In a bold move to position South Korea as a leader in artificial intelligence, SK Group and Amazon Web Services (AWS) have announced plans to build the largest AI data center in South Korea. The multibillion-dollar project will feature 60,000 GPUs and will be located in Ulsan’s Mipo National Industrial Complex.
Massive Investment in AI Infrastructure
The new facility will begin construction this August, starting with a 41 MW capacity by 2027 and scaling up to 103 MW by 2029. With long-term ambitions to expand to 1 gigawatt, the center is expected to become a major AI computing hub for Northeast Asia.
SK Group is partnering with AWS on this high-stakes project. AWS alone is contributing $4 billion, while SK Telecom, SK Broadband, SK Gas, and SK Hynix are jointly investing 3.4 trillion won ($2.5 billion) into AI initiatives by 2028.
Check out our BLOG on: Valves and Actuators the Silent Enablers of Data Centers
Strategic Location and Energy Supply
To ensure reliable power for the energy-intensive operations, the facility will be built near SK Gas’ LNG combined-cycle power plant. This proximity guarantees a stable energy source—vital for a center demanding nearly 100 MW, equivalent to a small power station.
South Korea’s AI Ambitions
With only 15,000 GPUs planned at the government’s National AI Computing Center, SK’s installation dwarfs public efforts, highlighting the company’s intent to lead in AI infrastructure. SK Telecom also plans to offer specialized AI data services for businesses of all sizes.
A Growing Turf War in Korea’s Cloud Market
AWS is aggressively expanding in South Korea, adding this new center to existing ones in Seoul and Incheon. The country’s cloud and AI infrastructure market is heating up, with local tech giants like Naver, Kakao, and NHN competing alongside global players like Google, Microsoft, and Alibaba.
This new partnership between SK and AWS marks a pivotal step for South Korea in the global race for AI supremacy—setting the stage for a digital transformation powered by high-performance computing.