewell to low-price involution competition.
Fourth, adhere to the dual track of "domestic scenario iteration + systematic going global." Domestically, rely on multiple scenarios such as intelligent transformation of manufacturing, elderly care services, and cultural tourism commerce to continuously iterate and polish models, forming mature and replicable commercial implementation solutions. On the going global front, adopt the model of industrial chain cluster going global and localized operations, exporting the entire industrial system of complete machines, components, and automation solutions. Integrate into the industrial development systems of countries worldwide, completing the leap from "selling hardware products" to "exporting the entire robotics industry ecosystem," and ultimately achieving comprehensive leadership across the entire industrial chain and in all aspects of software and hardware.
-END-
Editor-in-Charge / Layout: Zhu Wenfeng
Proofreader: Mei Yaxin
Supervisor: Liu Qicheng