Data centers and AI city infrastructure
Straight News held the 2026 AI City Forum at the Korea Chamber of Commerce and Industry in Seoul at 9 a.m. on November 3. Speaking in the final session, titled Sustainable Growth and New Investment in AI Cities, Seo Ji-hoon, a team leader at Korea Development Bank's Industry Technology Research Center, identified data centers as core infrastructure for AI cities.
Seo emphasized that data centers, electricity and telecommunications infrastructure are becoming increasingly important to building AI cities as artificial intelligence expands into urban transport, safety, administration and energy management. In particular, Seo explained that data centers responsible for training and inference for large-scale AI models are central facilities for AI cities, and that the large investments and financing required to build them are major challenges to the expansion of AI cities.
Seo explained that AI cities are realized not merely by introducing AI services, but through coordinated integration across the value chain, connecting infrastructure and data with computing, AI models, urban operating platforms and services whose benefits residents can experience.
Within this structure, data centers are core facilities that provide large-scale computing capacity. They are essential infrastructure not only for the stable operation of AI cities but also for securing national AI competitiveness and sovereignty.
Demand for data centers is growing rapidly with the spread of generative AI. According to Korea Development Bank's presentation materials, domestic data center IT power capacity is expected to expand from 3.7 GW in 2023 to 6.2 GW in 2028. However, increasing supply faces considerable constraints because data center construction requires substantial electricity supplies, land acquisition and cooling installations.
Seo particularly identified regional disparities in AI infrastructure and imbalances in electricity supply and demand as challenges arising from the concentration of domestic data centers in the Seoul metropolitan area. Seo's position is that, for AI cities to strengthen industrial and urban competitiveness nationwide, core infrastructure including data centers must be expanded to suit regional conditions, with long-term investment funding secured to support that expansion.
Policy financing and investment methods
Seo introduced the role of policy financing, including Korea Development Bank, in building this large-scale infrastructure. The bank presented examples of approximately 2.4 trillion won in financing provided to 14 domestic data centers and outlined financing support through large-scale projects including the National AI Computing Center and the Smilegate AI Data Center.
The presentation also emphasized that financing AI data centers requires a different approach from conventional financing focused on real estate development. Data centers are complex infrastructure projects that must secure not only buildings and land but also expensive core computing equipment such as GPUs, reliable electricity supply networks and cooling facilities at the same time. It is therefore important to combine financing methods such as direct investment, leasing, joint ventures (JVs), project finance (PF) and private lending, taking account of project characteristics and investment risks.
Support for AI infrastructure through the National Growth Fund and direct investment in companies across the AI value chain were also presented as major roles for policy financing. The intention is for policy financing to stimulate AI infrastructure investment by complementing private finance in meeting large upfront investment needs and long capital-recovery periods that private finance alone finds difficult to accommodate.
Regional projects and risk management
Regional project concepts for expanding AI infrastructure outside the Seoul metropolitan area were also introduced. These involve establishing AI data centers and industrial hubs in regions including Gangwon, Chungcheong, Honam and Gyeongsang, and linking AI computing infrastructure with advanced industries such as semiconductors, robotics and autonomous driving by using regional electricity, land and water conditions. The concept is to develop data centers not merely as computing facilities but as core hubs supporting regional industrial ecosystems.
Moderator Go Ran asked about the central challenges of investing in AI infrastructure, which requires vast amounts of capital. Korea Development Bank team leader Seo replied: “There is a time lag before added value is generated, so it is important to maintain appropriate valuations through sound risk management until then.”
The future role of policy financing is drawing attention because the practical competitiveness of AI cities depends not only on the quality of the AI services they adopt, but also on how much data center capacity they secure to run those services reliably and how they raise the financing required for substantial infrastructure investment.
Original Korean article: Straight News
English brief by Korea Wind Intelligence, koreawindintel.com
