SKT, Once a Mobile Carrier, to Collect 'Server Rent' in the AI Era [Weekend Money]

Leasing Everything from Space to GPU Computing Power
Securing Electricity and Long-term Clients Is Key

Even artificial intelligence (AI) needs an office to work. While AI answers questions and creates images, countless servers run behind the scenes, supported continuously by electricity and cooling systems. The business of providing and leasing this space is called the data center industry. SK Telecom, familiar to many as a mobile phone service provider, is now gaining attention in the AI era as a company that also collects ‘server rent’.

On the 18th, Mirae Asset Securities forecast that SK Telecom would be a major beneficiary as domestic data center expansion accelerates. The reasons cited include the company’s ability to scale both space leasing and computing power rental. SK Telecom’s experience operating telecommunications networks is evaluated as a strength, along with the semiconductor, energy, and construction capabilities found within the group.

There are two main revenue streams. 'Colocation' refers to providing clients with server space, power, and a cooling environment. 'GPUaaS (GPU as a Service)' rents out the computational power of graphic processing units (GPUs) needed for AI learning and services. In effect, SK Telecom is in the business of both providing the digital ‘offices’ and power, as well as high-performance computing resources to its customers.

Revenues are already increasing. Last year, SK Telecom’s AI data center sales reached 519.9 billion won, a 34.9% increase from the previous year. This was driven by higher utilization rates at the Gasan and Yangju data centers and the impact of acquiring the Pangyo data center. Even before constructing new buildings, increased occupancy at existing facilities is boosting revenue.

SK Horizon is the new corporation being established by splitting off SK Broadband’s data center and submarine cable businesses. The company plans to expand by dividing roles with SK Hyper, which will be in charge of developing large-scale new data centers. It is anticipated that the operational experience and customer base from existing centers will be leveraged for these new facilities.

Group-level collaboration is also expected. This is because data centers must reliably secure large amounts of electricity and provide efficient server cooling. This context highlights the importance of SK hynix’s AI semiconductors, SK Gas’s energy business, and SK ecoplant’s construction competency. It is analyzed that such a division of labor can reduce setup costs and shorten preparation time for facilities.

Of course, simply constructing data centers does not guarantee a steady stream of rent. The key factors are whether sufficient power has actually been secured and how many customers commit to long-term contracts. However, the outlook is that if SK Telecom succeeds in acquiring a significant number of new clients, its corporate value will be clearly re-evaluated.

Choi Yujin, a researcher at Mirae Asset Securities, commented, "There is potential for synergy, as SK Horizon can leverage its existing data center operational capabilities and customer base in the development of new, large-scale AI data centers with SK Hyper." She added, "SK Telecom will be able to maintain stable cash flows from its existing data center business while also securing growth options to meet large-scale AI computing demand."

This content was produced with the assistance of AI translation services.