SK Telecom is joining forces with NVIDIA to establish a global artificial intelligence (AI) infrastructure.
Specifically, they are collaborating on a full-stack AI cloud based on the NVIDIA DSX platform, encompassing everything from chips to data center operations.
As part of this initiative, they plan to expand their AI Factory, a specialized data center for AI tasks, with the goal of reaching gigawatt-scale operations.
On Monday, SK Telecom announced that SK Group Chairman Chey Tae-won and NVIDIA Chief Executive Officer (CEO) Jensen Huang agreed to review the AI infrastructure roadmap for both companies and pursue group-level collaboration.
SK Telecom aims to spearhead the development of Asia’s largest AI infrastructure, working closely with NVIDIA as the primary implementation partner.
Aiming for Gigawatt-Scale… Phased Expansion Through the Global AI Factory Project
The AI Factory functions as an intelligent facility that continuously produces tokens, the fundamental units of AI, using power and data as raw materials. SK Telecom will construct this based on NVIDIA’s DSX infrastructure, representing a next-generation concept that surpasses traditional data centers limited to general computing and data storage.
The AI Factory is scheduled to commence operations in South Korea in 2027, likely becoming the first case to validate the governance and operational structure of both companies’ AI clouds.
SK Telecom plans to gradually scale this model to gigawatt-level infrastructure and expand AI infrastructure across Asia.
To achieve this, SK Telecom will join the NVIDIA Cloud Partner program, a global ecosystem initiative that provides high-performance cloud AI services using NVIDIA’s AI infrastructure and software. This move aims to secure the lowest token costs and highest performance per watt, crucial factors for competitiveness in the AI infrastructure market.
The collaboration with NVIDIA is expected to form the cornerstone of SK Telecom’s AI cloud business. Starting with NVIDIA’s Blackwell graphic processing unit (GPU), SK Telecom plans to support AI training and inference, and will progressively integrate the latest NVIDIA Vera Rubin platform, slated for release in the second half of this year.
Through this partnership, SK Telecom will gain access to NVIDIA’s AI infrastructure, including computing resources and specialized software, while NVIDIA will benefit from SK Hynix’s memory leadership and SK Telecom’s expertise in building and operating AI factories.
This collaboration signals a shift in the AI market’s supply paradigm, evolving beyond mere component and infrastructure provision to a model that generates high-value intelligence.
The emphasis on the AI Factory, following a meeting between Jensen Huang and Chairman Chey Tae-won in Taiwan, can be seen as part of this strategic vision.
While NVIDIA has previously focused on developing hardware-centric supply chains with Korean companies, such as high-bandwidth memory (HBM), it now aims to create a next-generation AI service hub that integrates infrastructure and software.
SK Telecom aims to rapidly expand its AI cloud business through this partnership.
According to SK Telecom, the AI cloud represents a new business model offering services tailored for AI tasks such as training, inference, and data processing, distinguishing itself from existing cloud businesses that provide general cloud solutions.
This market has recently emerged as a critical component of the global AI infrastructure supply chain, with major tech giants leasing large-scale AI infrastructure.
SK Telecom envisions establishing itself as a key partner providing the AI infrastructure and business network NVIDIA needs across Asia, aiming to become one of the region’s leading AI cloud providers.
SK Group and NVIDIA announce joint research to develop next-generation AI factory design and operational systems
SK Group and NVIDIA have unveiled plans to conduct research aimed at jointly developing the architecture for next-generation AI factories, expanding their collaboration beyond infrastructure construction.
Previously, cooperation between SK and NVIDIA centered on semiconductors, such as SK Hynix’s HBM. This new agreement broadens their partnership to encompass all areas of AI infrastructure, including AI factory construction and operation.
The new R&D collaboration will involve joint research on a novel computing architecture that enhances GPU and memory performance from the design phase. The companies plan to establish a joint council for this purpose.
They have also been collaborating continuously in the field of physical AI. During NVIDIA’s keynote speech at GTC Taipei on June 1, SK Telecom showcased large-scale digital twin technology built on NVIDIA Omniverse, which they are applying to SK Hynix’s semiconductor manufacturing processes.
SK Telecom is also partnering with NVIDIA in the robotics field. They are enhancing their robot simulation and training platform based on NVIDIA’s robotics platform NVIDIA Cosmos and the humanoid AI model NVIDIA Isaac GR00T.
Chairman Chey Tae-won stated that the close partnership with NVIDIA has enabled them to develop full-stack AI infrastructure capabilities, from chips to data center operations. By going beyond simple service provision and jointly addressing GPU, memory, and energy challenges, we aim to become a leading AI cloud provider driving the development of the AI ecosystem across Asia.
Jensen Huang, CEO of NVIDIA, remarked that telecommunication networks are evolving into national AI infrastructure, with communication networks connecting people, businesses, and devices now forming the foundation of AI clouds. Through the NVIDIA DSX platform, SK Telecom will build a large-scale AI cloud, providing agent AI, enterprise AI, and physical AI to companies and industries leading Korea and the world.
In a related development, SK Telecom CEO Ryu Young-sang met with Huang at Kkanbu Chicken in Gangnam on the evening of June 7 during Huang’s visit to Korea.