• China's MIIT has launched a national pilot initiative to build high-quality industrial datasets for AI across key manufacturing sectors, targeting data infrastructure completion by 2026.
• The program follows a "1+4+N" architecture featuring a trusted data interconnection platform, four core resource systems, and multiple AI applications across manufacturing operations.
• Beyond AI training, China is developing a complete industrial data economy including data rights, value assessment, trading mechanisms, and quality evaluation systems for datasets.

---

China’s Ministry of Industry and Information Technology (MIIT) has launched a national pilot initiative (opens in a new tab) to build high-quality industrial datasets for artificial intelligence across key manufacturing sectors. The program, called the Industrial Data Foundation Action, is designed to organize companies, platforms, manufacturing clusters, and SME digital-transformation pilot cities into consortia that can collect, standardize, share, and apply industrial data. The stated goal is to break through bottlenecks in how industrial data is gathered, connected, circulated, and used by the end of 2026. For U.S. industry and policymakers, the signal is clear: China is moving to build not just smarter factories, but the data infrastructure, standards, and governance systems needed to train industrial AI at scale.

Data Becomes Industrial Policy

MIIT’s notice makes clear that Beijing sees industrial data as strategic infrastructure. The initiative targets manufacturing sectors with strong digital foundations and high data-value potential, naming steel and automotive as examples while leaving room for broader sector inclusion.

Four types of entities are eligible to lead pilot consortia:

• major industry companies
• platform institutions such as industrial internet platforms and big data centers
• advanced manufacturing clusters
• cities piloting SME digital transformation

The program is not just about collecting data. It is about creating mechanisms for trusted interconnection, standards development, scenario-based deployment, and controlled data circulation.

The “1+4+N” Blueprint

At the center of the plan is a “1+4+N” architecture.

The “1” is a trusted industry data interconnection platform.

The “4” are four core resource systems:

• an industry data resource library
• a key data-technology research library
• an industrial data standards library
• a high-quality dataset library

The “N” refers to multiple downstream applications, including industry large models and industrial AI agents deployed across R&D, pilot testing, manufacturing, operations, and supply-chain collaboration.

The documents also call for industrial data production and training bases, simulation-data capture, professional labeling, and synthetic data generation for extreme scenarios.

More Than AI Training

Rare Earth Exchanges™ reviewed and translated MIIT documentation, revealing broader ambition. China is also exploring:

• data rights and value assessment
• revenue-sharing mechanisms
• data trading and productization
• quality evaluation systems for datasets
• open data zones and ecosystem building
• cultivation of industrial data service firms and hybrid data-industry talent

That is not a narrow AI project. It is an attempt to build an industrial data economy around manufacturing. This dovetails with stated plans to advance the economy through digitization and greenification (e.g., further moving away from petrol) across urban coordinators.

Why the West Should Pay Attention

This matters well beyond software. If China can standardize and operationalize industrial data at scale, it could strengthen its advantage in advanced manufacturing, robotics, materials processing, EV supply chains, and potentially rare earth and critical minerals processing.

The real story is not that China wants more AI. The real story is that China is trying to build the data rails, standards stack, and governance model that make industrial AI commercially usable.

Disclaimer: This report is based on Chinese government documents and state-linked distribution channels. The information should be independently verified before drawing investment, policy, or strategic conclusions.