China creates first national standards for humanoid robots to support industry scale-up

China has launched its first complete nationwide customary framework for humanoid robots, marking a big step towards scaling the {industry} from early experimentation to large-scale commercialization.

The framework – titled the “Humanoid Robotic and Embodied Intelligence Commonplace System” (2026 Version) – was introduced at a technical committee assembly in Beijing on February 28, 2026, bringing collectively greater than 120 researchers, executives, and policymakers.

From silos to requirements

For an {industry} that has spent the previous couple of years in what one professional known as a “from zero to at least one” section of dazzling demonstrations and fragmented prototypes, the discharge of this framework alerts a decisive shift. China’s humanoid robotic sector is now being requested to develop up.

To know why this issues, contemplate the state of play. In response to the Ministry of Industry and Information Technology (MIIT), by the top of 2025, China was house to over 140 humanoid robotic producers, which collectively launched greater than 330 completely different fashions. The creativity was exceptional, however so was the chaos.

“Once we analyzed industrial situations,” Peng Zhihui, co-founder of Shanghai-based Agibot and a deputy director of the technical committee, defined on the Beijing convention, “we discovered that just about 80 p.c of duties the place people excel however conventional automation struggles are strongly associated to tactile sensing. The bottleneck outcomes from the absence of standardized technological pathways for tactile sensors.”

That is the issue the brand new customary system is designed to unravel. Developed collaboratively by the MIIT’s HEIS technical committee – drawing on experience from 120 analysis establishments, enterprises, and {industry} customers – the framework is structured round six pillars:

1. Foundational and customary requirements: The fundamental guidelines of the street, offering common steering and compliance assurance for technological growth.
2. Neuromorphic and clever computing: Requirements for the “cerebrum and cerebellum” of embodied intelligence, governing every thing from information administration to mannequin coaching pipelines.
3. Limbs and elements: Specs for the bodily constructing blocks – torsos, dexterous palms, actuation items, and notion modules – aimed toward enabling modular, interchangeable growth.
4. Full-system integration: Requirements for a way {hardware} and software program come collectively, making certain {that a} robotic from one producer can, in concept, combine easily with methods from one other.
5. Software: Pointers for a way robots ought to function in particular situations, from manufacturing unit flooring to houses, offering a blueprint for deployment.
6. Security and ethics: Maybe probably the most crucial pillar, overlaying bodily security, cybersecurity, and moral frameworks as autonomous machines enter human environments.

Liang Liang, secretary-general of the technical committee, described the imaginative and prescient succinctly. By unifying technical specs and analysis standards, the brand new system is meant to “scale back coordination and adaptation prices throughout the commercial chain, promote the modularization and generalization growth of upstream elements, and information R&D sources to focus on core and key areas, avoiding low-level redundant work.”

In sensible phrases, this implies encouraging interoperability. A high-quality actuator from one provider ought to match seamlessly right into a robotic chassis from one other, very similar to standardized components within the international automotive {industry}.

The ‘1 to 10’ problem

The timing of the requirements launch displays a broader inflection level for the sector. Jiang Lei, one other deputy director of the technical committee, noticed that whereas the {industry} has efficiently moved “from zero to at least one,” it now faces the extra advanced problem of scaling “from one to 10.”

This transition entails shifting from prototypes in analysis labs to mass manufacturing, and from managed demonstrations to dependable efficiency in real-world environments. Trade leaders usually describe this shift as shifting robots from “kung fu” to “working.”

Wang Xingxing, founder and CEO of Unitree Robotics and likewise a deputy director of the technical committee, introduced footage of robots performing meeting duties in factories. His message was direct: “To allow humanoid robots to genuinely work, notably on long-sequence duties, industry-wide requirements are completely important.”

The dearth of such requirements has created hurdles throughout the provision chain. Zhao Tongyang, founding father of EngineAI Robotics, drew a comparability to the automotive sector: “After almost a century of growth, each auto element has dozens and even tons of of mature suppliers. For humanoid robotics, which has solely not too long ago accelerated, the provider base stays restricted.”

This shortage will increase prices and slows innovation. With a standardized framework, the expectation is {that a} broader provider ecosystem will emerge, bettering reliability whereas decreasing prices.

Security, ethics, and public belief

A notable facet of the framework is its deal with security and ethics. As Liang Liang famous in an interview with China Day by day, making certain protected deployment – together with cybersecurity and information safety – is a precedence.

The framework consists of provisions for battery administration, {hardware} reliability, and pointers governing autonomous decision-making and human intervention.

This work is supported by organizations such because the China CEPREI Laboratory, which has been appointed co-lead of the Security Working Group. The institute will deal with requirements for “bodily security, community and information safety, clever habits security, and ethics and social governance.”

The purpose is to construct public belief, which will likely be important for adoption in houses, faculties, and public areas.

A future in factories, then houses

In response to Liang Liang, implementation is already beneath manner. The committee plans to prioritize requirements for key elements, information, and fashions, with the aim of finishing standard-setting cycles inside six months. These are meant to be “usable requirements” with sensible implementation mechanisms.

Preliminary deployment is anticipated in industrial and industrial environments – so-called “semi-structured” settings comparable to factories, logistics facilities, and public venues. As Jiang Lei outlined, the main target will likely be on “unlocking high-value, scalable potential in situations like semi-structured industrial settings, combined sorting in logistics, and inspection and safety.”

Family deployment stays a longer-term goal. “Bottlenecks exist in information know-how, computing energy, the generalization means of fashions and the event of elements comparable to dexterous palms and digital pores and skin,” Jiang defined. “If these bottlenecks aren’t addressed successfully, will probably be tough for humanoid robots to function in strange households.”

With the publication of the brand new framework, China’s humanoid robotic {industry} now has a roadmap for addressing these challenges. The section of fragmented experimentation is giving approach to coordinated industrial growth, because the sector prepares to maneuver from demonstration to deployment.

Extra about the usual

The Humanoid Robotic and Embodied Intelligence Commonplace System (2026 Version) was developed by the Ministry of Trade and Data Know-how’s Humanoid Robots and Embodied Intelligence Standardization Technical Committee (HEIS). The framework was unveiled on the committee’s inaugural annual assembly in Beijing on February 28, 2026.

It establishes six core pillars – foundational requirements, neuromorphic computing, limbs and elements, system integration, functions, and security/ethics – designed to information the sector towards modular, interoperable, and protected large-scale manufacturing.