China is aggressively investing in the computing infrastructure that underpins its artificial intelligence (AI) industry, despite facing domestic economic challenges.
On October 20, Xie Chun, head of the Ministry of Industry and Information Technology's Information and Communication Administration, announced during a State Council press briefing that China's intelligent computing power reached 2,592 EFLOPS by the end of September, marking a 146% increase compared to the same period last year. Following a record of 725.3 EFLOPS at the end of 2024 and an expansion to 1,590 EFLOPS by the end of 2025, the latest figure indicates a rapid acceleration in the development of AI computing infrastructure. The Chinese government aims to increase its computing power to approximately 9,800 EFLOPS by 2030.
EFLOPS measures the number of floating-point operations per second, with 1 EFLOPS equating to 100 quadrillion operations. China's computing capacity of 2,592 EFLOPS as of September 30 means it can perform 2,592 quadrillion operations per second, highlighting the growing demand for computational resources necessary for AI model training and inference.
The expansion of China's AI infrastructure goes beyond merely increasing the number of data centers. It involves the installation of AI accelerators, servers, high-speed networks, storage systems, power supplies, and cooling facilities within these centers. A national computing network that connects resources across various regions is also a key component, enabling businesses and research institutions to access the necessary computational power.
Additionally, China is strengthening policies to integrate this infrastructure with its manufacturing sector. According to the Ministry of Industry and Information Technology, 5G technology has been applied across 97 major industrial categories in the national economy, while the industrial internet encompasses all 41 major industrial categories. China has established over 26,000 dedicated industrial 5G networks and 1,260 5G factories. The plan is to further expand the integration of AI and the industrial internet to enhance productivity and automation levels in manufacturing.
The Chinese government has projected a cumulative investment of 3.8 trillion yuan in information and communication infrastructure from 2026 to 2030. Alongside the goal of achieving 9,800 EFLOPS of intelligent computing power by 2030, it is also pursuing the establishment of large-scale AI accelerator clusters. Despite a slow recovery in domestic consumption and the real estate market, China's focus on AI computing infrastructure reflects its belief that the competitiveness of future industries hinges on securing computational resources.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.