Current State of AI in China

China’s artificial intelligence (AI) commercialization has reached a significant milestone. As of March this year, the daily average token usage surpassed 140 trillion, marking a growth of over 40% since the end of last year. AI development is empowering various industries and driving rapid growth in related fields. In the first quarter, the value added of the digital product manufacturing industry increased by 11.2% year-on-year, with the value added in the manufacturing of electronic materials and integrated circuits directly related to AI production and application growing by 32.5% and 49.4%, respectively. The impact of AI extends further upstream to the chemical and power industries, which provide raw materials and energy.

Experts indicate that AI is accelerating its deep penetration into the real economy, presenting significant innovation opportunities for high-quality economic and social development, and becoming a key driver for cultivating new productive forces and reshaping competitive advantages.

Formation of a Positive Cycle

Tokens are the smallest basic units of information processed by AI large models. Each interaction between users and AI essentially represents an exchange of computing power resources and data value, with tokens serving as the settlement unit.

This year, intelligent agents like “lobster” have significantly increased token consumption. Data shows that the daily average token usage in China was 100 billion at the beginning of 2024 and is expected to leap to 100 trillion by the end of 2025, having surpassed 140 trillion as of March this year.

“This data clearly indicates that large models and various AI applications have truly moved out of the laboratory and into enterprise production and public life. Concurrently, the related hardware industry has also achieved rapid growth, indicating that China’s AI development is no longer a singular ‘software boom’ but has formed a positive cycle from underlying computing power and hardware infrastructure to upstream application ecosystems,” said Wang Guoqing, Vice Dean of the AI Institute of Sichuan Province and Professor at the University of Electronic Science and Technology.

Not only is there growth within the country. In February, data from the world’s largest AI aggregation platform showed that China’s large AI models surpassed the United States for the first time with a token usage of 41.2 trillion. Notably, the platform’s users are mainly overseas developers, with American users accounting for 47.17% and Chinese developers only 6.01%. This makes the data more reflective of the true global appeal of China’s large models.

“The ’token export’ essentially transforms China’s electricity into high-value digital services that can be delivered globally. This not only reflects the development of China’s intelligent industry but also showcases the systematic advantages formed by China’s AI in terms of energy costs, open-source ecology, and industrial chain collaboration,” said Zhu Xufeng, Dean of the School of Public Management at Tsinghua University.

China possesses the world’s strongest industrial chain and manufacturing capacity, having formed a relatively complete AI industrial chain ecosystem from chip manufacturing, smart hardware, algorithm research and development to various downstream application scenarios and physical products. The country also boasts a robust infrastructure, with nationwide high-speed mobile communication (5G), cloud computing platforms, and big data centers providing hardware support for AI training and deployment. Additionally, China is the world’s largest producer and consumer of electricity, leading in installed capacity for renewable energy sources like wind and solar, and has the longest ultra-high voltage transmission network, providing ample and inexpensive power for the energy-intensive AI computing industry. China has also cultivated a large pool of undergraduate and graduate students in mathematics, computer science, and engineering, providing a strong talent reserve for the AI industry.

Strengthening Source Innovation

Driven by algorithm optimization, computing power enhancement, and data accumulation, AI has demonstrated strong versatility and penetration. Experts note that for AI to truly achieve comprehensive empowerment, several challenges remain.

Wang Guoqing analyzed that at the algorithm level, current AI models perform excellently under closed test sets, but face significant challenges in robustness, interpretability, and security in complex and dynamic “open world” environments. At the data level, although China has a vast amount of data, high-quality, finely labeled multimodal data in vertical fields remains scarce, directly limiting the development of professional-grade AI. At the computing power level, cloud computing costs remain high, necessitating breakthroughs in low-power, high-efficiency edge computing for practical applications like robot control and high-security identity authentication.

With support from policies and capital, some companies are accelerating their layout in cutting-edge technology fields to overcome computing power bottlenecks. Recently, iFlytek announced a strategic investment in the quantum computing team from Tsinghua University, establishing a joint venture to explore the synergy between AI and quantum technology.

“AI has become a key area of international competition. There is still a half to one generation gap between China and the US in top model capabilities, mainly due to computing power and data, as the latter has trained earlier and on a larger scale,” said Liu Qingfeng, Chairman of iFlytek. He believes that in the next decade of AI development, both the scientific and industrial communities must seek new development paths, with quantum computing potentially being one of the answers.

The 14th Five-Year Plan clearly states the need to implement strategic deployments for AI and quantum technology, placing quantum technology at the forefront of future industrial “new economic growth points.” Liu Qingfeng emphasized the importance of not being limited to existing technological iterations but actively planning for the next generation of AI, particularly in disruptive fields like “AI + quantum,” to strengthen source technology innovation and establish a more robust mechanism to encourage original innovation, laying a solid foundation for the next generation of AI development.

A Clearer Blueprint

China possesses several innate conditions and structural advantages for developing the AI industry and should vigorously promote innovation in AI technology, industry, and market applications, enabling AI to empower various sectors, according to Zhu Xufeng.

With proactive policies, the blueprint for top-level design is becoming clearer. In August 2025, the State Council issued the “Opinions on Deepening the Implementation of the ‘AI+’ Action Plan,” emphasizing not only AI technology itself but also how AI can empower industrial development. Recently, the Ministry of Industry and Information Technology and eight other departments jointly issued the “Implementation Opinions on the ‘AI + Manufacturing’ Special Action,” proposing that by 2027, China will achieve a safe and reliable supply of key core AI technologies, maintaining a leading position in industrial scale and empowerment levels.

Li Lecheng, Secretary of the Party Leadership Group and Minister of the Ministry of Industry and Information Technology, stated that the ministry will promote the intelligent upgrade of the entire manufacturing process, deeply embed AI technology into core production processes, and expand application scenarios such as intelligent auxiliary design, virtual simulation, and fault warning, comprehensively transforming innovation paradigms, production methods, and management models. The ministry aims to accelerate the iteration and innovation of smart products and equipment, promote the replacement of AI smartphones, AI computers, and other consumer terminals, and accelerate the research and application of new-generation intelligent terminals like humanoid robots and brain-computer interfaces, driving deep integration of large models with smart connected vehicles and CNC machine tools.

“The ‘AI+’ action can greatly stimulate the innovation vitality of the digital economy and accelerate the development of new productive forces,” Zhu Xufeng stated. In emerging digital economy sectors, the deep integration of AI with frontier technologies such as big data, the Internet of Things, and blockchain is giving rise to new business formats. In traditional manufacturing, the introduction of technologies like intelligent robots and machine vision can automate and intelligently control production processes, enhancing industry flexibility and competitiveness, and pushing manufacturing toward high-end and intelligent development.

Zhu Xufeng suggested that the government should play a leading role, concentrate advantageous resources, and increase funding for top research teams to lay a solid foundation for AI development. At the same time, it should encourage market and social capital to actively invest in technology development and industrial innovation, forming a diversified AI industry investment landscape. By gathering government and business resources, the development of the industry can be promoted, enhancing China’s competitiveness in the global AI landscape.

Wang Guoqing believes that more supportive policies for basic research in AI application fields should be introduced, and a smoother platform for the transformation of production, education, research, and application should be established. Focusing on national major needs and market pain points, collaboration between universities and enterprises should be encouraged to shorten the cycle from core technological breakthroughs to the landing of terminal products, allowing AI to truly play a greater role in practical scenarios such as hospitals, factories, and schools.