China's National Data Administration Promotes Token‑Based Data Trading

The National Data Administration (NDA) issued a plan to build over 120,000 high‑quality data sets by June 2026, totaling more than 1,565 PB. At a July 28 press conference, Deputy Director Yu Ying announced that the NDA will encourage commercial models based on data tokens, exploring token‑trading mechanisms to foster a consensus for paying for premium data. The agency aims to integrate token applications with industrial, medical and financial sectors to scale AI deployments and create a “data‑flywheel” linking model use with data supply.

The NDA also led the drafting of five national standards on data‑set construction, format, classification and quality testing, applying a combined “data quality check + benchmark model verification” method. By June end, the NDA’s Data Set Management Service System had certified 578 supply‑demand entities and listed 1,433 data sets, facilitating orderly matchmaking and accelerating the data‑exchange ecosystem.