top of page
AI伺服器設廠設計 (15).png

AI server factory

The site design of the AI ​​server needs to consider multiple factors to ensure the efficiency and stability of the server operation. First, the site should have enough space to accommodate the server cabinets and related equipment, and allow air circulation to help dissipate heat. Secondly, the venue should have a stable power supply and backup power system in case of power outage. In addition, the venue should have an efficient cooling system, such as air conditioning or liquid cooling, to control the heat generated by the servers when they are running.high resolution.

AI伺服器設廠設計 (1).png
AI伺服器設廠設計 (2).png
AI伺服器設廠設計 (3).png
AI伺服器設廠設計 (4).png
AI伺服器設廠設計 (5).png
AI伺服器設廠設計 (6).png
AI伺服器設廠設計 (7).png
AI伺服器設廠設計 (8).png
AI伺服器設廠設計 (9).png
AI伺服器設廠設計 (10).png
AI伺服器設廠設計 (11).png
AI伺服器設廠設計 (12).png
AI伺服器設廠設計 (13).png
AI伺服器設廠設計 (14).png
AI伺服器散熱液冷系統 (6).png

AI server cooling liquid cooling systemactory

As the demand for AI servers continues to increase, cooling technology is also constantly evolving. The future design trend will shift from air cooling to liquid cooling, and liquid cooling technology will be widely used in AI servers.

AI伺服器散熱液冷系統 (2).png
AI伺服器散熱液冷系統 (3).png
AI伺服器散熱液冷系統 (6).png
AI伺服器散熱液冷系統 (4).png
AI伺服器散熱液冷系統 (1).png

1. High-performance computing requirements: The thermal design power consumption (TDP) of AI server chips continues to increase, making the heat dissipation efficiency of the cooling module critical. The TDP of the new generation of servers has exceeded the 500W limit of traditional air cooling, so air cooling is no longer sufficient.

2. Liquid cooling trend: With the rapid growth of AI servers, liquid cooling will gradually become the mainstream. Liquid cooling is divided into water cooling and oil cooling. Water cooling is currently the most used. Through liquid cooling, the heat dissipation efficiency can be effectively improved and the chip can be cooled down.

3. Performance and energy saving: Compared with air cooling, liquid cooling can save energy and improve computing performance. According to NVIDIA's tests, the amount of cabinets required for liquid cooling can be reduced by 66%, energy consumption can be reduced by 28%, and computing performance can be improved.

 

In short, as the demand for AI servers continues to increase, liquid cooling will become mainstream and is expected to explode in the next few years.

 

We add a cooling system to the cabinet design to ensure a significant improvement in chip performance

bottom of page