· Policy & regulation· 日本語で読む
China’s Ministry of Industry and Information Technology Issues Five Standards Related to Humanoid Robots
China’s Ministry of Industry and Information Technology has issued five standards related to humanoid robots. They cover capability assessment, training facilities, training data, electric-drive joints and dexterous hands. The capability assessment standard establishes six levels, from L0 to L5, and four evaluation dimensions.
What happened
humanoid.guide reported that China’s Ministry of Industry and Information Technology issued five standards related to humanoid robots on November 16, 2026. They cover capability assessment, training facilities, training data, integrated electric-drive joints and dexterous hands. According to the report, the standards took effect upon issuance.
Why it matters
The standards cover not only capability assessment but also training facilities and data, as well as joints and hands. This means they relate not just to comparing robot performance but also to development environments and component connectivity. humanoid.guide reported that their scope includes addressing inconsistent capability claims and mismatched equipment interfaces, among other issues, but actual compatibility and adoption require separate verification.
Analysis
One notable aspect of the five standards is that they cover not only capability assessment for completed robots but also training facilities and data, as well as electric-drive joints and dexterous hands. Placing standards for evaluation alongside those for development and training environments and components shows a scope of standardization that extends beyond the performance claims of individual robots. However, the issuance of standards does not mean that products from different companies are actually compatible or that they can be operated using common procedures.
The capability assessment includes levels from L0 to L5, as well as four dimensions: perception and cognition, decision-making and learning, execution performance, and coordination and interaction. The article’s descriptions of a capability profile, a basic level and an overall level indicate a design that differs from a system that represents capability using a single level alone. At least based on the report, the apparent intent is to distinguish multiple areas of capability rather than consolidate assessment results into a single ranking. The article provided, however, does not specify measurement methods for each dimension or the criteria for distinguishing levels.
The industrial issue is that robot testing and training, and component connectivity, are presented as standardization targets within the same framework rather than as separate challenges. How far the standards for joints and hands will address mismatched equipment interfaces, and how standards for training data will apply to data collection and management in practice, cannot be determined without reviewing the standards themselves and their implementation. Even if evaluation criteria are established, the extent to which comparability can be ensured when testing facilities or data acquisition conditions differ will require separate verification.
The next points to check are the official names and numbers of each of the five standards, assessment procedures, methods for confirming compliance, data handling, and specific requirements for component interfaces. The report says it is based on 36Kr, and the article text available here does not include the full standards or details of their implementation. The fact that the standards were issued and took effect does not establish their adoption by companies or interoperability in the field.
Sources
Translated from the Japanese original: /news/chuugoku-kougyou-jouhouka-bu-hitogata-robotto-kanren-no-5-ki-a366d7. The Japanese version is authoritative if they differ.