en.Wedoany.com Reported - China's General Administration of Customs stated at a press conference held by the State Council Information Office on July 14 that in the first half of 2026, China's imports and exports of computing hardware, including electronic components and computer parts, reached 5.13 trillion yuan, a year-on-year increase of 56.6%. As demand for computing resources grows for artificial intelligence training, inference, and industry applications, servers, chips, storage devices, optical communication products, and related components are becoming key equipment sources for computing infrastructure construction.
Computing hardware is not a single product but a complete equipment system required for the operation of data centers and intelligent computing centers. Its scope includes computing chips, AI servers, general-purpose servers, storage devices, printed circuit boards, power modules, cooling systems, optical modules, and internal server components. These devices must be deployed in conjunction with power supply and distribution, cooling, fire protection, network connections, and computing scheduling platforms to form a functional computing infrastructure.
The import and export scale of 5.13 trillion yuan reflects a growing demand for hardware circulation driven by China's computing infrastructure construction. On one hand, newly built and expanded intelligent computing centers require centralized procurement of servers, accelerator cards, storage, and high-speed network equipment; on the other hand, server and electronic component manufacturers also need to procure some equipment, materials, and components from the global supply chain. Computing hardware trade thus encompasses equipment exports, key component imports, and cross-border production collaboration.
Currently, China's computing infrastructure has shifted from single data center construction to coordinated development of computing centers, communication networks, energy facilities, and scheduling systems. A spatial structure of "8+10+3" has been formed nationwide, focusing on 8 national computing hubs, 10 national data center clusters, and 3 computing-energy collaborative development regions. As of the end of March 2026, China's built intelligent computing capacity was approximately 2.5 times that of the same period in 2025, and the national integrated computing network monitoring and scheduling platform has connected about 70% of intelligent computing resources.
Servers and computing chips are only the computing end of the infrastructure. After large-scale AI clusters are put into operation, high-speed optical fibers and optical modules are needed to connect different cabinets, data centers, and computing hubs. In June 2026, Qingdao, Shandong, completed an S+C+L triple-band ultra-low-loss multi-core optical cable line; in the "East Data West Computing" industrial park in Qingyang, Gansu, the data center and power center of the first phase of the Smart Blueprint Western Intelligent Computing Center are under construction, with plans to deploy over 8,800 high-density cabinets and a computing capacity of 108,000 PFlops.
In Zhengzhou, Henan, the core node of the National Supercomputing Internet went online in July, providing computing resources formed by over 100,000 domestic AI accelerator cards. This node also undertakes functions such as cross-regional computing operations, resource scheduling, and supply-demand matching, with construction content covering accelerator card clusters, storage systems, network equipment, and a unified scheduling platform.
These projects generate continuous procurement demands for hardware equipment. Large computing clusters require a large number of AI servers and accelerator cards; high-density cabinets need higher-power power supply and cooling systems; and cross-regional computing scheduling requires high-speed optical modules, optical cables, and network switching equipment. As the scale of data center construction expands, the circulation volume of electronic components and computer parts will also increase accordingly.
Previously released monthly data has already shown a growth trend in computing hardware trade. In April 2026, China's integrated circuit exports reached $31.09 billion, approximately doubling year-on-year; exports of automatic data processing equipment and parts reached $23.81 billion, a year-on-year increase of 47.6%. Related products include computers, laptops, servers, and their components.
Demand for computing hardware is also extending from centralized data centers to edge nodes and intelligent terminals. The General Administration of Customs mentioned at the press conference that products such as AI glasses, AI translators, and mechanical exoskeletons are rapidly updating. Although these devices are at the end of the computing system, they still require chips, sensors, storage, and communication modules, and complete model invocation and data processing through cloud or edge computing facilities.
In terms of construction scope, investment in the computing network includes not only computing equipment procurement but also data center buildings, power centers, power transmission and distribution facilities, communication lines, and software scheduling platforms. Xinhua News Agency reported on July 14 that, including information equipment such as AI servers, high-speed optical modules, and computing scheduling platforms, as well as civil engineering construction, the direct investment scale of China's computing network is expected to reach the trillion-yuan level.
The 56.6% growth in computing hardware imports and exports in the first half of the year indicates that the expansion of computing facilities has transmitted to sectors such as servers, electronic components, computer parts, and optical communication equipment. As more intelligent computing centers, computing hubs, backbone optical cables, and cross-regional scheduling nodes are put into construction, hardware equipment will continue to be a major component of AI infrastructure investment.
