Beijing wants software that can fuse data across the state, the military, and industry—but the institutions that hold that data do not trust anyone else to see it.

By Probe International

China wants its own version of Palantir.

It does not have one, and a recent Jamestown Foundation report argues that the obstacle is trust.

An American company founded in 2003 in Palo Alto, California, Palantir builds software that turns fragmented, messy data into something institutions can actually act on. Its main products are Gotham, used by governments and intelligence agencies, and Foundry, used by companies. The software does not collect the data. It sits on top of what the client already has, organizes it into a working model of the organization, and makes it usable across teams that normally cannot share it easily.

For a military client, that can mean combining intelligence reports, logistics, and targeting data into a single picture commanders use to plan and adjust operations. For investigators, it can mean connecting case files, financial trails, and identity records that would otherwise sit in different agencies. For a company, it can mean giving plant managers, suppliers, and executives one live view of production, inventory, and equipment so problems show up before they cascade.

Few governments have stated ambitions that align more closely with that capability than China’s. Beijing has designated data a “new factor of production,” invested heavily in surveillance and digital governance, and pursued military-civil fusion and industrial upgrading that depend on coordinating information across ministries, state-owned firms, local governments, and the security apparatus. A system that could reliably join those streams would sharpen both coercive capacity and technocratic planning. On paper, the demand is obvious.

In May 2026 it became explicit: the National Data Administration held a lecture titled “Viewing China’s Data Industry Through the Palantir Phenomenon.” In “Low Trust Stymies Emergence of a ‘Chinese Palantir,’” published in Jamestown’s China Brief on September 16, Aidan Powers-Riggs and Leon Li argue that the demand has not produced an equivalent firm. Chinese companies have built pieces of the stack. Utenet, Jing’an, Huaru, and MizarVision sell military-intelligence and simulation tools to the PLA, defense firms, and public-security departments. Transwarp markets a data-integration platform akin to Foundry. MiningLamp’s founder has cited Palantir’s ontologies and forward-deployed engineers as a model.

In January 2026, state-owned SDIC Intelligence unveiled an ontology-based system aimed at what it called “controllable action.” Huawei, Baidu, and Alibaba bundle data-fusion tools with their own clouds. None of them spans sensitive government, military, and commercial systems the way Palantir does. The binding constraint, the report argues, is not engineering talent or machine capacity. It is trust—or the lack of it—inside the institutions that would have to grant any such firm access.

Cybersecurity fears make organizations reluctant to open core databases to an outside vendor. Bureaucratic rivalries lead ministries and security services to treat data as institutional power and to resist pooling it. Political sensitivity about what the data reveal—about officials, localities, or the Party’s own performance—raises the cost of letting any single company see across silos.

National Data Administration director Liu Liehong has described an “impossible triangle” in which safety, compliance, and efficient circulation cannot be reconciled. The result is a low-trust environment in which broad, cross-institutional integration remains risky for both the customer and the vendor.

Palantir’s model depends on customers willing to hand a private firm unusually wide visibility. That precondition has not held in China.