LG CNS (LG CNS) is partnering with Naver Cloud to build next-generation liquid cooling infrastructure at the Samseong Data Center in Goyang, Gyeonggi Province. The strategy aims to secure a leading position in the AI factory infrastructure market by leveraging advanced cooling technology designed to control heat generated by high-performance GPUs (graphics processing units).
LG CNS announced on the 25th that it will apply direct-to-chip (DTC) liquid cooling technology at the Samseong Data Center. Both companies plan to complete the construction and verification of the liquid cooling infrastructure by 2027.
DTC involves attaching a cooling plate through which coolant flows directly onto the GPU chip to absorb heat. By utilizing liquids with higher heat transfer efficiency than air, this method effectively controls heat generated in highly integrated GPU servers.
The cooling capacity that conventional air-cooling systems can handle is typically around 20–30 kW per rack. In contrast, the power density of next-generation GPU servers is rising to over 200 kW per rack. Consequently, liquid cooling is gaining attention as an essential technology for AI factories operating large-scale GPUs.
The liquid cooling infrastructure being built at the Samseong Data Center will be used to operate NVIDIA's next-generation GPU "Rubin," secured by the Korea Information and Communications Technology Industry Development Institute under the Ministry of Science and ICT. Naver Cloud will operate these GPUs. Rubin's inference performance is expected to be approximately 3.3 times higher than that of Blackwell Ultra.
The Samseong Data Center is one of the largest in Korea, with a total power capacity of 80 MW. LG CNS is responsible for its entrusted operation and has signed a co-location contract with Naver Cloud.
Based on its capabilities in data center design, construction, and operation, LG CNS will integrate liquid cooling, GPU server operating environments, power supply systems, and AI platform operational foundations. Naver Cloud plans to provide high-performance AI services to public institutional investors and enterprises by leveraging its experience in large-scale GPU operations.
Cho Heon-hyeop, Executive Vice President of LG CNS responsible for the Data Center Business, stated, "Data center cooling technology is becoming a core competitive factor that determines the quality of AI services and operational efficiency." He added, "We will build infrastructure to support Korea's AI ecosystem based on our data center operation capabilities and next-generation cooling technologies."
in TECH
