 |
Photo courtesy of NVIDIA |
[Alpha Biz= Kim Jisun] Sejong, South Korea — July 29, 2025 — The Ministry of Science and ICT (MSIT) has announced the selection of Naver Cloud, NHN Cloud, and Kakao as partners for a large-scale project to secure next-generation graphics processing units (GPUs) to strengthen Korea’s AI computing infrastructure.
Backed by a ₩1.46 trillion (approx. USD 1.1 billion) supplementary budget approved in May 2025, the initiative will acquire more than 13,000 NVIDIA GPUs, including 10,080 B200 units and 3,056 H200 units.
Key Allocations
NHN Cloud: 7,656 B200 GPUs
Kakao: 2,040 B200 GPUs
Naver Cloud: 3,056 H200 GPUs
Approximately 10,000 GPUs will be used for government AI research and national projects, while the remainder will be deployed for operational support by the partner companies.
Building a National AI Infrastructure
The three selected companies will jointly develop a “GPU Unified Support Platform”, a centralized system through which universities, research institutes, and private companies can apply online for GPU resources. Applications will be evaluated and allocated based on project needs and national priorities.
Background
Four companies submitted proposals, including Coupang, which was not selected. The decision was based on the applicants’ ability to operate large-scale GPU clusters and support the government’s AI initiatives.
Government Statement
Minister of Science and ICT Baek Kyung-hun stated:
“This procurement of advanced GPUs is the catalyst for addressing Korea’s AI computing shortage. It marks the starting point for building an AI superhighway and positioning South Korea as a global AI powerhouse.”
Next Steps
The MSIT plans to finalize agreements with the selected companies in early August 2025, with phased GPU deployment and support for key AI projects beginning later this year.
Alphabiz Reporter Kim Jisun(stockmk2020@alphabiz.co.kr)
