China scales up network to fuel growth of AI

China is accelerating the construction of an integrated national computing power platform, aiming to enhance the quality and efficiency of computational resources and strengthen the country's capabilities in core technologies such as AI chips, said officials and experts.
As the backbone of AI development, computing power has become the battleground for global technological supremacy, and China is charging ahead with unprecedented momentum despite challenges, they said.
The move is China's latest plan to scale up the digital infrastructure that is key to buoying its AI economy and to hone its technological prowess in strategic industries.
Vice-Minister of Industry and Information Technology Xiong Jijun emphasized the government's commitment to "guiding the orderly development of computing infrastructure and improving the supply quality of computing resources" at the 2025 China Computational Power Conference, which concluded on Sunday in Datong, Shanxi province.
He emphasized the need to accelerate breakthroughs in key technologies such as graphic processing units and algorithms for large language models, which are crucial for sharpening China's competitive edge in AI.
Roughly defined as the ability to process data, computing power includes technologies such as chips and data centers to support information processing, data storage and network capacity in the digital economy era.
Wu Hequan, an academician at the Chinese Academy of Engineering, said that computing power is an invisible yet indispensable force behind any AI-driven metamorphosis, and that without enough computing power to process data, it is impossible to support large-scale AI applications.
"Every yuan ($0.14) invested in computing power drives 3 to 4 yuan in GDP growth," Wu said. "In the global race for AI leadership, expanding computing power supply is critical."
China's total computing power has been growing at approximately 30 percent annually, driven largely by soaring demand for intelligent computing fueled by AI applications, data from the Ministry of Industry and Information Technology showed.
Rao Shaoyang, director of the strategic development research institute of the China Telecom Research Institute, said, "AI is expected to contribute over 11 trillion yuan to China's GDP by 2035."
"That will account for 4 to 5 percent of the total economy," Rao said. "This growth could multiply computing power demand by 10 or even a hundred times."
To address this demand, China is building a nationwide computing power platform that integrates resources across regions and industries. The platform, described as a "unified network" for computing power scheduling, already includes subplatforms from 10 provinces and municipalities such as Shanxi, Liaoning, Jiangsu and Shanghai, according to the ministry.
Its intelligent scheduling capabilities allow efficient matching of supply and demand. For example, enabling businesses in eastern China to leverage idle computing resources in western China, significantly reducing costs and improving efficiency.
In one notable case, a Beijing-based healthcare company completed an AI computing task using resources from the Ningxia Hui autonomous region at a cost of just 10,000 yuan, compared with the typical expense of over 2 million yuan for purchasing and maintaining servers.
According to Wei Liang, vice-president of the China Academy of Information and Communications Technology, more than 100 computing service providers have joined the platform, along with over 1,000 registered enterprise users, while nearly 100 large language models have been connected, providing diverse online services to more than 1,000 developers.
Meanwhile, Chinese AI chip companies are also working hard to beef up their prowess in graphic processing units, prompting their stocks to be favored by investors, as the US company Nvidia's H20 chips face weak demand in China.
On Monday, Chinese AI chip stocks continued their strong upward trajectory, with Cambricon Technologies emerging as a potential challenger to beverage giant Kweichow Moutai for the title of the highest-priced stock on the A-share market.
Cambricon closed at 1,384.93 yuan per share on Monday, surging 11.4 percent for the day. Since mid-July, its share price has more than doubled, and it has skyrocketed by more than 562 percent since September 2024. The company now trails only Moutai, which ended the session at 1,490.33 yuan per share.
The AI chip intellectual property products of Cambricon, which was founded in 2016, have powered smartphones and data server chips, including those made by Huawei Technologies Co and Alibaba Group.
Chen Tianshi, co-founder and CEO of Cambricon Technologies, said in an earlier interview with China Daily that "there is rigid demand for AI chips in China. We don't worry about orders, as long as our technologies are strong enough".
Pan Helin, a member of the Ministry of Industry and Information Technology's Expert Committee for Information and Communication Economy, said, "The Nvidia incident offers a strategic window for ambitious Chinese AI chip designers to expand their market share."
masi@chinadaily.com.cn