As the climate crisis reshapes the economics of energy and the politics of disaster, China has moved meteorology from a public service into a strategic infrastructure — embedding AI-powered forecasting into the operational logic of wind farms, airlines, and power grids, while extending that same capability to dozens of nations through satellite networks and early-warning systems. At a trade fair in Beijing this September, the country's meteorological institutions signaled that this integration is no longer experimental but foundational, touching decisions from turbine shutdowns during typhoons
China expands meteorological services into operational decision-making globally
Forecasts have become operational instructions, not warnings.
Mark
So this is about weather forecasting getting better at predicting storms?
Mimi
It's that, but the real story is that forecasts have become operational instructions. A wind farm doesn't just know a typhoon is coming—it knows exactly when to shut down turbines and when to restart them based on minute-by-minute updates.
Luke
How accurate are these forecasts, though? The article mentions 75 percent accuracy for one model and says another ranked first in an evaluation, but it doesn't say what the baseline was or what "first" actually means in terms of real-world performance.
Mimi
Fair point. The Envision model does provide 45-day global forecasts, which is genuinely ambitious. But the article doesn't tell us how often those long-range forecasts are actually used for operational decisions versus the shorter-term ones.
Mark
Why does China care about exporting this technology to other countries? Is it just goodwill?
Mimi
Partly, but it's also about positioning. If China's systems become the standard for early warning in Southeast Asia, the Middle East, and Africa, that's soft power. It's also a way to demonstrate leadership in climate governance.
Luke
The article says MAZU is deployed in seven countries and used by 40 more through cloud services. But we don't know how deeply integrated it is in those 40 countries or whether they're actually relying on it for critical decisions.
Mark
What's the business model? Are companies like Moji Weather and Envision making money from this?
Mimi
Yes. Moji works with airlines and energy companies on subscription or contract bases. Envision sells its platform and AI models. But the article focuses more on the capability than the economics.
Luke
Right. We know Moji helped Xizang Airlines avoid 44 diversions, but we don't know the financial value of that, or whether it's typical or exceptional.
Mark
So what's the actual significance here?
Mimi
Weather forecasting is becoming infrastructure for the energy transition. You can't run a renewable grid without precise, real-time weather data. China is building that infrastructure domestically and exporting it globally.
Luke
Which means China gets to set standards and maintain relationships with countries that depend on its systems. That's the real story underneath the technology story.
Le Pouls
- Renewable energy operators can no longer afford to treat weather as background noise — a single imprecise typhoon forecast now translates directly into lost revenue, unnecessary downtime, and miscalculated risk across entire power grids.
- China's meteorological establishment has responded by building a nationwide wind and solar resource survey at one-kilometer resolution, giving developers not just averages but decade-long hourly profiles of what any given site will actually face.
- Private AI models are sharpening the edge further — Envision Group's Yuanjing Tianji now issues 45-day global forecasts updating every ten minutes, outperforming Google and Microsoft benchmarks and guiding decisions across the full lifecycle of major storms.
- Beyond its borders, China is deploying MAZU, an AI-integrated early-warning system, in seven countries spanning four continents, while Fengyun satellite data flows in real time to meteorological agencies across Southeast Asia.
- The trajectory is clear: weather intelligence is being recast as diplomatic currency, with data-sharing agreements and early-warning deployments positioning China as an indispensable partner in global climate governance.
As the climate crisis reshapes the economics of energy and the politics of disaster, China has moved meteorology from a public service into a strategic infrastructure — embedding AI-powered forecasting into the operational logic of wind farms, airlines, and power grids, while extending that same capability to dozens of nations through satellite networks and early-warning systems. At a trade fair in Beijing this September, the country's meteorological institutions signaled that this integration is no longer experimental but foundational, touching decisions from turbine shutdowns during typhoons to capital allocation for renewable energy projects. The expansion outward — through partnerships with ASEAN nations and deployments in seven countries via the MAZU early-warning platform — suggests that mastery of weather intelligence is becoming a new axis of geopolitical influence, one measured not in warships but in forecast accuracy and data-sharing agreements.
Weather forecasting has migrated from the evening news into the operational core of industrial decision-making. At China's International Fair for Trade in Services in Beijing this September, the country's meteorological institutions made clear that this transformation is complete — and now moving outward across borders.
The shift is most visible in renewable energy. Wind farm operators no longer simply shut down turbines when a typhoon approaches; granular forecasts now tell them precisely when to halt and when to restart, eliminating unnecessary downtime and revenue loss. The China Meteorological Administration has expanded its wind and solar resource survey nationwide, delivering what officials call 'meteorological genetic profiles' for prospective projects — built on more than a decade of hourly data at one-kilometer resolution. Association president Xu Xiaofeng described the survey as foundational infrastructure for the energy transition, providing a unified data basis for decisions on where to build and how to configure wind, solar, storage, and transmission capacity together.
Private companies are following the same logic. Moji Weather, once a consumer app, now serves airlines, energy companies, and rail operators with industry-specific guidance. For Xizang Airlines, operating at high altitude where plateau storms are both frequent and dangerous, Moji built an AI model that forecasts storms 120 minutes ahead, updating every six minutes, with accuracy above 75 percent — resulting in 44 fewer diversions in a single season. The company also launched ElectroMetis, an energy platform integrating AI forecasts, satellite imagery, and on-site sensors to support generation, grid operations, and maintenance decisions.
Artificial intelligence is pushing the frontier further. Envision Group's Yuanjing Tianji generates global forecasts up to 45 days ahead at five-kilometer resolution, updating every ten minutes, and ranked first in the WeatherBench 2 evaluation ahead of models from Google and Microsoft. During this year's typhoon season, it guided decisions across the entire storm lifecycle — from pre-landfall inspections to turbine restarts and post-storm maintenance.
China is now exporting this capability. Through partnerships with ASEAN nations, it has established data-sharing and joint analysis mechanisms focused on typhoons, floods, and torrential rains. Fengyun satellites deliver high-frequency remote-sensing products to Vietnam, Thailand, Myanmar, the Philippines, and Indonesia during major weather events. The CMA's MAZU system — combining satellite observations, numerical prediction, and machine learning — had been installed in seven countries by July 2026, including Pakistan, Ethiopia, and Sri Lanka, while cloud-based services reached meteorological agencies in more than 40 nations. Xu framed China's role as both a contributor to other countries' disaster response and a more active voice in global climate governance — a standing built not through declarations, but through shared data and deployed systems.
Weather forecasting has moved from the evening news into the engine room of industrial operations. At China's International Fair for Trade in Services in Beijing this September, the country's meteorological establishment made clear that the shift is now complete—and accelerating outward across borders.
The transformation is most visible in renewable energy. A wind farm no longer simply receives a typhoon warning and shuts down wholesale. Instead, operators use granular forecasts to decide precisely when to halt turbines and when to restart them, eliminating unnecessary downtime and the revenue losses that follow. The China Meteorological Administration announced the nationwide expansion of its wind and solar resource survey, which now delivers what officials call "meteorological genetic profiles" for prospective projects. Built on more than a decade of hourly data at one-kilometer resolution, the survey has grown from pilot programs in six provinces to cover the entire country, helping developers assess not just how much wind or sun a site receives, but whether resources are stable, complementary, and what hazards they might face.
Xu Xiaofeng, president of the Chinese Meteorological Service Association, framed the work as foundational infrastructure for the energy transition. The survey, he said, provides "a unified and reliable data foundation for renewable energy planning," supporting decisions on where to build projects and how to configure wind, solar, storage, and transmission capacity in concert. This is no longer meteorology as a service to the public. It is meteorology as a tool embedded in capital allocation.
Private companies are following the same logic. Moji Weather, once known primarily as a consumer weather app, has built a business around translating meteorological data into industry-specific operational guidance. The company works with airlines, energy companies, and rail operators, combining weather intelligence with clients' own business data to generate tailored solutions. For Xizang Airlines, which operates at high altitude where thunderstorms are both frequent and dangerous, Moji developed an AI model that forecasts plateau storms 120 minutes ahead, updating every six minutes, with accuracy above 75 percent. The result: 44 fewer flight diversions and returns in a single season. The company also unveiled an energy platform called ElectroMetis that integrates high-resolution weather data, AI forecasts, satellite and radar imagery, and on-site sensors to support decisions on power generation, grid operations, trading, and maintenance.
Artificial intelligence is pushing the frontier further still. Envision Group's Yuanjing Tianji, a weather foundation model, generates global forecasts up to 45 days ahead at five-kilometer resolution, updating every ten minutes. It ranked first in the latest WeatherBench 2 evaluation, ahead of models from Google and Microsoft. During this year's typhoon season, it guided decisions across the entire lifecycle of storms—from pre-landfall equipment inspections and resource deployment to turbine shutdown and restart calls, and post-storm maintenance planning. The value, according to Envision executives, lies in converting meteorological uncertainty into better operational choices. More precise weather information allows operators to balance safety and financial returns with greater precision.
China is now exporting this capability beyond its borders. Rather than simply selling weather data, the country is offering forecasting, early-warning, and decision-support systems tailored to other nations' needs. With ASEAN countries, China has established regular mechanisms for meteorological data sharing, joint analysis, and technical cooperation focused on typhoons, torrential rains, and floods. Fengyun satellites provide high-frequency remote-sensing products to Vietnam, Thailand, Myanmar, the Philippines, and Indonesia during major weather events. The CMA developed MAZU, an AI-powered integrated early-warning system that combines satellite observations, numerical weather prediction, and machine learning models. It can be deployed through cloud services or installed locally, depending on each country's infrastructure and disaster risks.
As of July 2026, MAZU had been installed in seven countries—Pakistan, Ethiopia, the Solomon Islands, Jordan, Sri Lanka, Mongolia, and Djibouti—while its cloud-based services were in regular use by meteorological authorities in more than 40 countries. Unny Sankar Ravi Sankar, Malaysia's minister for economic affairs, called the initiative a demonstration of international cooperation in building multi-hazard early-warning capacity. Xu framed China's role as a contributor to other nations' disaster response capabilities and a more active participant in global climate governance. By sharing data, technology, and early-warning systems, he said, China helps other countries respond to extreme weather while advancing its own standing in international climate discussions.
Citations marquantes
By putting meteorological capabilities and business data together, we can develop solutions tailored to specific needs.— Jin Ruichao, general manager of Moji Weather's government and enterprise business department
By quantifying meteorological uncertainty, we can strike a better balance between safety and returns.— Zhang Pu, senior executive from Envision Group
Envie de l'histoire complète ?
Lire l'original sur China.org.cn ↗
Economy & Finance