[Edaily Reporter Kim Hyun-ah ] NVIDIA, a leader in the global race for artificial intelligence (AI) supremacy, has designated South Korea as a key AI hub in the Asia-Pacific (APAC) region.

#While SK Group and Naver (NAVER(035420)) have joined forces with NVIDIA to enter the AI data center (AIDC) race, the scale of their investments and business strategies differ markedly.

SK’s strategy is to leap forward as an “AI infrastructure provider” by building an AI factory with a capacity of up to 2 GW (gigawatts) to supply AI computing resources to global companies and nations. In contrast, Naver plans to start with a 200 MW AIDC and expand into an “enterprise AI platform provider” that integrates its own AI models, cloud services, and enterprise solutions.

The difference in investment scale also stems from the nature of their businesses. SK’s AI factory is a massive project centered on the construction of large-scale physical infrastructure—including GPUs, data centers, power, and networks—requiring an investment of at least $100 billion (approximately 146 trillion won). Naver’s $10 billion (approximately 14.6 trillion won) investment is focused on securing data centers to operate AI services and strengthening the competitiveness of its platform.

Industry observers view this collaboration as an opportunity for South Korea to emerge as a key pillar of the global AI supply chain. Under this structure, while SK builds the “computing body” of the AI era, Naver will serve as the “platform brain,” leveraging it to implement industry-specific AI services.

SK to Build 2GW AI Factory… Leaping Forward as a Global AI Computing Provider

The scale of the partnership between SK and NVIDIA

,

recently announced in San Francisco

,

exceeds $500 billion (approximately 731 trillion won). However, this figure represents the scope of a long-term partnership that encompasses the construction of AI factories, next-generation memory (HBM), the GPU ecosystem, and AI services.

SK’s partnership with NVIDIA is interpreted as a strategy to leap forward as an AI computing infrastructure provider, going beyond simply expanding HBM supply.

An SKTelecom official explained, “The $500 billion figure represents the total scope of the partnership, including NVIDIA GPUs, memory, AI factories, and a diverse partner ecosystem,” adding, “Looking solely at the AI factory sector, it is a project requiring an investment of at least $100 billion.”

A 2GW AI data center is a massive infrastructure project valued at approximately 140 trillion to 150 trillion won as of now. SK plans to expand this beyond meeting domestic demand to a business that supplies AI computing resources to global markets, including Southeast Asia.

SK hynix(000660)The strategy is to establish itself as a global AI infrastructure provider by combining SK’s HBM competitiveness, SKTelecom(017670)’s experience in operating AI data centers, and the group’s energy and infrastructure capabilities.

SKTelecom has also launched a specialized subsidiary, “SK Hyper,” to enhance the execution capabilities of its AI data center business. SK Hyper is an operational organization responsible for all aspects of the AI data center business, including securing sites, building power infrastructure, acquiring customers, and developing new business opportunities.

SKTelecom plans to first establish a foundation for an AI data center business with a capacity of 5 GW centered on SK Hyper, and then expand it to a total of 15 GW by 2035, depending on market conditions. The 2 GW-class AI factory being developed in partnership with NVIDIA is a core pillar of this mid- to long-term AI infrastructure strategy.

SK’s AI factory aims to be more than just an in-house AI development facility; it seeks to become a global computing platform in the form of GPUaaS (GPU as a Service), providing GPU computing resources to companies and institutions.

An SKTelecom official emphasized, “The core of sovereign AI lies not only in the models but also in where the data and AI infrastructure are located,” adding, “It is crucial to secure a foundation that allows us to produce and service AI within our own country.”

Jensen Huang, CEO of NVIDIA, also spoke highly of SK’s role. CEO Huang emphasized that SK hynix’s competitiveness in HBM is a key foundation enabling the current development of the AI industry, and highlighted that SK is a vital partner in the fields of memory and AI infrastructure.

Naver Launches 200MW AI Data Center… Expanding AI Platform Competitiveness

Naver’s collaboration with NVIDIA takes a different direction from SK’s.

While SK is focused on building a large-scale AI computing infrastructure supply chain, Naver is concentrating on expanding its platform ecosystem that connects AI models, cloud services, and enterprise solutions.

Naver is pursuing a $10 billion (approximately 14.6 trillion won) AIDC investment partnership involving NVIDIA and global investment firm Brookfield. Under this structure, NVIDIA will invest $1 billion and Brookfield $9 billion in Naver’s business.

This amount represents nearly half of Naver’s market capitalization (approximately 32 trillion won), making it one of the largest AI infrastructure investment projects by a South Korean company.

A Naver official expressed confidence, stating, “Based on this $10 billion investment, we plan to first build a 200 MW AI data center and expand it to 1 GW in the long term,” adding, “We will be able to launch the AI Factory service first.”

In particular, Naver is set to target the enterprise AI market by leveraging not only its hyper-scale AI model “HyperClova X” but also the data and cloud capabilities it has accumulated through operating large-scale services such as search, advertising, and commerce.

Through this collaboration, the company aims to strengthen the competitiveness of its sovereign AI platform, which integrates everything from AI model development to data center operations and enterprise AI services.

As an NVIDIA Cloud Partner, Naver plans to leverage GPU infrastructure to expand its AI services for enterprises, public institutions, and startups.

However, expanding its 1GW-class AI data center (AIDC) requires experience in building and operating data centers, as well as securing infrastructure capabilities such as power and cooling. Accordingly, in addition to attracting further investment, strategic collaborations with IDC operators, construction firms, and energy companies are likely to gain momentum.

CEO Jensen Huang has emphasized the importance of sovereign AI that reflects each country’s language, culture, and data, stating, “Every nation needs its own AI.”

This collaboration with Naver is also viewed as a key step toward building a Korean-style sovereign AI platform.

NVIDIA’s Choice of the “AI Factory + AI Infrastructure Platform” Combination

Analysts suggest that NVIDIA’s decision to partner with both SK and Naver stems from the balance inherent in South Korea’s AI ecosystem.

SK is responsible for AI infrastructure capabilities such as GPUs, HBM, data centers, and power, while Naver handles AI models, cloud, and service platform capabilities.

From NVIDIA’s perspective, meeting the rapidly growing demand for AI computing in the Asia-Pacific region requires both companies capable of building large-scale infrastructure and platform companies that can leverage it to create services.

CEO Jensen Huang recently described the Korean AI industry as a “Golden Age,” highly praising Korea’s competitiveness in memory, manufacturing capabilities, and the growth potential of its AI ecosystem.

The assessment is that AI competition is no longer determined solely by model performance. To secure competitiveness at the national level, it is necessary to secure the GPUs, data centers, power supply, and data management systems required to run AI models.