China's AI Server Market Size Expected to Reach $13.4 Billion by 2027
-
After the rise of ChatGPT, major tech companies have been focusing on large AI models, leading to a shortage of general-purpose computing chips. The AI era is surging, with massive data generating enormous computing power demands, driving the increasing need for AI servers and the rising demand for interface chips used in internal and external data transmission within servers.
China's AI Server Market Size to Reach $13.4 Billion by 2027
On November 29, at the 2023 Artificial Intelligence Computing Conference (AICC), IDC and Inspur Information jointly released the 2023-2024 China AI Computing Power Development Assessment Report. The report predicts that China's AI server market size will reach $9.1 billion in 2023, a year-on-year increase of 82.5%, and is expected to grow to $13.4 billion by 2027, with a compound annual growth rate (CAGR) of 21.8%.
The report highlights that AI is rapidly advancing from perceptual intelligence to generative intelligence, leading to a surge in computing power demands. It estimates that from 2022 to 2027, China's intelligent computing power will grow at a CAGR of 33.9%, while general-purpose computing power will grow at 16.6%. Globally, the AI hardware (server) market is projected to expand from $19.5 billion in 2022 to $34.7 billion in 2026, with a CAGR of 17.3%. Notably, the share of servers dedicated to generative AI will rise from 11.9% in 2023 to 31.7% in 2026 within the overall AI server market.
The report also notes that China's AI technology is accelerating into a comprehensive application era. In 2023, the top five industries with the highest AI penetration are internet, telecommunications, government, finance, and manufacturing, with telecommunications jumping from fourth to second place compared to the previous year.
AI servers are data servers capable of providing artificial intelligence (AI) services. They support local applications and web services while also delivering complex AI models and services for cloud and on-premise servers. AI servers enable real-time computing for various AI applications.
There are two primary architectures for AI servers: hybrid architecture, which stores data locally, and cloud-based architecture, which utilizes remote storage and hybrid cloud storage (a combination of local and cloud storage) for data management.
The cost of AI servers is primarily driven by chips like CPUs and GPUs, accounting for over 50% of the total cost.
A server consists of several components including power supply, CPU, memory, hard drive, fan, and optical drive. The cost of chips (such as CPU and GPU) accounts for a significant portion, ranging from 25% to 70%.
AI servers adopt a heterogeneous architecture, which can be categorized into combinations like CPU+GPU, CPU+FPGA, and CPU+ASIC. Currently, GPUs remain the preferred choice for data center acceleration, while the application of other non-GPU chips is gradually increasing. IDC predicts that by 2025, the share of non-GPU chips will exceed 20%. Generally, ASICs offer the best performance but have weaker programmability and flexibility. For training or general purposes, GPUs are the better choice.
AI Server Industry Market Analysis
China is at the forefront of the world in AI server technology. AI servers use a CPU+accelerator chip architecture, which provides efficiency advantages in model training and inference. Unlike the monopolistic situation of foreign AI chip manufacturers, China's AI server technology leads globally.
According to IDC data, in 2022, Inspur held the top position in China's AI server market (by sales) with a 46.6% share, followed by H3C and Ningchang with 11% and 9%, respectively.
The upstream of China's AI server industry chain includes core components and software supply; the midstream consists of AI server manufacturers; and the downstream covers various application markets, including internet companies, cloud computing firms, data center service providers, government departments, financial institutions, healthcare, and telecom operators.
From the perspective of Du Yunlong, an IDC China AI infrastructure analyst, the technical level of domestic chip manufacturers is relatively behind compared to international advanced standards, and the ecosystem is not yet fully developed. However, the supply-demand relationship in the chip sector is gradually changing. Many companies are shifting from "international procurement" to "local procurement" or "self-research and self-use," creating favorable conditions for the development of China's chip industry.
Over the past five years (2018–2022), cities such as Beijing, Hangzhou, Shanghai, Shenzhen, Guangzhou, Hefei, Suzhou, and Chongqing have shown outstanding performance in the AI field. Influenced by generative AI and investments in intelligent computing centers, the 2023 China AI City Evaluation Rankings have changed. Beijing remains in first place, while Hangzhou and Shenzhen rank second and third, respectively. Other cities in the top ten include Shanghai, Suzhou, Guangzhou, Jinan, Hefei, Chongqing, and Chengdu.