中国国家数据局近日召开新闻发布会宣布了数据基础设施建设的最新进展。[1]数据基础设施接入主体已达4万余个，较2025年底增长10倍；上架产品6万余个，增长4倍。[1]国家公共数据资源登记平台登记资源近30万项。[1]
在算力和数据集方面取得重要突破。[1]目前已有140万PFLOPS智能算力接入全国一体化算力网试验验证平台。[1]截至2026年6月，全国已建成高质量数据集超12万个，总体量超过1565拍字节，其中140家先行先试单位建成高质量数据集超106PB，72家行业链主单位建成高质量数据集超45PB，覆盖高价值场景超759个。[1]这些数据基础设施成果已成为人工智能快速发展的关键支撑。[1]
China's National Data Bureau announced a dramatic expansion of its data infrastructure ecosystem, revealing that the number of connected entities has reached over 40,000, a tenfold increase from the end of 2025 [1]. The number of listed products has also grown substantially to over 60,000, representing a fourfold rise during the same period [1].
The rapid scaling reflects accelerating progress in building foundational data systems. Intelligent computing capacity of 1.4 million PFLOPS has been integrated into the nationwide unified computing network experimental validation platform [1]. Over 120,000 high-quality datasets have been developed across the country, with a combined volume exceeding 1,565 petabytes, positioning data infrastructure as a critical enabler for artificial intelligence advancement [1]. The national public data resource registration platform now documents nearly 300,000 resources [1].
Progress extends across multiple sectors. More than 140 pilot units have constructed high-quality datasets totaling over 106 petabytes [1], while 72 industry-leading organizations have developed datasets exceeding 45 petabytes, covering over 759 high-value application scenarios [1]. The National Data Bureau has designated 2026 as the "year of data element value release," signaling intensified efforts to unlock the economic potential of data assets [1].