China's artificial intelligence industry is experiencing robust growth, with a surge in technological innovation marking an unprecedented period of activity. AI is increasingly becoming a powerful engine for high-quality economic development, with large language models, intelligent agents, and AI chips evolving rapidly, enriching product forms, and expanding application scopes, thus transitioning from theoretical to practical applications and reshaping the global productivity landscape and industrial supply chain structures.
Systematically monitoring the AI industry and quantifying its scale is crucial for scientifically and orderly promoting its high-quality development, serving as a fundamental task. The China Academy of Information and Communications Technology (CAICT) has been continuously researching the AI industry and has conducted a calculation of its 2025 scale. According to the estimates, China's AI industry exceeded 1.2 trillion yuan in 2025, marking a 40% increase year-on-year. By June 2026, the number of AI companies surpassed 6,600, accounting for 15% of the global total, forming a complete industrial system covering foundational infrastructure, model frameworks, and industry-specific applications.
Where to start
The AI industry chain is divided into three layers: foundation, model framework, and application. In 2025, the scale of China's AI foundation layer grew by 59% year-on-year, accounting for 38% of the total. The model framework layer saw a 189% increase, representing 7% of the total, while the application layer expanded by 22%, making up 55% of the total. The industry's scale is highly concentrated, with Beijing, Guangdong Province, Shanghai, Zhejiang Province, and Shandong Province being the leading regions, collectively accounting for over 80% of China's total AI industry scale.
The AI industry is characterized by its broad boundaries, rapid product evolution, and accelerating integration with traditional sectors. CAICT has established a dedicated research team, collaborating with industry, academia, and research institutions, organizing multiple rounds of discussions and validations, and driving the development of relevant standards. Building on existing industrial statistics and scale estimation methods, CAICT has proposed a methodology system for calculating AI industry scale. This method adheres to standard guidance and relies on publicly available data and first-hand research, using a four-step process of "defining boundaries," "screening companies," "calculating coefficients," and "measuring scale" to achieve full industry chain and regional scale estimation.
CAICT has led the development of two industry standards related to AI: "Artificial Intelligence - Basic Commonality - Definition of AI Industry Boundaries" and "Artificial Intelligence - Basic Commonality - Identification of AI Enterprises." These standards have already been the subject of specialized interpretation meetings, providing a unified foundation for industrial statistics, policy formulation, and corporate practices. Moving forward, CAICT will continue to strengthen the foundation of industry monitoring, clarify the industry's development landscape, track real-time changes, and accurately identify risks and opportunities, offering robust support for informed government decision-making and the healthy growth of the sector.