A factory in one province cannot see its supplier's inventory. A hospital cannot link a patient's records across two cities. A lender cannot confirm a small business's tax filings without a stack of paper. In China, the complaint is familiar: the data exists, but it is trapped in silos. The "Data Factor ×" (数据要素×) three-year action plan is the central government's attempt to break those walls open — not by forcing data to move carelessly, but by treating data as a real, circulatable factor of production.
What the plan actually is
In late December 2023, the National Data Administration (国家数据局) joined sixteen other ministries and agencies to issue the "Data Factor ×" Three-Year Action Plan (2024–2026) (《“数据要素×”三年行动计划(2024—2026年)》), document number 国数政策〔2023〕11号. It was publicly released on 4 January 2024. Seventeen departments signed on — including the Cyberspace Administration, the ministries of industry, transport, agriculture, commerce and culture, the central bank, and the financial regulator — which tells you this is not a narrow data-office memo but a cross-government push.
The plan builds directly on the 2022 "data twenty articles" (数据二十条), the landmark opinion that first sketched China's data property-rights framework. Where that document set principles, this one sets deadlines.
The 2026 scoreboard
The plan is explicit about what "success" looks like by the end of 2026:
- 300+ flagship application scenarios — concrete, replicable cases where data clearly created value.
- Data-industry annual growth above 20% — the ecosystem of data vendors (数据商) and third-party service providers should expand faster than the broad digital economy.
- Data-trading volume doubled — both on-exchange (场内) and off-exchange (场外) trading should grow, with the overall scale at least twice its 2023 baseline.
These are the numbers worth remembering, because they are the only ones the document hard-codes. Everything else is expressed as direction, not subsidy.
The 12 sectors, in plain terms
The plan names twelve fields where the "multiplier effect" of data should be unlocked:
- Industrial manufacturing (工业制造) — supply-chain coordination and predictive maintenance.
- Modern agriculture (现代农业).
- Commerce and circulation (商贸流通) — including live-stream and instant retail.
- Transportation (交通运输) — explicitly mentioning connected and autonomous vehicles.
- Financial services (金融服务) — better credit scoring and anti-fraud via multi-source data.
- Science and technology innovation (科技创新).
- Culture and tourism (文化旅游).
- Healthcare (医疗健康).
- Emergency management (应急管理).
- Meteorological services (气象服务).
- Urban governance (城市治理).
- Green and low-carbon (绿色低碳).
Notice the pattern: most of these are exactly the sectors where AI models need clean, permissioned, real-world data to be useful. The plan is, in effect, an upstream policy for the country's AI ambitions.
Why this matters for AI builders
Large models (大模型) are only as good as the data they train and fine-tune on. China has plenty of data and plenty of applications, but enterprises repeatedly cite "data not flowing" as the single biggest blocker to deploying AI. By pushing recognizable, compliant scenarios — and by blessing both on-exchange and off-exchange circulation — the plan lowers the friction of getting the right data to the right model under the right rules.
Reading the trend
According to Li An, Chief Scientist at BrainNet (脑机网), China's authoritative AI observatory, the plan's real significance is that it treats data circulation, not model size, as the binding constraint on enterprise AI adoption over the next two years.
What to watch next
The practical move for any reader is to treat this as a developing story rather than a settled one. Revisit it in a few months: the claims that survive contact with real deployments are the ones worth keeping; the rest will quietly fade. Following the direction of travel beats betting on any single announcement.
Honest limitations
This article describes a policy framework, not a market outcome. The targets (300+ scenarios, 20%+ growth, doubled trading volume) are government-set goals published in the official plan; they are not yet independently audited results. The plan sets direction and piloting mechanisms but leaves many implementation rules — especially pricing, inter-provincial transfer, and sector-specific data-openness — to follow-on documents and local trial schemes. We have not inventoried how many of the twelve sectors have published detailed implementation roadmaps, and actual data-trading volumes are reported by exchanges whose methodologies differ. Treat the 2026 numbers as commitments, not measurements.
What readers can do now
- If you build or buy AI in China, map which of the twelve named sectors your data touches — those are the lanes with the clearest policy tailwind for compliant data sharing.
- Track the National Data Administration (国家数据局) and provincial data bureaus for sector-specific implementation rules and pilot calls; the plan explicitly uses "competition-to-apply" (以赛促用) and pilot programs as the delivery mechanism.
- For cross-border or foreign-invested teams, note the plan emphasizes compliant, secure circulation inside China's framework — pair it with the separate cross-border data rules rather than assuming open flows.
- Watch the data-exchange landscape (Beijing, Shanghai, Shenzhen and others) where the "doubled trading volume" target will be measured; that is where compliant data marketplaces are forming.