Private-sector-led AI infrastructure investment combining financial and industrial capital
Companies use fund contributions to expand exports in their core businesses
Kyushu among candidate locations with power and industrial infrastructure
Samsung, SK and LG link AI investment to their core businesses. According to business community sources on November 1, major South Korean companies are already connecting AI infrastructure investment with market expansion in their existing businesses. In September, six companies, Samsung Electronics, Samsung C&T, Samsung SDS, Samsung SDI, Samsung Life Insurance and Samsung Fire & Marine Insurance, agreed to invest a combined $1 billion, or about 1.3358 trillion won, in US AI infrastructure company Helix. Helix develops data centers together with power and telecommunications infrastructure.
Through this investment, Samsung plans to combine its affiliates' capabilities in semiconductors, cooling equipment, construction, operations and batteries. Its strategy is to invest in an AI infrastructure company to secure opportunities not only to supply components and equipment but also to participate in facility construction and operations.
SK Group is pursuing an AI data center in Ulsan with Amazon Web Services, or AWS, targeting initial operation in 2027. The project combines SK Telecom and SK Broadband's construction and operating capabilities, the group's energy and telecommunications infrastructure, and AWS's cloud capabilities. Lee Sang-heon, an analyst at iM Securities, said, "As SK is also pursuing large-scale data center construction in South Korea, it could approach this as an opportunity to expand its business overseas."
LG is also supplying a combination of LG Electronics' cooling technology, LG Energy Solution's batteries and LG CNS's design, construction and operating capabilities to data centers in Indonesia. LG Electronics is responding to AI data center cooling demand through investment totaling 150 billion won, including a new chiller plant in Virginia in the United States and production line expansions in Pyeongtaek and Changwon in South Korea.
Opportunities extend beyond construction to operations and use, with location also drawing attention. Companies could seek opportunities to participate in the Korea Investment-SBI joint fund based on their respective capabilities. Lee said, "If South Korean companies participate in a joint South Korea-Japan data center project, they could do so through equity investment or as limited partners in a fund. Samsung C&T could participate in data center construction, Samsung SDS could take part in related business, and Samsung Electronics could also approach it as an opportunity to use data centers in Japan in the future."
Once facilities begin operating, operations and computing services also generate revenue. In a recent report on Samsung SDS, iM Securities categorized data center investment approaches as direct asset ownership, equity investment, and design, construction and operations for customers. Direct asset owners earn revenue by providing server space, power and cooling or renting out GPU computing resources. Although this approach requires substantial investment, returns can grow depending on facility utilization and service demand.
Securing computing resources can also be an important investment objective for companies that use data centers. Lee said, "Samsung Electronics also needs to use data centers for areas such as physical AI. Investment in Japanese data centers can likewise be viewed as an extension of these activities."
The location of the data center facilities is also attracting attention. Access to enormous amounts of power is an important location requirement for large-scale AI computing facilities.
Kyushu, one of the candidate locations for this project, has a base of nuclear and renewable power generation. It also hosts production facilities for semiconductor companies, including TSMC subsidiary JASM in Kumamoto, Sony and Mitsubishi Electric. The concentration of advanced industries is increasing as TSMC's arrival has prompted investment by related companies and the development of supply chains.
In a February report, Mitsubishi UFJ Trust and Banking said, "Facilities for AI training face relatively fewer constraints from communications latency and can therefore also be located in regional areas where power is easier to secure."
Original Korean article: Financial News
English brief by Korea Wind Intelligence, koreawindintel.com
