China is moving from showcasing humanoid robots to defining the rules by which they will be measured, integrated, and governed. A consultation document issued by China’s industry ministry sets a 2028 goal of no fewer than one hundred major standards for the humanoid-robot sector. Xinhua reported that the proposal spans performance assessment, core technology, platform architecture, operating environments, and safety oversight.
The proposal is still a draft. Public comments are open from Aug. 25 to Sept. 23, and the eventual standards may differ from the current plan. Yet the direction is clear. China is trying to give a rapidly expanding field a shared language for performance and safety before companies deploy machines at a larger scale. The draft aims for more than 200 enterprises to promote and implement the standards by 2028, tying technical rules to industrial participation.
Why Humanoid Robots Need Common Measures
Humanoids are difficult to compare because a demonstration can conceal as much as it reveals. One machine may walk quickly on a clear stage, another may manipulate objects well, and a third may operate in a factory cell. Without common tests, buyers cannot easily judge task success, reliability, energy use, safety behavior, environmental tolerance, or the limits of a system’s claimed intelligence.
The MIIT draft addresses that problem through a framework for terminology, interfaces, evaluation methods, and application requirements. Xinhua says it calls for common methods to grade robot capabilities and for compatible formats covering machine IDs, coding conventions, and recognition systems. These details may sound administrative, but they matter when a manufacturer, a software provider, a factory, an insurer, and a regulator need to understand whether the same system can be deployed safely.
The proposal also distinguishes between a robot’s technical components and its operating context. For industrial use, the draft would cover operational abilities and environmental adaptability in structured and semi-structured settings. That emphasis reflects a practical reality: factory floors involve dust, reflections, heat, shifts in lighting, moving equipment, and continuous operation. A robot that performs a polished sequence at an event has not necessarily met those conditions.
Standards Could Turn Conference Momentum Into Procurement Rules
The timing follows a crowded season of Chinese robot announcements. The Beijing conference that closed with 44 new embodied-AI technologies showed the breadth of current competition, from bodies and hands to world models and sensing systems. The draft standards offer a way to connect that product energy to procurement, testing, and safety expectations.
This does not mean that a standard system will choose winners. It may, however, affect which companies can qualify for industrial pilots, public-sector projects, supply-chain partnerships, or longer-term deployment contracts. Businesses with good interfaces, documented test results, service procedures, and safety controls could gain an advantage over competitors that rely on narrow demonstrations.
The draft’s references to ethics and social-impact assessment are also significant. Humanoids will operate close to people in workplaces, public areas, and potentially homes. Safety cannot be treated only as a question of whether a robot falls down. It includes how systems respond to people, what data they collect, how faults are recorded, who can intervene, and how responsibility is assigned after an incident.
The 2028 Goal Measures China’s Deployment Ambition
Xinhua reported that standard-setting work is already underway across China and other markets, with more than 100 related projects in development or early research. The draft’s 2028 target is therefore not a starting point from zero. It is an attempt to coordinate work that has already begun and direct it toward a common national structure.
That effort arrives as companies seek to turn technical gains into actual use. The recent Tiangong Ultra story underscored the intensity of China’s performance competition. Standards will ask a different question: can a robot’s capability be measured consistently enough for a customer to trust it in a defined task?
The answer will not be settled by one document. Standards must be updated as sensors, models, batteries, actuators, and safety practices improve. The draft recognizes that need by calling for a dynamic system rather than a fixed checklist. Standards can also help buyers distinguish a robot’s own controls from achievements dependent on carefully arranged surroundings, remote supervision, or scripted sequences. That transparency could make pilot results more comparable across suppliers and industries.
For now, the proposal signals that China sees the next stage of humanoid development as industrial infrastructure, not just product launch cycles. The companies that can meet common tests, explain their limitations, and operate safely in real settings will be better positioned when interest turns into orders.