China unveils AI-powered satellite toolbox to help developing countries' weather forecasts

China launched an artificial intelligence-powered weather satellite toolbox on Friday, aiming to help developing countries turn satellite observations into actionable weather forecasts and early warnings.

The MAZU-FengYun Satellite AI Box was unveiled by the China Meteorological Administration during the meteorological forum of the 2026 World AI Conference and High-Level Meeting on Global AI Governance in Shanghai.

The launch marks a new stage in MAZU's intelligent early-warning initiative, which was unveiled last year, shifting from providing shared meteorological products to delivering AI-enabled forecasting capabilities, according to the administration.

The toolbox combines observations from China's Fengyun meteorological satellites with artificial intelligence to provide an integrated satellite application platform for international users.

Built on the Fengyun satellite system's global observation network, the AI toolbox integrates satellite data reception, AI inference and operational forecasting into a platform embodied with both hardware and software, Wang Jinsong, director of the National Satellite Meteorological Center, said.

The system can be rapidly deployed to meet the operational needs of different countries and regions, supporting application scenarios including internet users, Fengyun direct-broadcast receiving stations, users of the CMA's satellite broadcasting system, known as CMACast, and emergency field operations.

Wang said the toolbox addresses a longstanding challenge faced by many developing countries that can access satellite observations but lack reliable internet connectivity, computing capacity or technical expertise needed to convert raw data into operational weather services.

The AI models are deployed directly onto local devices through edge computing technology, allowing satellite observations, forecast products and local weather observations to be processed on-site.

The localized processing enables users to generate meteorological products tailored to local conditions while reducing dependence on external computing infrastructure, officials said.

"The toolbox provides not only a technical product but also a practical and adaptable application capability," Wang said.

He added that the approach allows advanced meteorological technologies to reach frontline forecasters and disaster-response agencies more efficiently, improving services for disaster prevention.