Transitioning from the "show-off" faction performing backflips and street dances to the "practical" faction deployed in ultra-high voltage inspection, heavy-load handling, and intelligent refueling scenarios, embodied intelligent robots are shedding their "performance attributes" and accelerating their deployment on the front lines of key industrial sectors such as power, petrochemicals, and steel. In this process, central and local state-owned enterprises (SOEs), known as the "national team," have played a pivotal role. Reporters learned from the organizers of the 2026 World Robot Conference that the State-owned Assets Supervision and Administration Commission of the State Council (SASAC) will unveil the Central Enterprise Robot Innovation Consortium during the conference. This initiative aims to integrate central SOE resources across the industry and address the bottlenecks in the mass production of embodied intelligence. In fact, recently, the Ministry of Industry and Information Technology (MIIT) and SASAC jointly launched the 2026 Special Action for Real-Scene Training of Humanoid Robots and Embodied Intelligence, proposing to "activate the 'operation mode'" and "drive the formation of a deployment capacity at the scale of 10,000 units" by the end of this year. As the "national team" of the industry, major central SOEs are taking the initiative to shoulder the heavy responsibility of tackling key challenges in the industrial chain. From fundamental technology research and large-scale scenario deployment to the construction of a localized ecosystem, they are continuously driving embodied intelligent robots to deeply empower the industrial front lines and serve major national projects.
The "National Team" Demonstrates Responsibility in High-Threshold Fields Over the past few years, the focus of China's robotics industry has been highly concentrated on private and sci-tech innovation enterprises. Companies such as Unitree Robotics, Zhiyuan Robotics, and UBTECH have successively gained widespread attention and secured extensive favor from the capital market. However, it cannot be ignored that a vast number of domestic application scenarios, complex industrial data resources, and substantial financial reserves are largely held by central SOEs. Compared with private enterprises, central SOEs possess the unique advantage of penetrating high-threshold, high-investment, and high-risk core tracks, making them a crucial force for high-end breakthroughs in the industry.
The 2026 Government Work Report proposed deepening and expanding "AI+" to promote the commercial and large-scale application of AI in key industrial sectors. During interviews, reporters found that State Grid Corporation of China (State Grid) has millions of kilometers of transmission lines requiring inspection, PetroChina has tens of thousands of kilometers of oil and gas pipelines needing maintenance, and CRRC Corporation Limited (CRRC) has a massive amount of rail transit equipment requiring intelligent manufacturing. Ensuring safe production, improving quality, and reducing costs in these key industries creates a genuine and urgent demand for "replacing humans with robots." The underground utility tunnels in substations feature narrow spaces and complex working conditions. Traditional manual inspection is not only time-consuming and labor-intensive but also entails high operational risks. To address this challenge, State Grid launched the country's first hexapod bionic inspection robot specifically designed for power scenarios. Integrating multi-source sensing detection, 5G quantum encryption, and the Guangming Large Model for intelligent analysis, the robot achieves a 100% pass rate in complex, narrow, and low-clearance spaces. It can accurately identify three types of defects: partial discharge, sulfur hexafluoride (SF6) leakage, and equipment fluid leakage. Furthermore, it enables encrypted data transmission, automatic fault identification, intelligent assessment of hidden hazards, and automatic generation of disposal recommendations. The front-line working environment in traditional heavy industry is harsh and extremely labor-intensive, long plagued by the dilemma of recruiting and retaining workers. Leveraging their scenario advantages, central SOEs are accelerating the replacement of high-intensity manual positions with intelligent robots, effectively mitigating labor shortage pressures and optimizing the industrial employment structure. Today, these robots have become indispensable "steel comrades-in-arms" in nuclear island construction.
Traditional manual rebar tying requires workers to carry heavy loads in sweltering work environments, leading to immense physical exhaustion, a high risk of heatstroke, and relatively limited efficiency. Today, equipped with "crawler tracks" and "Qilin arms," robots autonomously perform high-difficulty tasks such as grabbing, positioning, and twisting rebars. Through efficient human-robot collaboration, the tying efficiency for both horizontal and vertical rebars has increased fourfold, the installation efficiency of embedded parts has increased eightfold, and the quality qualification rate has reached 99.8%. This drives the transformation of industrial workers from high-intensity physical laborers to technical managers.
Advancing R&D Through Real-Scene Field Testing To transition robots from merely "usable" to "highly effective" in industrial scenarios, laboratory prototypes must not be disconnected from real-world applications, and "show-off" style R&D must be abandoned. "Conducting technology iterations detached from real industrial working conditions is destined to fail in achieving true industrialization and deployment. Only by advancing R&D aligned with the actual needs of the industry can the barriers between technology and application be broken," Wang Zhenlin, Secretary of the Party Committee and Chairman of Beijing Research Institute of Mechanical & Electrical Technology (hereinafter referred to as "BRIAM"), told reporters. Wang Zhenlin stated that BRIAM's development of humanoid robots differs from the R&D path of some industry players that focus on general-purpose humanoid robots and pursue motion demonstrations. Instead, BRIAM deeply integrates its accumulated advantages in system integration with real industrial operation scenarios. For instance, the wheeled-arm humanoid robot developed for the complex environments of the food industry integrates visual perception, flexible manipulation, autonomous navigation, and human-robot collaboration. It can adapt to the humid and low-temperature conditions of meat processing scenarios, compensating for the lack of flexibility in traditional automated equipment. Robot training requires massive amounts of high-quality data for "feeding." Central SOEs operate millions of factories, power plants, grain silos, mines, and substations nationwide. Holding a vast array of scarce real-world scenarios, they can resolve the disconnect between simulated data and actual working conditions, thereby accelerating the iterative optimization of embodied large models. The humanoid robot industry is a high-tech field requiring heavy investment. The R&D cycles for core components such as main control chips and reducers are long, and the investments are massive. Relying on stable funding and scientific research platforms, central SOEs possess the capability to tackle "choke point" bottlenecks. It is reported that relying on its "one entity, two levels, multiple platforms" innovation system and a dedicated R&D team of nearly 200 personnel, BRIAM continues to tackle core technologies such as embodied intelligence, high-precision motion control, and human-robot collaboration. It shifts pilot verification and engineering adaptation forward into the R&D phase, proactively considering practical issues such as cost control, mass production, and production line compatibility. This approach gradually breaks down the transformation barriers from laboratory prototypes to large-scale industrial applications, systematically resolving the three core industry challenges: engineering adaptation, cost control, and mass production. Real-scene training is a crucial means to propel humanoid robots and embodied intelligence into "operation mode." Earlier this year, the MIIT and SASAC officially launched the 2026 Special Action for Real-Scene Training of Humanoid Robots and Embodied Intelligence. Regarding this, industry experts told reporters, "The launch of this special action is perfectly timed." Through repeated refinement and debugging within the massive and complex real-world scenarios of central SOEs, it can help humanoid robots and embodied intelligence products quickly rectify issues and continuously iterate and optimize. This addresses the shortcomings in real-machine data and scenario adaptation, building an industrial closed loop of "training-iteration-application-re-optimization." This serves not only as an accelerator for technology deployment but also further reduces enterprises' trial-and-error costs.
Central and Local Entities "Team Up" to Jointly Address Industrial Shortcomings The year 2026 is regarded by the industry as a critical year for humanoid robots to transition from "prototypes" to "large-scale mass production." Guolian Minsheng Securities assesses that the focus of industrial development has shifted from technology R&D to supply chain capacity building and mass production delivery. In June this year, at the PowerChina Supply Chain Cooperation and Development Conference held in the Xiong'an New Area, Chengdu Humanoid Robot Innovation Center Co., Ltd., PowerChina Chengdu Survey and Design Research Institute Co., Ltd., a major central SOE, and Zhicheng Ruijin Robot Technology (Chengdu) Co., Ltd. jointly signed a strategic procurement order for 5,000 embodied intelligent robot units. The AI-powered integrated solution for monitoring and O&M of long-distance belt conveyors by robots is one of the core products in this procurement. This solution systematically integrates independently developed long-distance belt conveyors with multi-modal intelligent inspection robots, forming intelligent equipment and operational services tailored for engineering construction scenarios such as water conservancy and hydropower, transportation engineering, and green mines. This strategic cooperation achieves complementary advantages among multiple parties: PowerChina Chengdu Survey and Design Research Institute possesses full industry chain capabilities in planning, design, and construction within the energy and power, water conservancy and water environment, and urban construction and infrastructure sectors; the Chengdu Humanoid Robot Innovation Center focuses on the R&D of humanoid robot "big and small brains," core components, and embodied intelligent engineering robots; Zhicheng Ruijin is a mixed-ownership high-tech enterprise and the physical operational entity of the country's first "Central and Local Co-built Embodied Intelligent Engineering Robot Innovation Center," boasting rich scenario advantages and resource integration capabilities. The Central Enterprise Robot Innovation Consortium, to be unveiled at the 2026 World Robot Conference, will further integrate the comprehensive resources of central SOEs, focusing on resolving the mass production bottlenecks in the humanoid robot and embodied intelligence industries. It is understood that this innovation consortium will build a collaborative platform where "central SOEs provide scenarios and requirements, private enterprises provide technology and products, and research institutes provide talent and algorithms." All participating entities will leverage their respective strengths, implementing tiered matching among large, medium, and small enterprises. They will jointly research to solve "choke points," share scenarios to solve "nowhere to test," and share outcomes to solve the "reinventing the wheel" issue, addressing these three major pain points. Industry experts stated that the establishment of the Central Enterprise Robot Innovation Consortium will propel China's robotics industry from individual enterprises fighting alone into a completely new stage characterized by national overall planning and full-chain synergy. Central SOEs will fully leverage their role as "ballast stones," driving embodied intelligence to become a new hallmark of China's high-end manufacturing exports, and continuously fostering new quality productive forces.
Author | Qi Xu, Editor | Qiu Jiangyong, Art Editor | Maliya, Supervisor | Lian Xiaodong