How AI and Satellite Images Created China’s First Nationwide Map of Solar and Wind Power

This approach differs from traditional energy statistics, which often depend on reporting from regional authorities or estimates derived from installed capacity. Instead, the system identifies infrastructure directly from imagery, producing geolocated data points for individual installations.

Handling the scale of the imagery was a major technical challenge. China’s diverse terrain—from deserts to dense urban regions—means renewable facilities appear in many different contexts, requiring models that can recognize installations across varying landscapes and lighting conditions.

What the Nationwide Renewable Map Reveals

The completed inventory provides a detailed spatial snapshot of China’s renewable infrastructure. According to reported results, the AI system detected:

• About 319,000 solar photovoltaic installations
• Around 91,600 wind turbines
• Facilities distributed across 1,915 counties nationwide

Because each asset is geolocated, the dataset reveals regional patterns of renewable deployment and provides a foundation for energy‑system analysis.

Previous national datasets often relied on lower‑resolution satellite imagery or partial reporting, making it difficult to precisely track where installations were built. Remote‑sensing studies have increasingly used machine learning to map solar and wind infrastructure, but comprehensive high‑resolution inventories remain relatively rare.

Why a “God’s‑Eye View” Matters for Grid Planning

Wind and solar power are inherently variable. Their output fluctuates with weather conditions and time of day, which can lead to curtailment—when renewable electricity cannot be used or transmitted and is therefore wasted.

A nationwide, geolocated map of installations helps researchers analyze how different renewable sources interact across regions. Using the new dataset, the team studied solar‑wind complementarity, finding that wind and solar generation often peak at different times—for example, solar producing more during the day while wind can remain stronger at night.

Understanding those patterns allows planners to explore strategies such as:

• Coordinating renewable generation across provinces
• Expanding transmission links between regions
• Adding energy storage where renewable output fluctuates

Researchers report that wider regional coordination of wind and solar resources could significantly increase the amount of renewable electricity that can be absorbed by the grid.

Linking Renewable Energy to the AI Data‑Center Boom

The mapping project also has implications beyond traditional grid planning. China’s rapidly growing data‑center and AI computing infrastructure is driving large increases in electricity demand.

To address that demand while pursuing decarbonization goals, the country has begun linking renewable generation more directly to computing infrastructure. One example is a project in Ningxia that supplies large‑scale photovoltaic power directly to data‑center operations.

A comprehensive map of wind and solar assets could help policymakers decide:

• Where new data centers should be located relative to renewable resources
• Where transmission or storage infrastructure is needed
• How renewable power can be coordinated to support energy‑intensive computing workloads

In that sense, the AI‑generated inventory does more than catalogue infrastructure—it creates a data foundation for managing the next phase of China’s energy transition.

A New Way to Track Renewable Infrastructure

As renewable energy systems expand worldwide, the ability to monitor them accurately becomes increasingly important. Satellite imagery combined with AI offers a scalable way to track infrastructure across vast areas and update inventories as new facilities are built.

For China—home to the world’s largest renewable‑energy system—the new nationwide map provides something policymakers and researchers have long lacked: a precise, continuously updatable picture of where solar and wind power actually exist on the ground.