China’s AI infrastructure plan appears to be part of a broader national infrastructure strategy announced by the Chinese leadership in April this year, says Omdia
In sum – what to know:
National initiative – China is reportedly developing a plan to build interconnected computing hubs to support domestic AI development.
Telco leadership – Omdia said state-owned telecom operators could take a more central role in China’s AI ecosystem under the initiative.
Domestic ecosystem – The effort aims to coordinate computing, networking, and energy resources while relying heavily on homegrown technologies.
China is reportedly preparing to invest about CNY2 trillion ($295 billion) over the next five years in a nationwide network of AI-focused data centers and computing hubs, as Beijing seeks to strengthen domestic AI capabilities and reduce reliance on foreign technology.
According to a Bloomberg report, government agencies including the National Development and Reform Commission are drafting a plan under which state-owned operators such as China Mobile and China Telecom would operate much of the AI infrastructure. The report also said at least 80% of the technology used in the initiative, including AI chips, would come from domestic suppliers such as Huawei.
Guang Yang, senior principal analyst at Omdia, told RCR Wireless the reported plan appears to be part of a broader national infrastructure strategy.
“The plan should be part of the six-network initiative that was announced by China’s top leadership in late April. The computing network is one of the six networks. The others include water supply, electricity, new-generation telecom networks, etc.,” the analyst said.
According to Yang, the initiative is intended to coordinate computing resources, communications infrastructure, and power availability rather than simply build additional AI data centers.
“The goal is not only to build AI data centers but also to achieve efficient coordination among computing, networks, and electricity to match AI demand with computing resources and cost-effective energy supply, e.g., to connect cost-sensitive AI workload to AI data centers with low-cost renewable energy supply or to connect quality-sensitive AI workload to edge AI data centers for low latency,” Yang said.
The analyst said the architecture naturally positions telecom operators to play a leading role. “As networks play a critical role in the architecture, Chinese telcos will lead the initiative, potentially creating new revenue opportunities for them, but also imposing a heavy capex burden.”
Yang added that the initiative could alter the structure of China’s AI market, where hyperscalers have traditionally held a dominant position.
“More importantly, this may be an opportunity for state-owned telcos to take a central role in the Chinese AI ecosystem, which is currently dominated by privately owned hyperscalers like Alibaba, ByteDance, and Tencent.”
The reported plan also places significant emphasis on domestic technology. Yang noted that Chinese operators have already been deploying AI infrastructure based largely on Chinese AI chips.
However, he said adoption challenges remain. “But obviously, it is complex and time-consuming for customers to migrate their AI workloads from the CUDA ecosystem to homegrown Chinese systems. So, the utilization of Chinese telcos’ AI data centers does not look as good as their competitors, which use Nvidia’s chips.”
Looking ahead, Yang said government-funded sectors such as healthcare, transportation, and smart city projects could help drive utilization of the planned infrastructure.
“So, the first step should be to leverage demand from these sectors to fill in telcos’ AI data centers, creating a closed loop from investment to revenue.”
