NHN Cloud completed infrastructure with 7,656 Nvidia B200 graphics processing units at its AI-only data centre in Yangpyeong-dong, Yeongdeungpo district, Seoul, and has begun full-scale supply of GPU-as-a-service to the public sector and industry-academia-research users, it said on Tuesday.

The buildout was 추진 through the Ministry of Science and ICT and the National IT Industry Promotion Agency's programme to strengthen the foundation for using AI computing resources. NHN Cloud was selected as the largest operator, responsible for more than 1 trillion won of the total project budget of 1.46 trillion won. The infrastructure will be managed as a national asset and allocated and used under government policy.

The infrastructure configured about 4,000 GPUs into a single cluster to create a large-scale high-performance computing environment. It fully adopted a liquid-cooling system, cutting energy use by 15 to 20 percent compared with air cooling. It also has a system that monitors coolant pressure, flow and temperature in real time and automatically blocks off the affected section if an abnormality occurs.

NHN Cloud said it reflected its experience in high-density design during the buildout of its Pangyo NCC1 and its three years of operating the Gwangju National AI Data Center to improve operational stability. About 500 specialists were deployed to build a complex infrastructure including power, cooling and networks.

Of the total 7,656 GPUs, 6,120 are set to be used for national-led AI projects overseen by the ministry. NHN Cloud has been supplying some of them on a priority basis to industry-academia-research users since April 1, alongside the centre's start-up.

An NHN Cloud official said, "With the start-up of the Yangpyeong data centre, we have 마련 practical groundwork to respond to large-scale demand for AI computing." The official added, "We will advance customised GPU environments tailored to the characteristics of each industry and lead the expansion of AI services and growth of the ecosystem in Korea."