ChatGPT Ignites The Server Computing Infrastructure Sector

     ChatGPT model training involves a large amount of vector and tensor operations, and AI servers have the advantage of computational efficiency. Large model training is expected to drive the increasing demand for AI server procurement. A server is a high-performance computer that serves as a node in a network, storing and processing 80% of the data and information on the network. Therefore, it is also known as the soul of the network. Usually, a website server can be a single machine or a group of machines, which can be composed of a single or multiple high-performance computers. Compared to ordinary PCs, servers require higher stability, security, performance, and other aspects.

  Industry insiders have pointed out that the power consumption of servers is generally around 400 watts or even higher. When the CPU gets hotter, system manufacturers must solve the problem of heat sources. Zhejiang Securities stated that air-cooled cooling has reached the capacity ceiling, and there is an urgent need to upgrade liquid cooling. It is expected that the demand for AI server chip level liquid cooling will reach tens of billions of yuan: the estimated global and Chinese AI server liquid cooling market size in 2025 is 22.3-33.3 billion RMB and 7.2-10.8 billion RMB.