SK Hyper seeks funding for $350 billion data-center plan

SK Hyper is in talks with large technology companies and financial investors to fund its planned data-center buildout.[1] It estimates the first 5-gigawatt phase will cost US$350 billion and plans a total of 15 gigawatts of capacity in South Korea by 2035.[1] The company aims to raise its first outside capital by the end of 2026 or early 2027, followed by additional fundraising over the next three to four years.[2] SK plans to share the overall cost with data-center tenants and external investors.[2]

SK Hyper says it will adjust investment according to actual demand rather than commit to the entire plan upfront.[1] It plans to begin phased operations in 2029 as it works toward the initial 5-gigawatt capacity.[3] Chung said the company could generate revenue as early as 2027, before those initial data centers are scheduled for completion.[2] The plan targets AI training and operations for Asian markets including Japan and Taiwan, rather than depending only on domestic demand.[2] Securing global demand and the capital to build the centers are described as the project’s biggest challenges.[4]

SK Hyper is developing a 900-megawatt data-center project in Ulsan alongside a previously announced 100-megawatt project for Amazon Web Services.[1] It is also in talks with South Korea’s government to secure additional sites.[1] The company plans a data-center cluster exceeding 2 gigawatts in the Yeongnam region, starting with its Ulsan facility.[4] SK Hyper is responsible for securing sites, constructing and operating substations, acquiring customers, and commercializing the data-center business.[3] SK Telecom will retain full ownership and plans to invest 750 billion won in SK Hyper by 2030.[2][3]

SK Hyper is exploring work with governments in Asia and the Middle East, offering computing capacity and AI models as a package.[2] Chung said SK’s expertise in chips, energy, and telecommunications should help offset relatively high electricity costs by improving computing efficiency.[2] The initial 5-gigawatt target is roughly five times the domestic data-center market Chung cited.[4] Meanwhile, investors are nervous that spending on future AI services may not justify investment in the global computing buildout.[1] These factors make customer demand and the staged funding plan important considerations when assessing the project.