In May 2026 the National Data Administration issued the 2026 Digital Economy Work Points, organising the year’s digital economy agenda around eight areas of priority tasks. For companies, two lines stand out: deepening the market-based allocation of data elements, and strengthening data’s empowerment of AI development.
The eight tasks and two main lines
The Work Points cover the main fields of the digital economy across eight priority tasks. Two of them — “deepening the market-based allocation of data elements” and “strengthening data’s empowerment of AI development” — sit on the “circulation” and “supply” sides of data respectively, and are the key to reading the document.
The first answers how data becomes an element that can be owned, traded and priced; the second answers how data turns into the capability to train models and solve industry problems. Together they point to one goal: making data actually flow and be used.
What a nationally unified data property rights registration system means
On the market side, the Work Points call for accelerating the establishment of a nationally unified data property rights registration system, issuing registration guidelines, and promoting a price-formation mechanism for the authorised operation of public data.
Registration is the precondition for determining rights. Data has long faced a “hard-to-own” problem — unclear ownership makes trading and circulation impossible. A unified registration system and guidelines aim to give data a verifiable proof of rights. For companies this means data moves from “usable internally” toward “manageable as an asset”, and inventorying, registering and keeping ledgers of data resources will become routine. The price-formation mechanism for public data authorised operation also matters: it affects whether companies can obtain public data on market terms and the cost of bringing public data into industrial use. For our earlier review of the parallel datasets and computing-network deployment, see the National Data Administration’s datasets and integrated computing network.
Data for AI: the six special actions
On the AI side, the Work Points call for six special actions — strengthening the base and expanding capacity, enabling applications, improving quality and efficiency, managing and servicing, releasing value, and attacking annotation — to form a set of benchmark high-quality datasets that meet “AI readiness” requirements, effectively train advanced models and genuinely solve industry problems.
These six actions are consistent with the National Data Administration’s June 2026 implementation plan for building high-quality industrial datasets. “AI readiness” is the key phrase: a dataset must not only exist but be trainable, licensable and traceable. Beyond scale, quality and compliance become hard measures of a dataset’s value.
How the two lines interlock, and the compliance prerequisites
Market-based allocation and AI data supply are not separate. Rights determination and registration answer “who owns the data and whether it can circulate”; dataset building answers “whether the data can be used for training”. If the first is unclear, the second cannot proceed — a dataset of ambiguous ownership, however large, cannot safely enter model training.
So when companies take part in the data-element market or dataset building, several compliance prerequisites are unavoidable: lawful sourcing, the boundaries of personal information processing, data security and classification, and compliance at the annotation stage. These prerequisites bear directly on circulation arrangements, as we discuss in our pieces on the token economy and data-element compliance and AI data-compliance risks. Rather than remediating after circulation stalls, companies are better off building a source ledger, authorisation chain and compliance assessment at the collection and registration stage. For support on data asset registration, dataset-building compliance and data circulation arrangements, see our Data Compliance service.
This article is based on public information from the National Data Administration’s 2026 Digital Economy Work Points, issued in May 2026. It is for reference only and does not constitute legal advice.