South Korean IT company NHN Cloud has launched an AI cluster located in its Seoul data center.

The company revealed today, May 13, that it has deployed and started operating a GPU cluster made up of 7,656 'high performance GPUs' at its dedicated AI data center in Yangpyeong-dong, Yeongdeungpo-gu, Seoul.

The cluster is accessible via a GPU-as-a-Service to public sector customers, along with industry and academia research institutions.

The cluster is part of South Korea's Project to Strengthen the Foundation for Utilizing AI Computing Resources, which was launched in 2025, aiming to increase AI compute capacity within the country.

NHN secured the contract for the project from the Ministry of Science and ICT in July 2025. At the time, it said that the cluster would comprise Nvidia B200 GPUs.

The cluster saw an investment of around one trillion won ($671m).

NHN's Yangpyeong Data Center was designed to meet the requirements of high-density compute environments, and the cluster uses a liquid cooling system enabling the GPUs to be grouped up to around 4,000 chips. According to NHN, the cooling solution further enabled them to reduce energy consumption by between 15 and 20 percent compared to conventional air cooling.

The company operates four data centers in South Korea, plus a site in Tokyo, Japan. NHN previously developed the Gwangju National AI Data Center, from which it also offers a GPUaaS. From the NGN Cloud Center Pangyo, the company provides Blackwell Ultra GPU infrastructure to Krafton, which signed on to lease 1,000 Blackwell Ultra chips from the cloud company in January of this year.

Gcore has also worked with NHN to bring its cloud platform to a data center in Incheon, South Korea, based on more than 1,000 GPUs, including the Nvidia H100s, H200s, and GB200s.

The South Korean government has teamed up with Nvidia to significantly expand AI infrastructure within the country. In October 2025, Nvidia agreed with the South Korean Ministry of Science and ICT to deploy 250,000 GPUs across the country, including up to 50,000 of the 'latest' chips across its National AI Computing Center and Korean cloud services, while the likes of Samsung, SK Group, and Hyundai will also deploy 50,000 GPUs each in planned ‘AI factories.’ Naver Cloud also said it would deploy 60,000 GPUs.

NHN Corp. was founded as a gaming company called Hangame, and launched in 1999. In 2000, Hangame was merged with Naver and changed its name to Next Human Network (NHN). In 2013, NHN was split into two, with Naver Corporation becoming a separate entity. Since 2014, NHN has focused on cloud servers and data centers. Naver itself has a cloud division as well - Naver Cloud.