President Xi announced at the opening ceremony of WAIC 2026 plans to deploy the intelligent early warning system MAZU in 30 nations over five years. This transforms a relatively modest meteorological project into a key initiative of China's global cooperation strategy in artificial intelligence.
Shortly after, a version of MAZU Djibouti 2.0 was presented at a specialized meteorological forum. This demonstration showed the initiative's readiness to operate and showcased the integration of space, artificial intelligence, and ground observation capabilities into a single export system.
The MAZU system is not just a single satellite or radar, but a comprehensive infrastructure named after the Chinese sea goddess. The name MAZU itself reflects its four-layer architecture: covering multiple threats, including typhoons, storms, floods, and sandstorms; generating alerts, which converts meteorological data into actionable warnings; seamless delivery, ensuring that alerts reach governments, businesses, and residents without failure; and universal accessibility, allowing developing countries vulnerable to weather conditions to access basic early disaster warning capabilities and make them available.
The technology integrates various Chinese space and AI assets. Fengyun meteorological satellites provide observation over vast areas from space, tracking large-scale weather phenomena through cloud layers, water vapor, typhoons, and strong convection. Ground meteorological radars conduct high-frequency local monitoring for ports, airports, and urban areas. Artificial intelligence models process data from satellites and radars to forecast storm intensification, precipitation distribution, and wind force impact. Parallel to MAZU, the Fengyun Satellite AI Toolbox was announced, which packages satellite data reception, AI output, multi-source data fusion, and operational applications into a unified hardware and software system. As one official noted, previously China exported individual ingredients, but now it exports the entire kitchen, recipes, and the chef's expertise together.
The deployment of the Djibouti 2.0 version illustrates this evolution. Version 1.0 provided an analytical meteorological center capable of processing data and generating forecasts. Version 2.0 supplements this with local intelligent terminals containing meteorological chips and forecasting models that work with deployed agent systems, creating an integrated measurement, forecasting, and alerting system. The World Meteorological Organization estimates that early warning of hazardous weather conditions 24 hours in advance can reduce losses by approximately 30%, and early warning systems can provide up to 10 times return on investment. Many developing countries have access to satellite cloud imagery but lack full networks of ground radars, local forecasting models, and warning delivery systems. MAZU is designed to fill this entire chain—from satellite observation to last-mile warning delivery. Although the direct construction cost may only amount to hundreds of millions of yuan in 30 countries, the strategic significance lies elsewhere: MAZU demonstrates how China combines space assets, AI capabilities, and system integration into an export product of management infrastructure that competes not on component price, but on the value of an integrated, functioning early warning system that strengthens national resilience to natural disasters in the developing world.
Organizers announced that the World Artificial Intelligence Conference (WAIC) will take place in Shanghai from July 17 to 20, 2026. At a press conference on Tuesday, it was stated that the exhibition, new product presentations, and forum sessions will reach record levels.
The conference will cover three locations in the metropolis, utilizing four exhibition halls with a total area exceeding 100,000 square meters for the first time. Over a thousand exhibitors will present more than 3,000 advanced technologies, and over 300 AI-based products will debut globally. The event is scheduled to host more than 140 forums with the participation of over 1,400 experts from China and abroad.
For the first time, the conference will feature a specialized academic track. WAIC Academic, a high-level international academic meeting on AI led by Turing award winners and international academics, will serve as a platform for scientific exchange.
The conference focuses on two key areas: intelligent computing and embodied intelligence. Each of these areas will attract over 200 enterprises. Industry representatives, key supply chain players, state-owned enterprises, and leading international companies will demonstrate the application of AI across various sectors.
Organizers noted on Tuesday that the National Development and Reform Commission (NDRC) reported that the AI-related industry in China exceeded one trillion yuan (approximately $147 billion) last year, with projected growth of over 30% in 2026. The NDRC has established over 30 national AI pilot bases and stimulated state-owned enterprises to open over 1,000 application scenarios.
The penetration of general AI into key industries has exceeded 80%. Official NDRC spokesperson Wang Jiumeng stated that the agency will present two documents at the conference: a collection of successful cases covering over 20 countries in agriculture, industry, energy, and research, and an action plan for AI cooperation and development that covers eight areas, including intelligent computing, open ecosystem, AI empowerment, and security management.
According to NDRC data, in 2025, China exported over 100 million AI-based smart devices, and it is expected that sales of such devices will surpass non-AI model sales for the first time this year. In China, native AI office agents record over 20 million monthly visits, with daily token consumption reaching hundreds of trillions, representing a sharp increase compared to the previous year.
Under the fifteenth five-year plan (2026–2030), Wang noted that China will accelerate independent innovation in models, computing power, and data, find high value-added scenarios, and create benchmark applications. Furthermore, the impact of AI on employment will be studied, and its use in dangerous, dirty, and labor-intensive tasks will be promoted.
