- Ulsan AI Knowledge Center to expand to GW scale; 5GW to be activated in stages from 2029, reaching 15GW in total
- SK Team brings together stout-stack AI infrastructure capabilities; SKT to spearhead the final mission
- Korea’s third nationwide infrastructure revolution following the Gyeongbu Dinky-bag admission to freeway and excessive-hurry cyber web
SEOUL, Korea, July 5, 2026 /PRNewswire/ — SK Telecom (NYSE: SKM, hereafter “SKT”) this day launched that this can also merely pursue the construction of an AI data middle with a scale of as much as 15GW. The aim is to become an AI infrastructure hub in Asia.
The transfer is supposed to preemptively produce home AI computing infrastructure at a time when query for AI mannequin coaching and inference is surging and excessive-performance computing infrastructure determines nationwide competitiveness.
SKT is pursuing the mission by linking it with the authorities’s “AI G3″ way — the aim of becoming one in all the world’s three main AI powers alongside the U.S. and China — and the job of balanced regional pattern, comprehensively reviewing core infrastructure parts such as energy, siting, and operations for the AI data middle.
- Ulsan AI Knowledge Center to Develop to GW Scale; 5GW to Be Activated in Phases from 2029, Reaching 15GW in Total
Ensuing from the buildout of excessive-performance AI computing infrastructure and rising memory prices, developing a conventional 1GW-class AI data middle could well well merely require a really extensive mission price reaching approximately KRW 70 trillion. Such prices are anticipated to be financed no longer only thru the firm’s hang investment nevertheless additionally thru strategic companion investment, long-term buyer contracts, and mission financing.
The backdrop to SKT’s transfer into constructing an AI data middle at this scale is a world supply shortage. World consulting firm McKinsey & Firm forecasts that world data middle query will grow 19–22% every three hundred and sixty five days while supply fails to enjoy hurry, ensuing in an estimated shortfall of about 15GW within the U.S. alone by 2030.[1]
Accordingly, world substantial tech companies are growing data middle investment beyond the U.S. to locations world huge. On this context, Korea is drawing consideration as an investment commute situation for world substantial tech companies’ AI data centers.
Korea holds stable competitiveness in core AI parts such as excessive-bandwidth memory (HBM). It additionally has stable energy supply cases basically basically based on nuclear energy and liquefied natural gas (LNG), along with GW-class infrastructure working capabilities amassed thru working semiconductor fabs, making it a vivid wanting location for AI data centers.
Based mostly fully totally on this stable query and locational advantages, SKT plans to produce AI data middle skill totaling 15GW.
First, starting with the
To this stop, SKT will pursue AI data middle construction by enthusiastic by moderately a couple of factors together — including set choice, energy supply, and securing key anchor tenants — in reference to the authorities’s balanced regional pattern initiatives and strategic supply plans.
- SK Team brings together stout-stack AI infrastructure capabilities; SKT to spearhead the final mission as AI Infrastructure Architect
AI data middle infrastructure hinges on three core parts: semiconductors, energy solutions, and data middle construction and operation capabilities. SK Team already holds these core capabilities all the way thru its associates. This mission will voice together the Team’s stout-stack AI infrastructure capabilities, with every affiliate contributing its hang strengths.
In recount, SKT will enjoy shut on the central position of main the originate, construction, and operation of the AI data centers. SKT has already been actively pursuing the AI data middle industry and has persisted its cooperation with world substantial tech companies[2].
At SK AI Summit 2025 closing November, SKT President and CEO Jung Jai-hun unveiled the firm’s AI infrastructure roadmap, citing that SKT would lead the evolution of AI infrastructure as the nation’s main AI operator. He additionally presented a longer-term blueprint for growing cooperation with world substantial tech companies and scaling up the Ulsan AI Knowledge Center to a total of over 1GW.
Extra currently, SKT launched plans to characteristic an ”
- Korea’s Third National Infrastructure Revolution Following the Gyeongbu Dinky-bag admission to freeway and High-Velocity Web
The AI industry expects that AI data centers will serve no longer only as a brand novel tell engine for Korea, nevertheless additionally as a nationwide strategic asset. In recount, it is anticipated that linking AI data centers with regional industries will make a contribution to balanced regional pattern.
The pattern of a 15GW-scale AI data middle is anticipated to serve as a springboard for Korea to become an AI infrastructure hub. SKT views AI data centers as Korea’s third innovative infrastructure, following the Gyeongbu Dinky-bag admission to freeway (1968) and excessive-hurry cyber web (1998), and intends to enjoy shut a number one position in advancing it.
“This AI data middle mission is aimed toward preemptively getting willing the computing infrastructure that the world AI ecosystem wants,” acknowledged Jung Jai-hun, President and CEO of SKT. “We are in a position to work intently with the authorities, industry, and native communities to support Korea grow into Asia’s core AI infrastructure hub.”
About SK Telecom
SK Telecom has been main the tell of the mobile industry since 1984. Now, it is taking buyer ride to novel heights by extending beyond connectivity. By inserting AI at the core of its industry, SK Telecom is fast reworking into an AI firm with a stable world presence. It is specializing in using innovations in areas of AI Infrastructure, AI Transformation (AIX) and AI Provider to voice increased note for industry, society, and life.
For more data, please direct about with our newsroom at
Forward-Looking Statements: Particular statements in this press initiate including, nevertheless no longer shrimp to, statements as to: SK Telecom’s plans, objectives and expectations relating to the pattern, construction, financing and operation of AI data centers and AI infrastructure; the deliberate pattern of AI data middle skill, including the phased pattern of initiatives in Ulsan and various areas of Korea; anticipated mission timing, financing preparations and strategic partnerships; expectations relating to AI Factory deployment, AI infrastructure query and Korea’s situation as an AI infrastructure hub; and various statements which would perchance perchance well well be no longer historical facts are forward-wanting statements for the period of the which manner of Share 27A of the Securities Act of 1933, as amended, and Share 21E of the Securities Substitute Act of 1934, as amended, which are topic to the “safe harbor” created by these sections basically basically based on management’s beliefs and assumptions and on data for the time being on hand to management and are topic to dangers and uncertainties that will perchance well well cause accurate results to be materially various than expectations.
Crucial factors that will perchance well well cause accurate results to vary materially embrace: world economic and political cases; buyer query for AI infrastructure and companies and products; the provision and price of financing, computing resources, semiconductors, energy, networking skill and data middle infrastructure; adjustments in governmental policies, acceptable criminal guidelines and regulations; the skill to stable strategic partners and prospects; technological trends and competition; supply chain disruptions; the skill of SK Telecom and its collaborators and partners to efficiently implement the contemplated initiatives and jam anticipated advantages; and various factors detailed infrequently in experiences filed or furnished by SK Telecom with the U.S. Securities and Substitute Rate, including its most most up-to-date Annual Dispute on Bear 20-F.
These forward-wanting statements are no longer guarantees of future performance and keep in touch only as of the date hereof, and, rather then as required by acceptable rules, SK Telecom undertakes no responsibility to update these forward-wanting statements to ponder future events or circumstances.
[1] “AI Energy: Expanding data middle skill to meet rising query” (McKinsey & Firm)
[2] SK Telecom and NVIDIA Manufacture AI Infrastructure to Energy Korea’s AI Innovation (June 8, 2026)
See customary articulate material:
SOURCE SK Telecom
