Naver expands AI infrastructure supply
South Korea's leading cloud service providers (CSPs) are continuing to grow through their artificial intelligence (AI) infrastructure businesses. Growing demand for graphics processing unit (GPU) services and data centers is improving financial results, while providers are accelerating large-scale infrastructure investment and competition for customers.
According to industry sources on November 16, the three major domestic CSPs, Naver Cloud, kt cloud and NHN Cloud, reported revenue growth in the third quarter of this year. They are strengthening their ability to supply AI infrastructure and seeking new revenue sources by expanding AI data centers and introducing next-generation GPUs.

Naver's enterprise segment posted third-quarter revenue of 193.3 billion won, up 33.0% year on year and 25.8% from the previous quarter. Business-to-business operations, including a digital twin project in Saudi Arabia, contributed to growth.
Naver Cloud is developing its global AI factory business as a new growth driver beyond its existing cloud operations. It recently secured an initial customer by signing a computing services supply contract worth approximately 2 trillion won with a US AI inference company, including next-generation GPUs.
Naver is working with global asset manager Brookfield and Nvidia to expand its AI infrastructure supply base. It plans to start with 55 megawatts (MW) in 2027 and expand supply capacity to 200 MW by 2028. Related revenue is expected to begin contributing to results progressively from the first half of 2027.
Naver chief executive Choi Soo-yeon said during a conference call, "We plan to secure both stability and high profitability by pursuing medium- and long-term contracts with large customers while expanding our customer base to global AI companies with high growth potential."
KT plans additional data center capacity
kt cloud also continued to improve its results, supported by growth in data centers and public-sector cloud services. Third-quarter revenue rose 7.7% year on year to 268.2 billion won. Growth was supported by higher colocation revenue as utilization increased at the Gasan data center, demand for cloud migration among public institutions and high-performance computing support programs.
KT plans to substantially expand its infrastructure business around AI data centers. In addition to the 152 MW currently in operation, it aims to secure an additional 1 gigawatt (GW) of AI data center capacity by 2031, based on actual demand. It plans to invest about 5 trillion won in the related business over the next five years.
To that end, KT has secured a project pipeline of 911 MW across 20 primary sites and 500 MW across 10 reserve sites. Investment decisions have been completed for 200 MW at the primary sites, with a further 60 MW expected to be finalized by the end of 2026. kt cloud also plans to participate on a design-build-operate (DBO) basis in AI data center projects of 60 MW in Gunsan, North Jeolla Province, and 50 MW in Cheongna, Incheon.

Kim Bong-kyun, chief executive of kt cloud and head of KT's enterprise division, said during a conference call, "We currently operate 152 MW and plan to secure an additional 1 GW based on actual demand. We plan for 80% of the 1 GW capacity to be leased and 20% to be our own."
NHN Cloud expands GPU services
NHN Cloud recorded a high revenue growth rate among the three providers, supported by the expansion of its GPU services business. Consolidated third-quarter revenue rose 75.5% year on year and 10.3% from the previous quarter. It also remained profitable as AI GPU service revenue increased. NHN's technology segment revenue rose 53.6% year on year to 171.7 billion won.
In particular, it sold all GPU capacity available to external customers at FactoryX Seoul in Yangpyeong, Seoul, for supply over the next five years. The facility is operated through a government program to strengthen the foundations for the use of AI computing resources. It houses 7,656 Nvidia B200 GPUs. In the private sector, revenue from building and operating Krafton's GPU cluster began contributing in earnest, supporting growth.
To meet growing GPU demand, NHN Cloud plans to operate a total of 100 MW of AI data centers, including existing facilities, by the end of 2027. It is pursuing leases for a 20 MW facility within the Pohang AI data center and a 26 MW data center in Jangseong, South Jeolla Province, and plans to build its own 10 MW data center in the Gwangju Research and Development Special Zone.
In the Seoul metropolitan area, it is also securing additional data centers in Gasan, Seoul, and Gajwa, Incheon, to install more than 1,700 Nvidia B300 GPUs. It plans to launch related services progressively from 2027 and secure liquid-cooling equipment in preparation for the introduction of next-generation Vera Rubin GPUs.
The expansion of AI infrastructure businesses by South Korea's three CSPs is expected to affect competition in the cloud market. As they step up efforts to secure large-scale data centers and GPUs, maintaining stable customer demand and recovering investment costs are emerging as key challenges for medium- and long-term growth.
NHN chief executive Jung Woo-jin said, "NHN Cloud will firm up its plans to secure AI data center infrastructure and obtain additional data center space to become an operator running approximately 100 MW of data centers by the end of 2027. As we are steadily securing private-sector customers following government projects, we will continue to strengthen our market competitiveness."
Original Korean article: ZDNet Korea
English brief by Korea Wind Intelligence, koreawindintel.com















