China Is Collecting Militarily Useful Weather Data Abroad

A global meteorological warning system has become an excuse for China to embed AI-feeding weather stations on foreign territory, gathering data that would help its armed forces in war.

Solomon Islands and countries surrounding India are among the many locations at which the system is collecting weather data, and perhaps other information, then sending it back to China.

By embedding what it calls the Mazu AI Weather Forecasting System in foreign countries, including several developing nations which have all signed up to the Belt and Road Initiative, China has found another way to increase their dependency on it, this time in the vital area of weather warning and disaster mitigation.

But civil authorities aren’t the only people who are interested in the weather. It affects military operations in myriad ways – for example, in determining what kind of reconnaissance or surveillance equipment should be used, or whether and when a contemplated operation can succeed. Atmospheric conditions affect homing sensors and anything that flies, including missiles.

Moreover, the better Mazu’s forecasting, the higher its military as well as civilian value.

So the Chinese armed forces are surely involved closely in Mazu.

In August, the meteorological administration said that customised Mazu systems had been installed in Pakistan, Ethiopia, Jordan, Sri Lanka, Mongolia and Djibouti and elsewhere. In Pakistan, for example, five China-supplied weather stations reportedly transmit real-time observations to Mazu.

Of even more concern for Australia is that the administration’s list included Solomon Islands. In July 2023, Solomon Islands signed a memorandum of understanding with China on disaster and emergency response plans, including early warning systems and cooperation on remote sensing. Papua New Guinea has also employed forecasting services from China and exchanges weather data with it, although no local Mazu facility has been installed.

The Mazu service combines meteorological satellite observations, supercomputing and AI with meteorological ground stations, equipment and training to enable monitoring of local weather.

While regional governments want a greater ability to monitor and forecast their local weather, it is at least plausible that the stations can be used for other purposes, such as collecting radio intelligence. Host countries may know little of what’s connected to the antennas or exactly what’s transmitted back to China.

At the World Artificial Intelligence Conference in Shanghai in July 2026, Xi Jinping pledged to expand Mazu to 30 countries in the next five years. Indo-Pacific nations struggling with real-time forecasting of floods and cyclones have understandable reasons to accept what China describes as a global public good.

But if Australia and its allies, distracted by international and domestic problems, pay little attention to this development, Xi will achieve his unstated goal: Chinese dual-use surveillance technology will have surrounded Australia and uprooting it will be improbable. To prevent the technology’s spread, Australia and its partners, especially in the Pacific, need to have some difficult conversations now to avoid harsher ones later. This entails explaining why a useful weather forecasting system can be a Trojan horse.

The Mazu system has at least one overt military connection. It uses weather satellites of the Fengyun series developed by the giant, state-owned civil and military space contractor China Aerospace Science and Technology Corporation, particularly its subsidiary Shanghai Academy of Spaceflight Technology (SAST). The latest model in the series is the Fengyun 4.

The US and Japan have sanctioned SAST for its weapons research and aerospace development.

The system’s space-weather forecasting also warrants scrutiny for its use of several Feng series meteorological AI models. These predict solar and geomagnetic phenomena that can disrupt satellites, navigation and radio links. Again, military applications are obvious.

Beyond the security problem, there’s a familiar risk of China achieving a monopoly and resulting local influence in the host countries. Mazu’s foreign hosts are likely to become dependent on it. Relying on China for software, updates, repairs and specialist expertise, they’ll have little control over the output. They will also find that changing systems is expensive and disruptive. We see a similar pattern in various countries’ reliance on Chinese telecommunications equipment.

The time for Canberra and regional capitals to act is while recipient governments still have options. Three practical actions can be implemented immediately.

First, Australia’s Department of Foreign Affairs, the Bureau of Meteorology, the Department of Defence and the National Intelligence Community should jointly assess Chinese-supported weather installations and agreements with Pacific partners.

Second, the issue could form part of discussions by members of the Quad – Australia, India, Japan and the United States – if it does not already. All four nations share concerns about both China’s dual-use technology and its influence across the region.

Third, given that the World Meteorological Organization (WMO) has already partnered with the China Meteorological Administration at the World Artificial Intelligence Conference in Shanghai in both 2025 and 2026, Australia should work with like-minded partners and the WMO to ensure there is increased awareness and transparency of Mazu AI. This might be done through funding equipment and technology audits to ensure that host governments control credentials and data-sharing agreements and have the ability to change suppliers without losing essential services.

These measures are needed to better understand how nominally civilian Chinese technologies are being supplied and applied globally and how they might be used militarily in any future conflict. There is no need to wait for such a conflict or do nothing in the mere hope there will not be one.

Tilla Hoja is a China analyst and Justin Bassi is executive director at ASPI.

Image (top) of a satellite view of Tropical Cyclone Debbie: NASA/NOAA/Suomi NPP via Flickr.

AI contributed no ideas to this article – Tilla Hoja and Justin Bassi.

This article appeared originally at The Strategist (ASPI).