Table of Contents
SK Telecom cited continued supply constraints in the global AI infrastructure market as a key driver behind the initiative
In sum – what to know:
15GW AI infrastructure target – SK Telecom plans to develop up to 15GW of AI data center capacity, with the first 5GW scheduled to come online in phases beginning in 2029.
Regional AI expansion – The company will expand its Ulsan AI data center into a multi-gigawatt cluster while adding new capacity in southwestern Korea to support national AI infrastructure.
Full-stack AI approach – SK Group plans to combine semiconductor, power, and data center expertise across its affiliates to support large-scale AI computing infrastructure.
Korean carrier SK Telecom (SKT) has unveiled plans to develop up to 15GW of AI data center capacity in South Korea, outlining one of the largest AI infrastructure initiatives announced in Asia as demand for AI model training and inference continues to accelerate.
The company said the initiative is intended to expand domestic AI computing infrastructure while positioning South Korea as a regional hub for AI infrastructure. SKT is aligning the project with the government’s AI G3 strategy, which aims to make Korea one of the world’s three leading AI nations alongside the United States and China.
The first phase centers on the AI data center currently under construction in Ulsan. SKT plans to expand the site into a cluster exceeding 2GW across the Gyeongsang region while developing an additional 1GW facility in the Jeolla region. Combined, these projects will contribute to an initial 5GW of AI data center capacity that the carrier expects to activate in stages beginning in 2029.
The longer-term objective is to expand total capacity to 15GW. SKT noted that building AI infrastructure at this scale will require substantial investment. Funding is expected to combine company investment with strategic partners, long-term customer agreements, and project financing.
The company cited continued supply constraints in the global AI infrastructure market as a key driver behind the initiative. According to SKT, global consulting firm McKinsey & Company forecasts annual data center demand growth of 19% to 22%, with supply expected to lag demand, resulting in an estimated 15GW shortfall in the U.S. market by 2030.
SKT also pointed to South Korea’s competitive position in AI infrastructure, highlighting its leadership in high-bandwidth memory (HBM), stable electricity supply supported by nuclear, and liquefied natural gas generation, and experience operating gigawatt-scale semiconductor manufacturing facilities.
The company said site selection, power availability and securing anchor tenants will all be considered as part of the deployment strategy.
Rather than relying solely on its telecom business, SKT plans to leverage capabilities across SK Group affiliates. The company said the project will combine semiconductor technologies, energy solutions, and data center construction and operational expertise to support AI infrastructure deployment.
SK Telecom will lead the overall design, construction, and operation of the facilities. The company has previously outlined plans to launch an “AI Factory” in 2027, describing it as a next-generation AI data center platform that it intends to expand to gigawatt scale over time.
The broader initiative is also intended to support regional economic development by locating AI infrastructure outside the Seoul metropolitan area and connecting future facilities with local industries.
“This AI data center project is aimed at preemptively preparing the computing infrastructure that the global AI ecosystem needs,” said Jung Jai-hun, president and chief executive officers at SKT. “We will work closely with the government, industry, and local communities to help Korea grow into Asia’s core AI infrastructure hub.”
The latest announcement builds on a partnership unveiled by SK Telecom and Nvidia in June, when the companies revealed plans to develop a gigawatt-scale AI Cloud in South Korea based on Nvidia’s DSX AI factory architecture. The first AI factory is scheduled to begin operations in 2027.
According to the companies, the AI Cloud will provide GPU-based infrastructure for AI model training, inference, and agentic AI, supporting sovereign AI initiatives as well as enterprise and industrial applications. SK Telecom also said the platform is intended to expand beyond South Korea to serve other Asian markets over time.
Beyond deploying infrastructure, Nvidia and SK Group announced plans to jointly research next-generation AI factory architectures, focusing on areas including accelerated computing, memory technologies, and data center operations to improve the scalability and efficiency of future AI infrastructure.
