National Data Resource Survey Report (2025) released: The total annual data production reached 52.26 zettabytes in 2025, a year-on-year increase of 27.28%.

On April 29th, the "National Data Resource Survey Report (2025)" was officially released at the 9th Digital China Construction Summit. In 2025, the national annual data production reached 52.26 zettabytes (ZB), a year-on-year increase of 27.28%, with a growth rate 2.28 percentage points higher than the previous year. Our country's data production accounts for approximately 27.44% of the global total. In terms of entities, enterprises have become the main force in data production, with about 90% of the data production increment coming from enterprise data, demonstrating significant results in the transformation of industry intelligence. In terms of industries and sectors, the data production volume of industries such as industrial manufacturing, transportation logistics, software, and information technology services has increased significantly, with year-on-year increases of 1.27 zettabytes (ZB), 1.22 zettabytes (ZB), and 0.92 zettabytes (ZB) respectively, effectively playing a role in leading and supporting development. Emerging fields such as embodied intelligence and low-altitude economy have seen a rapid increase in data production, with growth rates of 477.78% and 75% respectively. The original text below: "National Data Resource Survey Report (2025)" Officially Released On April 29th, the "National Data Resource Survey Report (2025)" was officially released at the 9th Digital China Construction Summit. The survey results show that the market value and commercialization of data elements in China have significantly accelerated. The data resource supply system, data circulation system, and data development and utilization system are evolving in a coordinated manner, shifting from the expansion of data resource scale to the release of data element value, presenting the following significant characteristics: First, the scale of data resources maintains a high growth trend. In 2025, the national annual data production reached 52.26 zettabytes (ZB), a year-on-year increase of 27.28%, with a growth rate 2.28 percentage points higher than the previous year. Our country's data production accounts for approximately 27.44% of the global total. In terms of entities, enterprises have become the main force in data production, with about 90% of the data production increment coming from enterprise data, demonstrating significant results in the transformation of industry intelligence. In terms of industries and sectors, the data production volume of industries such as industrial manufacturing, transportation logistics, software, and information technology services has increased significantly, with year-on-year increases of 1.27 zettabytes (ZB), 1.22 zettabytes (ZB), and 0.92 zettabytes (ZB) respectively, effectively playing a role in leading and supporting development. Emerging fields such as embodied intelligence and low-altitude economy have seen a rapid increase in data production, with growth rates of 477.78% and 75%. The total national data storage reached 2.53 zettabytes (ZB), a year-on-year increase of 21.05%. The structured data storage volume was 0.56 zettabytes (ZB), a year-on-year increase of 43.59%, accounting for 22.13% of the total data storage volume. Data quality continues to improve, accelerating transformation into forms that are computable and analyzable. Second, the construction of computing power infrastructure is steadily advancing. The "Eastern Calculation and Western Algorithm" project is fully implemented, the national integrated computing power network is accelerating its construction, and the level of intelligent computing power supply is steadily increasing. By the end of 2025, the national intelligent computing power scale reached 1.59 million petaflops (FP16), with general computing power being replaced by intelligent computing power, becoming a key infrastructure supporting the development of artificial intelligence. The concentration of intelligent computing resources is gradually emerging, with more than 80% of the national intelligent computing scale located in eight national computing hubs (including ten major clusters). Third, data resource development and utilization are becoming more efficient. With the deepening of actions such as the "Data Element " initiative, the demonstration scenario construction of public data, the improvement of data efficiency in state-owned enterprises, the leading trial of national data infrastructure construction, and the development plan of a trusted data space, the integration of data application and scenario construction continues to deepen, and the value of data elements is accelerating. The development and utilization of public data resources have shown significant effects. The amount of data used for sharing, open access, and authorized operation has rapidly increased, with the number of data sets applied for sharing growing by nearly 30% year-on-year, and the amount of open data and authorized operational data increasing by 31.71% and 53.96% respectively. The development and utilization of public data resources are driving the faster integration and application of data in various industries, covering scenarios in industrial development, education technology, medical health, public services, grassroots governance, and other fields. The vitality of enterprise data is being accelerated. In 2025, enterprise investment in data technology increased by 17.37% year-on-year. The number of data products and services offered by enterprises increased by 29.29% year-on-year, with transaction volumes increasing by 39.8%, transitioning from digital products to engines driving business growth. Fourth, the effects of data circulation and transactions are beginning to show. The accelerated construction of a national integrated data market has further stimulated market vitality, accelerated the release of data element value, and fostered a market consensus towards paying for high-quality data. Data circulation activity continues to increase. In 2025, the total volume of cross-border data flow in the country was 142.34 exabytes (EB), a year-on-year increase of 14.88%. The total volume of inter-provincial data flow was 2949.12 exabytes (EB), a year-on-year increase of 19.01%. Guangdong, Zhejiang, Jiangsu, Shandong, Henan and other economically developed provinces ranked in the top five for inter-provincial data flow. The total volume of enterprise data circulation reached 1935.36 exabytes (EB), a year-on-year increase of 25.17%. The average data circulation scale of leading platform enterprises and central enterprises was more than 140 times that of other enterprises, and the role of data circulation hubs continued to strengthen. The willingness to pay for data continues to increase. Among the sample enterprises, 11.65% have purchased data, with the cost of purchasing data increasing by 22.36% year-on-year. The average cost of purchasing data for leading platform enterprises and central enterprises was 60 times that of other enterprises. Industries such as finance, software, and information technology services had enterprise data purchasing rates exceeding 30%, significantly higher than the industry average. Fifth, data empowering artificial intelligence has entered a new stage of scaled application. The development of artificial intelligence has shifted from generic large models to industry-specific models, and further to the evolution of proxy artificial intelligence forms, with data demands expanding from foundational linguistic data to high-quality industry data sets. The survey indicates that in 2025, the total data used for artificial intelligence training and inference was 199.48 exabytes (EB), a year-on-year increase of 42.86%, with the volume of inference data reaching 101.34 exabytes (EB), surpassing training data for the first time. The quantity of high-quality data sets exceeded 110,000, with a scale exceeding 908 petabytes (PB), increasing by 61.13% and 142.58% respectively year-on-year. Approximately 21.1 trillion word tokens were called throughout the year, with word tokens emerging as a new measure for artificial intelligence. 2026 marks the beginning of the "14th Five-Year Plan" and the "Year of Releasing Data Element Value." Looking ahead, China's advantage in data resource scale will accelerate its transformation into value advantages, with data elements playing a more crucial role in empowering economic and social development. This article is selected from the National Bureau of Statistics. Editor: Chen Wenfang.