China's daily average token usage increased by 40% over three months
In March this year, China's average daily token calls exceeded 140 trillion, an increase of more than 1,000 times from 100 billion at the beginning of 2024, and an increase of more than 40% in three months compared to 100 trillion at the end of 2025.
Liu Liehong, director of the National Data Administration, released the above data at the press conference of the State Council Information Office on the 24th. Mormones are the smallest information units for large models to process information, which has the characteristics of measurable, pricable, and tradable in the intelligent era. At present, a new value system is accelerating the evolution and formation of a new value system around the calling, distribution and settlement of tokens, and has become an important path for the commercialization of the artificial intelligence industry.
Liu Liehong said that the large increase in the average daily token call volume fully shows that China's artificial intelligence development has entered a stage of rapid growth and the industrial competitiveness has been significantly enhanced. It also marks a significant increase in the supply of data sets, the continuous release of the value of data elements, and the stage of benign interaction in which data elements empower the innovation and development of artificial intelligence.
This year's government work report proposes to create a new form of intelligent economy and deepen and expand "artificial intelligence +". Deepen the development and utilization of data resources, improve the basic system of data elements, and build high-quality data sets.
Data is a key production factor in the smart economy. This year is the first year of the "15th Five-Year Plan", and it has also been designated as the "year of release of the value of data elements" by the National Data Administration. Liu Liehong said that he will make greater efforts to promote the marketization and value of data elements. Accelerate the establishment of a unified national data property rights registration system, and speed up the issuance of policy documents for building a national integrated data market.
High-quality datasets empower the development of artificial intelligence
By the end of 2025, more than 100,000 high-quality datasets have been built across the country, with a total volume of more than 890PB, which is equivalent to about 310 times the total digital resources of the National Library of China.
"Wherever artificial intelligence develops, we will build high-quality data sets. Where the artificial intelligence + action goes, the construction and promotion of high-quality data sets in the industry will go. Liu Liehong said.
In response to the problem of "small and scattered" construction of high-quality datasets, the National Data Administration, together with 26 departments, organized and selected 72 high-quality dataset construction chain master units, 140 pilot work units and 104 typical cases, and built a high-quality dataset construction ecology driven by chain owners, multi-party participation, joint research, joint construction and sharing, and win-win cooperation, and continued to promote the construction of high-quality datasets.
In order to promote the development of the data annotation industry, the National Data Bureau has laid out seven cities in Chengdu, Shenyang, Hefei, Changsha, Haikou, Baoding and Datong that undertake the task of pilot construction of data annotation, issued the "Implementation Opinions on Promoting the High-quality Development of the Data Annotation Industry", selected 47 excellent cases of data annotation, and guided the holding of 7 supply and demand docking meetings for data annotation.
Liu Liehong said that in the next step, we will continue to promote the innovation and development of data-enabled artificial intelligence, and cooperate with all parties to deeply implement a new round of high-quality dataset construction action plans, including six special actions of strengthening the foundation and expanding capacity, labeling key problems, improving quality and efficiency, application empowerment, management services, and value release.
Build a national integrated data market
Since its establishment more than two years ago, the National Data Bureau has explored and formed a "5+3+1" work system for the reform of market-oriented allocation of data elements. "5" refers to improving the basic data system, building and operating data infrastructure, cultivating scenarios and data fusion applications, building a national integrated data market, and expanding the data industry; "3" refers to consolidating the three foundations of core technology research in the data field, data standardization, and digital talent training; "1" refers to one key point of data-enabled artificial intelligence innovation and development.
"With policies, how to use them well is the key." Liu Liehong said that recently, the publicity and interpretation of the data property rights system has been increased. Combined with the scenario, help all parties clarify the specific configuration plan of data "holding, use, and management rights", so as to clarify the data property rights and allow all kinds of subjects to dare to supply data and use data with confidence.
Liu Liehong said that in the next step, we will accelerate the establishment of a unified national data property rights registration system, pay close attention to the introduction of policy documents for the construction of a national integrated data market, promote the construction of data basic systems and data infrastructure as a whole, solve the "impossible triangle" problem of data security, compliance and efficient circulation, and provide a strong guarantee for the release of the value of data elements.
Scenario cultivation is the key to promoting the marketization and value of data elements. Liu Liehong introduced that in the past two years, the National Data Bureau has joined forces with more than 20 departments to jointly promote the "data element ×" action, which has stimulated a strong atmosphere of "data use" in the whole society. The "Data Element ×" competition has attracted more than 40,000 teams and 220,000 participants, excavated 302 excellent projects, 417 typical cases and 760 subdivided application scenarios, providing practical examples for data empowerment of thousands of industries. In the next step, we will work with relevant departments to cultivate high-value scenarios and better play the multiplier effect of data elements.
Data elements empower new industrialization
Industry is an important scenario for the release of the value of data elements. At present, the rapid development of a new generation of information technology represented by artificial intelligence is deeply integrated into all aspects of the whole process of manufacturing, and like computing power algorithms, data has increasingly become an important production factor and valuable strategic resource for industry.
Wang Yanqing, director of the Information Technology Development Department of the Ministry of Industry and Information Technology, said at the press conference that how to achieve efficient data collection, large-scale aggregation and in-depth application has become a must-answer question for the digital and intelligent development of the manufacturing industry.
In recent years, the Ministry of Industry and Information Technology has vigorously promoted the integration of "two modernizations", implemented the industrial Internet innovation and development project, and accelerated the large-scale application of 5G technology in the industrial and manufacturing fields, and achieved some remarkable results. By the end of 2025, the numerical control rate of key processes of enterprises in key industries will reach 68.6%, and a number of industrial 5G special chips, modules and terminal products will be developed.
In March this year, the Ministry of Industry and Information Technology launched the industrial data foundation building action to carry out the construction of high-quality industry datasets empowered by artificial intelligence, aiming to break through the bottleneck of industrial data "collection", "collection" and "use".
Wang Yanqing introduced that this pilot has a significant feature, which is to empower the manufacturing industry with artificial intelligence, and it is expected that by the end of this year, the goal of "six batches" can be achieved, that is, to cultivate a number of industry data cooperation consortiums, gather a number of industry data resources, tackle a number of key technologies of a batch of data, develop a number of industrial data standards, and create a number of high-quality standardized and circulating industry data sets, and finally empower the application of a number of industry large models and industrial agents.
Wang Yanqing said that in order to do a good job in the first and first trial, we must continue to do a good job in three aspects: first, we must strengthen support and guarantee; second, it is necessary to strengthen policy guidance, promote the issuance of policy documents on data elements to empower new industrialization, and issue reference guidelines for the application of data elements in industrial scenarios; Third, it is necessary to cultivate a good ecology, accelerate the development of industrial data standards, develop and expand data service enterprises such as data consulting, data governance, and data annotation, support the holding of a number of technical seminars, supply and demand matchmaking meetings, etc., and at the same time strengthen and optimize the artificial intelligence open source community to create a highland for high-quality open source data resources to gather.