From scale to strength

China needs to bolster its computing power to build a solid foundation for the development of its intelligent economy

Against the backdrop of the accelerating new wave of technological revolution and industrial transformation, computing power has become not only a cornerstone of national competitiveness but also a significant engine driving technological innovation, industrial upgrading and social transformation. As China enters the 15th Five-Year Plan (2026-30) period, amid rapid advances in global artificial intelligence and intensifying technological competition among major powers, making strategic breakthroughs in this field has emerged as a key priority for the country's overall development.

During the 14th Five-Year Plan (2021-25) period, China saw steady improvements in both computing power performance and scale. The construction of data centers, supercomputing facilities and intelligent computing platforms accelerated sharply. Increased R&D investment in chip architecture, core algorithms, basic software and high-performance computing frameworks helped China gradually narrow its performance gap with developed economies.

The latest data from the Ministry of Industry and Information Technology show that by September 2025, China's computing centers had operated more than 12.5 million standard racks, with intelligent computing capacity surging to 1,053 exaflops. Throughout last year, its market reached 835.1 billion yuan ($121.3 billion), posting a year-on-year rise of over 30 percent. China ranked second globally in both total computing capacity and intelligent computing capacity.

This expansion boosted digital capital accumulation, underpinned artificial intelligence model training, large-scale data processing and industrial digital transformation, and effectively supported the real economy's intelligent upgrading, thereby laying a technical foundation for higher productivity.

Equally important to the improvements in performance and scale has been the ongoing optimization of China's computing infrastructure layout. In early 2022, the "East Data, West Computing" initiative was launched, establishing a national integrated network of hub nodes in eight regions.

These nodes — spanning the Beijing-Tianjin-Hebei region, the Yangtze River Delta, the Guangdong-Hong Kong-Macao Greater Bay Area, the Chengdu-Chongqing region, Guizhou province, the Inner Mongolia autonomous region, Gansu province and the Ningxia Hui autonomous region — were selected based on factors such as local energy structure, industrial presence, market maturity and climate conditions.

The initiative has spurred the development of data center clusters, promoting more intensive, large-scale and green growth while fostering integrated innovation across computing power, algorithms, data and applications.

Today, China has built a nationwide platform with a "1+M+N" multitiered architecture. This integrates 149 computing power providers and aggregates total computing resources of 163 exaflops. Meanwhile, the identification system now covers 31 provinces, autonomous regions and municipalities, laying a solid foundation for cross-domain resource scheduling.

Leveraging its comprehensive industrial system and vast application scenarios, China has forged a two-way empowerment model between the computing industry and the real economy, establishing a competitive edge in the global computing landscape.

The latest data from the Ministry of Industry and Information Technology show that the country now operates over 35,000 basic-level, over 8,200 advanced-level, and over 500 excellence-level and flagship smart factories. These excellence-level facilities span 31 provinces, autonomous regions and municipalities, cover over 80 percent of manufacturing categories and support nearly 2,000 cutting-edge applications, such as intelligent warehousing, online inspection, digital R&D, production scheduling and quality traceability.

This strong industrial and scenario-based foundation has enabled a development path where computing power originates from real-world applications and flows back into industries. It has created synergies among computing infrastructure, large-model development and industry deployment — offering sustained momentum for China's competitiveness in this field.

However, progress notwithstanding, a substantial gap remains between China and leading nations in the computing power sector. A primary vulnerability is the relative dependence on foreign sources for some critical high-end chips. The software ecosystem presents another critical bottleneck, characterized by a pronounced hardware-software imbalance. Nvidia's CUDA ecosystem, honed over two decades, offers a relatively complete technology stack — from underlying drivers, compilers and library functions to high-level frameworks — accumulating a vast repository of optimized operators, toolchains and developer expertise, with millions of developers worldwide. This sharply contrasts with the state of domestic platforms, which still lag behind in compatibility, maturity and ease of use. The scarcity of advanced software platforms and application ecosystems leads to high migration costs for developers, ultimately undermining the effective utilization of China’s overall computing capacity.

At the infrastructure layer, the marketization and coordinated scheduling capabilities of computing power supply need further improvement. Cross-regional and cross-tier scheduling mechanisms have not yet been fully put in place, and problems such as network latency and bandwidth bottlenecks between the eastern and western regions remain unresolved.

As China enters the 15th Five-Year Plan period, strengthening foundational capabilities and addressing weaknesses are conducive to building a solid foundation for the intelligent economy, securing an edge in global computing power competition and injecting strong momentum into high-quality growth.

First, it is essential to address inadequacies in the software ecosystem and foster a coordinated hardware-software development model. Guided by open-source principles, efforts should be made to promote in-depth adaptation between domestic computing platforms and mainstream AI frameworks and development tools, lowering migration costs for developers. Support should be provided for building domestic developer communities, integrating ecosystem partners, and expanding operator libraries, model repositories, and application solutions.

Second, it is vital to optimize the infrastructure layout and set up an efficient, collaborative network. Priority should be given to accelerating the construction of a national integrated network to realize the flexible allocation of cross-regional and cross-tier resources. Additionally, diverse entities should be encouraged to participate in the construction of commercial data centers, with a focus on cultivating globally competitive cloud infrastructure service providers and elevating the market-oriented development of services.

Third, it is imperative to deepen computing power's enabling role and advance the in-depth integration of the digital and real economies. Support can be provided for the construction of industry-level service platforms, and initiatives can be launched to cultivate high-value application scenarios in key areas such as manufacturing, agriculture and healthcare. Besides, measures can be taken to strengthen the training of professionals in applications, improve digital literacy across society, and provide talent support for unlocking its value.

The author is a professor at the School of Economics and Management at the University of Chinese Academy of Sciences.

The author contributed this article to China Watch, a think tank powered by China Daily. The views do not necessarily reflect those of China Daily.

Contact the editor at editor@chinawatch.cn.