Following the United States, China is also experiencing a funding drought for artificial intelligence (AI) data centers (DCs) due to increasing financial risks associated with investment debts. In South Korea, major corporations are unveiling gigawatt (GW) scale AIDC plans, but with an additional 600 trillion won needed, the drying up of global funding is causing disruptions to initial plans.
According to investment banking sources on October 26, KT is in discussions with a global investment firm to secure up to 35 trillion won for AIDC investments.
KT reportedly plans to rely on external funding for over 80% of its 1GW AIDC construction, with internal cash contributions not exceeding 3 trillion won. The total cost for building a 1GW AIDC is estimated at around 87 trillion won. KT faces the challenge of raising over 80 trillion won, with few investors showing interest beyond one firm, and no specific amounts or timelines have been confirmed with the company under discussion.
Initially, KT's AIDC funding plan was expected to be announced in October, but it has now been postponed indefinitely. An internal source at KT indicated that while construction is planned to proceed next year, completing the 1GW project as originally intended is deemed impossible. Previously, SK Telecom's 8 trillion won Ul-san AIDC began construction in August last year, but it took a year to finalize its funding plan.
KT is not the only company struggling with AIDC funding. SK hynix needs to secure approximately 461 trillion won for its first phase plan of 5GW, while Naver must raise an additional 18 trillion won for a 200MW project. LG Uplus has only secured 2 trillion won of the 18 trillion won needed for its 200MW project, and Kakao has raised about 600 billion won of the 7 trillion won required for its 80MW project.
While the required funding reaches hundreds of trillions of won, the AI infrastructure market is tightening as financial risks grow. Morgan Stanley and Moody's estimate that around $3 trillion (approximately 3,964 trillion won) will be needed to build the AIDCs planned by global companies. A significant portion of this will need to be covered by bank loans and bonds, but it is increasingly difficult to find AI companies generating profits that exceed their capital expenditures (CAPEX). Last year, Alphabet, Amazon, Meta, and Oracle borrowed a total of $93 billion in the U.S. investment-grade bond market. As global funding and bonds concentrate on U.S.-centered AIDC construction, attracting investment for domestic companies is becoming increasingly challenging.
Securing funding from China and Singapore has also become difficult. The CSI AI index in China has fallen nearly 40% from its peak in June, and in Hong Kong, over 60% of newly listed companies in the third quarter traded below their offering prices, freezing AI investment sentiment across Asia. Singapore's sovereign wealth fund Temasek identified AI valuation as a major risk factor for the global market next year on October 7.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.