NHN Cloud Bets on Liquid Cooling to Lead AI Infrastructure Race

NHN Cloud identifies liquid cooling technology as a core competitive advantage for next-generation AI infrastructure. The company argues that as AI chip performance and density increase, the ability to reliably manage the heat generated by servers directly impacts service performance and stability. NHN Cloud points to its early development and operational experience with large-scale liquid-cooling infrastructure as a key strength.

The adoption of next-generation AI chips, such as Nvidia's Vera Rubin, is anticipated to make liquid cooling the standard for managing heat output. This transition requires more than just new cooling equipment; it demands structural changes to data centers themselves. Facilities will need to bear loads potentially exceeding 2 tons per square meter to accommodate the weight of these systems, coolant, and piping, a capacity few current Korean data centers possess.

The ongoing AIDC and GPU shortages in Korea have created a supply-side market dynamic, allowing cloud providers to prioritize customers. NHN Cloud has nearly fully allocated its GPU resources, including approximately 7,600 units acquired through a government program, to both external clients and national AI research initiatives. This tight supply has led to inquiries from overseas companies seeking large-scale AI resources.