[Edaily Reporters Yoon Jeong-hoon and Lee So-hyun] Leading domestic ICT companies #SK Telecom and #NAVER have joined forces with global AI semiconductor giant NVIDIA to compete for leadership in next-generation AI infrastructure. As Korean companies have been selected as key partners in NVIDIA’s “AI Factory” strategy—which goes beyond the simple construction of data centers—competition between the two companies is expected to intensify.

On the 8th, NVIDIA CEO Jensen Huang met successively with SK Group Chairman Choi Tae-won and Naver Board Chairman Lee Hae-jin in Seoul to unveil plans to build gigawatt (GW)-scale AI factories in Korea based on NVIDIA’s AI infrastructure platform, DSX.

According to industry sources, NVIDIA has recently been pursuing its “AI Factory” business—which goes beyond chip sales to provide the entire AI computing infrastructure—as a global strategy. In particular, the company has been seeking strategic footholds in the Asian market, where demand for data centers is surging, and ultimately chose South Korea.

NVIDIA has been discussing AI Factory collaboration plans with domestic companies. Naver has been in early discussions with NVIDIA regarding the construction and utilization of AI infrastructure and has been making related preparations. The company aims to strengthen the competitiveness of its AI services, including its own hyper-scale AI model, HyperClova X, as well as enhance cooperation in the AI cloud business.

SK Telecom (SKT) has also recently signed a Letter of Intent (LOI) with NVIDIA, officially joining the AI Factory initiative. SKT is cultivating its AI data center and cloud businesses as future growth engines and is exploring opportunities to enter the global AI Factory market through its partnership with NVIDIA.

NVIDIA also plans to make significant investments in the SKT and Naver AI Factories, which serve as its Asian hubs. With both companies participating in NVIDIA’s AI Factory strategy, competition for dominance in the domestic AI infrastructure market is expected to intensify.

Transforming from a Telecom Provider to an ‘AI Cloud Provider’… Targeting 2027 Launch

SK Telecom presented a blueprint to use this partnership as a springboard to leap beyond its role as a traditional telecommunications network operator and become a global AI cloud provider encompassing both agentic AI and physical AI.

SKT plans to build a gigawatt (GW)-scale AI cloud in Korea based on NVIDIA’s DSX full-stack reference architecture and launch its first AI factory in 2027. However, since securing land and building power infrastructure for new AI data centers (AIDCs) requires significant time, the company is expected to initially advance the project by upgrading existing facilities such as the Gasan AIDC.

SKT will also participate in NVIDIA’s global partner program, “NVIDIA Cloud Partner (NCP).” Through this, the company aims to reduce AI service operating costs and secure competitiveness by leveraging cutting-edge AI infrastructure and the software ecosystem.

The two companies will go beyond building AI infrastructure to jointly research next-generation AI factory architectures. The plan is to optimize the entire AI infrastructure by combining SK Telecom’s network operation capabilities, SK Hynix’s HBM (High Bandwidth Memory) technology, and the power infrastructure capabilities of group affiliates.

A prime example is a project that applied digital twin technology to SK Hynix’s semiconductor production facilities using NVIDIA’s digital twin platform, Omniverse.

They are also expanding cooperation in the field of robotics. SK Telecom is working to advance robot simulation and training platforms by utilizing NVIDIA’s robotics platform “Cosmos” and the humanoid AI model “Isaac Groot.”

Jensen Huang, CEO of NVIDIA, explained, “Just as electricity, water, and the internet transformed nations in the past, AI will become the driving force of nations in the future,” adding, “Through our partnership with SKT, we will build essential infrastructure that universities, research institutes, and startups can freely utilize.”

SK Group Chairman Choi Tae-won said, “We have secured full-stack AI infrastructure capabilities ranging from semiconductor chips to data center operations,” adding, “Together with NVIDIA, we will solve challenges in the fields of GPUs, memory, and energy, and grow into an AI factory company that leads the Asian AI ecosystem.”

Naver Expands AI Infrastructure Business Based on ‘Gak Sejong’

Through its collaboration with NVIDIA

,

Naver plans to establish an NVIDIA AI Factory in Korea. This initiative aims to enhance the competitiveness of its proprietary large language model (LLM), ‘HyperClova X,’ and accelerate its entry into the AI agent market, which is considered a next-generation AI service. In particular, the company intends to meet domestic AI computing demand while also expanding into the Middle East.

To this end, Naver will utilize ‘Gak Sejong,’ the country’s largest hyperscale data center, as a key hub for its global AI infrastructure expansion. By implementing NVIDIA’s DSX platform, the company plans to first launch a 55MW AI infrastructure in the first half of 2027 and expand it to 200MW by 2028. In the long term, it also aims to build a gigawatt (GW)-class mega AI factory based on this foundation.

Jensen Huang, CEO of NVIDIA, stated, “Naver is Korea’s leading cloud provider,” adding, “Based on our long-standing partnership, we will build large-scale AI cloud infrastructure in Korea.” Lee Hae-jin, Chairman of Naver’s Board of Directors, remarked, “Naver was the first company to build a SuperPod using NVIDIA GPUs,” and expressed his expectation that “the two companies will join forces to achieve meaningful results.”

The two companies will expand their collaboration not only in infrastructure but also in the AI software ecosystem. Naver recently became the first Korean company to join NVIDIA’s global AI alliance, the “Nemotron Alliance,” and is using this to strengthen the global competitiveness of HyperClova X.

They are also pursuing collaboration in the field of spatial intelligence. Naver is working to build the “Seoul World Model,” a virtual spatial simulation, by combining its spatial modeling technology and street view data with NVIDIA’s world foundation model, Cosmos.

Chairman Lee stated, “Through this collaboration, we are now able to present a new alternative that allows countries and companies to build their own independent AI capabilities,” adding, “This will serve as an opportunity to expand Naver’s AI technology and infrastructure competitiveness into the global market.”

Meanwhile, SK Telecom and Naver, which announced their collaboration with NVIDIA and AI Factory on the same day, were the only two stocks among the top 50 on the KOSPI by market capitalization to close higher. SK Telecom closed at 106,700 won, up 0.2% from the previous day, while Naver closed at 279,000 won, up 9.2%.