China’s AI Weather System Mazu Expands Across Global South

China’s AI-powered weather warning system, Mazu, is rapidly expanding across the Global South, with full deployments in seven countries and cloud-based testing by over 40 nations, marking a significant milestone in global climate resilience efforts. The system, developed by the China Meteorological Administration (CMA), uses advanced artificial intelligence to provide early warnings for floods, typhoons, heatwaves, droughts, and other extreme weather events.

Pan Jinjun, chief engineer at the CMA, described Mazu as “an international public good” that provides “tailored early warning services for developing nations,” according to the South China Morning Post. Speaking ahead of the World AI Conference (WAIC) in Shanghai, Pan said the system was the first national project responding to the UN Early Warnings for All initiative, which aims to ensure every person on Earth is protected by early warning systems by 2027.

What Is Mazu?

Mazu — an acronym for Multi-hazard Alert, Zero-gap and Universal, and named after the Chinese sea goddess — is not a single AI model but an integrated platform that combines Fengyun meteorological satellite data with ground-based observations and multiple AI forecasting models. Unlike conventional forecasting systems that rely on individual prediction models, Mazu integrates multiple AI models with meteorological observations and early warning tools on a single platform, allowing countries to adapt the system using their own weather data for localized forecasts.

As CGTN reported, the platform has been adapted to different climate conditions worldwide. In Mongolia, the customized version monitors snowstorms, strong winds, and dust events. Jordan’s version supports forecasts for floods, droughts, heatwaves, and cold spells, while Sri Lanka’s meteorological agency reports that locally trained AI models have improved forecast accuracy and strengthened disaster response capabilities.

Expanding Global Reach

At the WAIC 2026 opening ceremony on July 17, President Xi Jinping announced that China would promote Mazu’s deployment in 30 countries, a significant escalation from the current seven nations where the system is already operational on the ground: Pakistan, Djibouti, Ethiopia, Jordan, Mongolia, the Solomon Islands, and Sri Lanka. The public cloud version of Mazu, which offers quicker access to meteorological data, has already been tested by over 40 Global South countries.

Pakistan offers one of the most compelling examples of the platform’s impact. Chinese and Pakistani meteorological authorities began jointly developing a localized version of Mazu in 2023, tailoring the system to weather risks including monsoon rainfall, heavy precipitation, and glacial lake outburst floods. “Using this platform, we can make more accurate predictions of floods, droughts and other extreme weather events,” said Furrukh Bashir, head of the Research and Development Department at the Pakistan Meteorological Department. “We look forward to seeing more products like this that can benefit people around the world.”

Djibouti 2.0 and New Capabilities

A major highlight of the WAIC 2026 conference was the formal delivery of Mazu Djibouti 2.0, an upgraded version featuring 3-kilometer spatial resolution, a three-day forecast horizon, and a six-hour update cycle with a “one-click alert” function. Mohamed Ismail Nour, Director of Djibouti’s National Meteorological Agency, said the upgraded model “will greatly enhance our ability to respond to frequent disasters and is of great significance for disaster prevention and mitigation.”

Alongside the Djibouti upgrade, China and Thailand launched the world’s first bilateral laboratory dedicated to AI-powered meteorological forecasting at WAIC 2026. Built on the Mazu platform, the joint laboratory will develop AI technologies for forecasting typhoons, heavy rainfall, heatwaves, and droughts while supporting cross-border applications in agriculture, energy, and infrastructure.

UN Endorsement and International Support

UN Secretary-General António Guterres, speaking at the WAIC 2026 meteorological forum, praised the system, stating that Mazu “pushes forward the UN Early Warnings for All initiative.” Guterres noted that “China’s platforms, satellites, and AI models have helped dozens of countries improve their early warning capabilities” and described the cooperation model as “exactly the cooperation model the world needs — helping countries protect lives through technology transfer, joint research, and local capacity building.”

Wang Jinsong, Director of the National Satellite Center at CMA, emphasized that Mazu represents more than just technology transfer. “‘Mazu’ is not just an AI technology, but an exploration of the integration of global wisdom and local wisdom,” Wang said. “Mazu embodies not only China’s technology and service output, but also the ability to help local areas enhance their autonomous early warning and climate risk response capabilities through capacity sharing and localized application.”

Training and Capacity Building

Since 2024, nearly 1,000 participants from over 100 developing countries have received training in China on early warning technologies. Chen Zhenlin, head of the CMA, confirmed that under China’s 15th Five-Year Plan (2026-2030), the country will continue upgrading Mazu with improved precision and AI integration. Alongside Mazu, China also released the “Fenghe” (风和) AI meteorological service large language model and the Fengyun Satellite AI Toolbox, both opened to global users.

Strategic Implications

Mazu serves multiple strategic objectives for China: it projects soft power by positioning the country as a responsible global actor providing climate solutions to the Global South, showcases Chinese AI and satellite capabilities on a global stage, and strengthens diplomatic ties with developing nations through tangible climate resilience support. For recipient countries, the system provides access to advanced AI-powered weather forecasting that many could not develop independently, along with customized solutions tailored to local climate risks.

However, the expansion also raises questions about data dependency and the geopolitics of weather data. Countries using Mazu may become reliant on Chinese satellite data and AI infrastructure, and the proprietary nature of the AI models may limit external scrutiny. These concerns notwithstanding, the system directly supports the UN’s goal of universal early warning coverage by 2027 and represents a significant model of South-South cooperation in climate technology.

What’s Next

With a target of 30 countries and continued technological upgrades under the 15th Five-Year Plan, Mazu is poised to become a cornerstone of global early warning infrastructure. The coming months will reveal how recipient countries assess the system’s effectiveness independently and how Mazu integrates with or competes with existing early warning efforts by the World Meteorological Organization, developed countries, and private sector actors. For now, Mazu stands as China’s flagship contribution to global climate resilience — an AI-powered system that is already helping protect vulnerable communities from the accelerating impacts of extreme weather.