This week, Microsoft, Google, Amazon, and Meta will all announce their financial reports for the last quarter, and the market will focus on the progress of data center business, capital expenditure and AI-related layout, which also affects the follow-up expectations of the server chain; The market is optimistic about AI servers to drive product design changes, complexity and specification upgrades, coupled with the increase in load requirements, for chassis and rails is a positive trend, but in the first half of this year, the server industry is still facing a slow recovery in demand for general-purpose products, and CSP is in the transition period of switching to self-developed chips.
Server chassis and rail manufacturers such as Chencheng, Chuanhu, Yingguang, Shengmingdian, Nanjun International, etc., in the second half of last year, the revenue of the second half of last year was warmer than that of the first half of the year, especially the large manufacturers Chencheng and Chuanhu increased their contribution to AI products, and the growth rate in the Q4 quarter of last year was more prominent than that of their peers.
Generative AI has driven an AI investment boom, and the supply of NVIDIA AI chips has also increased the value of high-end AI servers, including heat dissipation, chassis, and rails, which are all named and are expected to benefit from the AI server trend.
Due to the high U-number design and scale upgrade of high-end AI servers, ASP will be greatly improved for chassis and rails, and with the gradual increase in AI server shipments this year, it will also help to increase the revenue contribution of AI servers for chassis and rails.
For parts manufacturers, the revenue comes from the combination of P (** multiplied by Q (quantity), the scale is still the focus of observation, and the contribution of AI servers to the overall server demand, as well as the future development direction of AI server architecture, there are still a lot of debates in the industry and the market, parts companies pointed out that it is becoming more and more difficult to define general servers and AI servers in the future, excluding high-end AI servers using NVIDIA chips, as well as ASICs used by CSP manufacturers, In fact, more and more manufacturers are beginning to use general-purpose server architecture to change, and as AI applications and scenarios become more and more diverse, high, medium, and low-end requirements are slowly emerging.
So how do you look at the growth opportunities in parts? In fact, it still depends on whether these parts factories can continue to maintain close cooperation with upstream chip and terminal server manufacturers, and continue to win new specification development projects and create moats.
In the long run, the application of AI servers will become more and more extensive, and it is becoming more and more difficult to define, and parts manufacturers expect that AI servers can become the catalyst for the rapid growth of the overall server output value in the next wave after cloud computing.
Header image**:p ixabay).