China has unveiled an ambitious plan to significantly expand its artificial intelligence (AI) computing capacity over the next five years, aiming to quintuple the nation’s current capabilities by 2030. The initiative, announced by the Ministry of Industry and Information Technology (MIIT), includes deploying AI computing clusters with up to 100,000 accelerator cards and increasing total computing capacity to 9,800 exa floating-point operations per second (ef10s).
The new five-year plan outlines a cumulative investment of 3.8 trillion yuan (approximately HK$4.43 trillion) in information infrastructure from 2026 to 2030. Key components of the strategy include the orderly construction of intelligent computing clusters equipped with at least 10,000 graphics processing units (GPUs), and developing inference computing facilities tailored to specific AI applications. The plan also emphasizes adapting infrastructure to support domestically produced AI chips.
China’s AI computing capacity has already seen rapid growth. By June 2024, the country’s capacity reached 2,185 ef10s—an increase of 177 percent compared with the previous year—and was supported by 52 facilities, each featuring over 10,000 accelerator cards. Data from the National Data Administration indicated that capacity had further increased to roughly 2,450 ef10s by the end of July.
This buildout is anchored in the “East Data, West Computing” initiative launched in 2022, designed to relocate energy-intensive data processing away from densely populated eastern regions to western areas with more affordable land and abundant energy supply. China has developed a network comprising eight national computing hubs, 10 data center clusters, and three regions coordinating computing capacity with power resources.
Integration efforts are also underway to link regional computing resources that were previously fragmented. By the first quarter of 2024, a government-supported platform aggregated approximately 316 ef10s of intelligent computing capacity, pooling resources from 155 companies with 578 resource pools, including around 720,000 GPUs. Over 80 percent of the country’s total intelligent computing power is concentrated within the eight national hubs.
Demand for AI processing has surged alongside this expansion. Token usage, a measure of AI workload, reportedly grew from about 100 billion tokens daily in early 2024 to 140 trillion tokens by March. The adoption of domestically designed AI chips is advancing, with Chinese-made chips accounting for 41 percent of the country’s domestic shipments in 2025. However, challenges remain due to software ecosystem fragmentation and issues with compatibility.
In addition to AI computing goals, China’s plan outlines priorities for next-generation communications technology. It calls for the timely commercialization of 6G services, enhanced integration of 5G networks with the BeiDou satellite navigation system, and increased use of embedded SIM cards (eSIMs). The plan also encourages the development of technologies that enable communication without reliance on traditional network coverage and promotes research into space-based computing.
China has already made significant strides in 5G deployment, hosting approximately 5.1 million 5G base stations at the end of June 2024—about 69 percent of the global total. With nationwide 5G coverage largely established, telecommunications operators are focusing on upgrading networks and deploying 5G-Advanced technology. For example, China Mobile’s 5G-Advanced network now covers over 330 cities.
Preparations for 6G technology are progressing through standards development and testing stages. In May 2024, China approved the use of the 6GHz frequency band for 6G trials. Meanwhile, the internationally recognized 3rd Generation Partnership Project (3GPP) has commenced work on Release 21, the first global 6G standard expected to be finalized by 2029.