China's MIIT Seeks Public Comments on National Standards Framework for Humanoid Robots
Nashnova编辑部
China's Ministry of Industry and Information Technology on August 24 released a draft guide for building a national humanoid-robot standards system, targeting at least 100 key standards by 2028 — a move that shifts the sector from open-ended experimentation into a rule-defined industrial phase.
Why set standards for humanoid robots now?
MIIT states plainly: no country has yet published formal humanoid-robot standards. Over a hundred are under development worldwide, but none have landed.
This means → the entire sector operates in a "everyone is building, nobody has defined what 'good' looks like" phase. China is choosing to move first on a systematic framework.
The guide's targets are concrete: ≥100 key standards completed by 2028, with over 200 companies participating in adoption and rollout.
In plain terms = within two years, "what counts as a qualified humanoid robot" gets written down — then 200-plus firms are expected to build to it.
What do the standards actually cover?
The framework spans six pillars: foundational commons, brain-inspired computing & intelligent computing, limbs & components, complete machines & systems, applications, and safety & ethics.
Foundational commons — the "bedrock" — includes terminology, classification, digital identity, and testing across nine categories. The first job: make sure everyone means the same thing.
The brain-inspired & intelligent-computing pillar covers AI training, foundation models, world models (AI models that let robots understand physical environments), and swarm intelligence — nine categories focused on data and model development.
This reflects a scope that goes well beyond hardware shells: the AI brain and software layer are equally subject to standardization.
How is safety and ethics handled?
Safety and ethics standards span the full product lifecycle, from design through decommission.
They break into seven categories: general safety principles, physical-body safety, cybersecurity & data security, model & algorithm safety, behavioral safety, application safety, and ethics & social governance.
This means → the regulatory approach is not "fix it after something goes wrong" but embeds safety requirements into every step from blueprint to market launch.
In plain terms = the robot's body must not injure people, its network connection must not leak data, its AI decisions must not go unchecked, and its role in society must not overstep — each layer gets its own enforceable standard.
What ensures the standards actually land?
The guide lays out five supporting measures, all aimed at closing the gap between writing standards and using them.
First, coordinate standards bodies so that drafting and industry rollout move together. Second, align academia, research, and industry so standards link to policy, R&D, testing, and commercialization.
Third, run training programs to guide companies toward compliance. Fourth, recruit and develop standards-focused talent. Fifth, track international standards and push for mutual recognition between Chinese and foreign frameworks.
This reflects an ambition beyond domestic rule-setting: China wants this framework to carry weight internationally.
What does this mean for the industry?
The draft marks China's humanoid-robot standardization entering a systematic build-out phase, moving from scattered preliminary research to structured framework construction.
This means → for companies, future market-entry barriers will rise significantly — products that fall short may not reach the market at all.
The key milestone to watch: whether the 100 standards can be completed on schedule by 2028 and enforced effectively will determine if this framework becomes real regulatory teeth or remains a paper plan.
Content is for reference only, not financial advice.