China's land registry, seabed maps and mineral exploration records are about to get a shared digital brain, if a new policy document delivers what it promises.
In November, the Ministry of Natural Resources and the National Development and Reform Commission jointly issued the Implementation Opinions on "AI + Natural Resources," a sector-specific blueprint for applying artificial intelligence across one of the state's most data-heavy domains.
The two milestone years are 2028 and 2030. Neither has happened yet, and the distinction between a plan and a result is the right lens for reading this document.
What the Document Is
The opinions were announced on November 10, with the ministry publishing details on November 12, and coverage followed in Economic Daily and Guangming Daily. As an implementation opinion (实施意见), it is a policy directive issued under the State Council's "AI+" action framework, directing how a specific ministry should apply artificial intelligence, rather than a law or a regulation binding private parties.
The ministry's rationale leans on an asset few other sectors have: natural resources agencies hold enormous, long-accumulated datasets covering land use, spatial planning, oceans, geology, minerals, forests and grasslands, plus a dense web of real application scenarios.
The 2028 Milestone: Demonstration, Not Transformation
By 2028, the document targets demonstrated application of AI in the natural resources sector. Specifically, it calls for building high-quality datasets, large models and AI agents (智能体) in five priority directions: spatiotemporal geographic information, territorial spatial planning and governance, oceans, geology and minerals, and forestry and grassland.
It also targets a comprehensive upgrade of the intelligence level of the "One Map" (一张图), the unified digital map that pools land, forest, grassland, wetland and marine data and on which government departments at various levels are expected to work collaboratively. Replicable, promotable typical application scenarios are part of the 2028 goal.
By 2030, the ambition widens: AI deeply integrated across the whole chain of natural resources management, an intelligent territorial spatial governance platform, secure and controllable key technologies and smart equipment, and an overall digital-intelligent governance capability that the document describes as internationally leading. That last phrase is a policy aspiration, not a measured claim.
Five Pilot-Testing Platforms
The most concrete mechanism in the document is a plan to lay out five AI pilot-testing platforms (中试平台), one per priority direction, designed as shared infrastructure combining computing supply, model training, algorithm iteration, scenario validation and ecosystem building.
Pilot testing, the stage between lab prototype and industrial deployment, is where many AI applications stall. By centralizing it per domain, the ministry is effectively building the intermediate infrastructure it expects industry and research institutes to use, and Economic Daily's coverage details what each platform focuses on, from AI-produced mineral exploration models to forest-fire prevention.
Eight Demonstration Applications
The document names eight key application areas:
- AI + integrated "sky-space-ground-sea-network" monitoring and regulation
- AI + territorial spatial planning and governance
- AI + natural resource asset management
- AI + natural resource disaster governance
- AI + ocean resource development and governance
- AI + geological and mineral exploration and development
- AI + forestry and grassland protection and utilization
- AI + earth science
The list spans operational needs, from disaster response to asset valuation, and scientific ones, including geoscience research. Each is meant to produce typical, replicable application results.
Reading It Correctly
Three cautions apply when interpreting this document.
First, stage. This is an issued implementation opinion with dated targets; the 2028 datasets, models and agents, and the 2030 integration goals, are objectives. Nothing in the document reports them as achieved.
Second, the phrase "internationally leading" by 2030 is a policy expression of intent. Whether governance capability can be benchmarked internationally, and by whom, the document does not specify.
Third, implementation depends on money and coordination that the document only gestures at. It calls for strengthening inter-ministry and province-level coordination, opening application scenarios, using central fiscal funds to lever social capital into technology development, scenario demonstration and commercialization, and imposing full-lifecycle supervision over data compliance and AI safety.
For international readers, the document is a signal of where a large state data holder intends to apply AI first: mapping, planning, disaster governance and resource exploration, all domains where the state is both the data owner and the primary customer.
Honest limitations
- This article summarizes the document as reported by the ministry's own site, Economic Daily and state broadcasters; the full text's detailed annexes, if any, were not independently reviewed.
- The five-platform and eight-application plans have no published budgets, timelines per platform or responsible implementing bodies in the available coverage.
- No independent assessment exists of current AI maturity in China's natural resources agencies, so the distance from today's state to the 2028 target cannot be quantified.
What readers can do now
- Watch ministry procurement announcements for the five pilot-testing platforms; tenders will reveal actual budgets and lead implementers.
- Track which of the eight application areas produces the first published, replicable case; that indicates where data readiness is highest.
- If you work in geospatial AI or mining technology, monitor the ministry's scenario-opening measures, which may define partnership or procurement channels for outside vendors.
