Beijing Sets Up a State Data Giant to Break Up Its Data Islands

China has launched a central state-owned data company to support AI and economic growth, giving Beijing a national operator for a resource still fragmented across organisations and regions. ^[2][4]

China National Data Group was inaugurated on October 20, 2026, in Xiongan New Area, where it is headquartered. Its mandate covers developing and using data resources, connecting data infrastructure and advancing data technology. Vice-Premier Ding Xuexiang attended the launch and called on government departments, local authorities and enterprises to cooperate. ^[2]

That request goes to the heart of the undertaking. Beijing can establish a company by administrative decision. Making data accessible, compatible and worth using requires agreements among organisations with different responsibilities and interests. Experts cited by the South China Morning Post warned that entrenched barriers to sharing could undermine the new group. ^[1]

The group's best chance is to make sharing worthwhile and safe for data holders, while producing something customers want. Central status gives it a stronger position from which to negotiate. It does not, by itself, turn records held by separate organisations into a useful national resource.

A national operator meets local interests

The group enters an institutional structure Beijing has been building for several years. China's current data-economy framework draws on the 2022 Twenty Data Measures and the establishment of the National Data Administration in 2023, according to a February analysis by Ran Guo of the Asia Society Policy Institute. That effort seeks to integrate fragmented markets across jurisdictions and sectors, often described as data islands. ^[4]

The intended division of labour is fairly clear. According to Zhang Xianghong, a member of a national advisory committee on data, the administration develops foundational systems and policy, while the new group will concentrate on commercial development, use and operation of data resources. ^[3]

One institution sets the framework. The other must find ways to do business within it.

Zhang said the group could integrate national data resources, encourage sharing across industries and regions, and support a unified national data market. That is the ambition. It is also a description of the coordination problem the company inherits. ^[3]

Beijing is backing the group with large state-supported enterprises, including major telecommunications carriers and power-grid operators, according to the South China Morning Post. Their involvement offers a potential starting point for cooperation across large organisations. ^[1]

For a data holder, sharing can mean surrendering some control over a commercially valuable resource or accepting responsibility for how sensitive information is used. The group's task is therefore partly a negotiation over benefits and risks. Who gains from access? Who pays to prepare the information? Who answers when its use goes wrong?

These are questions the business model must settle. A national mandate may bring organisations to the table, but a workable agreement has to give them reasons to stay there.

Access is only the first hurdle

Even successful negotiations would leave a second job: making information from different organisations work together. Guo's analysis identified uneven data quality and limited interoperability among China's domestic data markets as persistent obstacles. ^[4]

Interoperability means that systems can exchange information and use it meaningfully. Two organisations can agree to share records without those records describing the same things in the same way. Access to both does not automatically produce a reliable combined dataset.

That is why the government's infrastructure mandate matters. Ding urged the group to improve data services, develop infrastructure and make breakthroughs in key technologies. ^[2] Connections are necessary, but the commercial value comes from what can pass through them in a usable form.

For AI applications, this places the emphasis on preparation and permitted use. A large quantity of accessible information can still be unsuitable for a particular task. The operator has to establish what the data describes, how dependable it is and what users are allowed to do with it. Otherwise, it is selling access to a problem someone else must solve.

China's push therefore needs a more demanding measure of progress than the number of institutions participating. A useful service must let a customer accomplish something that was previously difficult, costly or impossible. An agreement creates the opportunity. A working product establishes the value.

The group's launch announcements establish its mandate, not operating results. ^[2][3] Guo's February analysis predates the company and assesses the wider policy framework. ^[4]

Somebody still has to want the product

There is a less glamorous problem than persuading organisations to release data: finding demand for it. Guo identified weak demand for data products among the challenges facing China's data-market effort. ^[4]

That finding changes how the new group's success should be judged. If the main obstacle were simply restricted supply, securing access would do much of the work. Weak demand means the operator must also discover which information has a customer, what form that customer needs and whether the service is worth its price.

A dataset's value depends on the job it helps someone do. The cost of collecting or holding it is no guarantee that another organisation will pay to use it. Nor does labelling data an economic resource create a market by itself.

The strongest case for a central operator is that it could reduce the work required on both sides. Instead of each user negotiating separately with several holders, the group could arrange access and offer services that are easier to buy and use. Zhang's assessment of its potential to integrate resources and encourage cross-sector sharing supports that possibility. ^[3]

But aggregation creates its own risk. Guo's analysis identified possible monopolistic tendencies in data franchising, arrangements that give an operator rights to develop and sell data products. ^[4] If those rights are exclusive, users may have only one supplier to bargain with, leaving that supplier considerable power over prices and access conditions.

The commercial test must therefore include access for users, not merely access for the group. Moving information into the orbit of a powerful intermediary would accomplish little if potential customers still faced restrictive terms or products they could not use.

Trust has to be part of the service

Security and privacy are built into the policy setting. Guo described China's post-2022 approach as combining economic-development goals with oversight by the Cyberspace Administration of China and other agencies, while identifying unresolved privacy and security risks. ^[4]

That makes unrestricted sharing an unsuitable goal. Some information requires controlled access, and some uses should face tighter limits than others. A credible operator needs rules that distinguish between them.

Safeguards also serve the commercial argument. If data holders cannot control permitted uses or understand their responsibilities, withholding information may remain their safest choice. Clear access conditions can make cooperation possible by reducing that uncertainty.

Ding linked the group's establishment to opportunities in AI, the construction of Digital China and new national competitive advantages. ^[2] Those are large objectives for an enterprise whose first practical challenge is getting other organisations to work with it.

Its first commercial offerings will reveal whether Beijing has made data easier to use or merely put a stronger company in charge of access.