SK Group and NVIDIA Sign USD 500B+ Mega-Deal to Build 2-Gigawatt AI Factory in Korea

SK Group, South Korea's second-largest conglomerate, and NVIDIA Corporation (NASDAQ: NVDA) have signed Letters of Intent (LOIs) formalising a strategic partnership valued at over USD 500 billion in comprehensive multi-year business commitments. The landmark alliance bridges the two primary physical constraints of modern artificial intelligence: the construction of hyperscale, gigawatt-class "AI Factories" and the guaranteed production of high-density High Bandwidth Memory (HBM).

The partnership establishes a unified ecosystem connecting three core entities: NVIDIA supplying its next-generation Vera Rubin full-stack computing architecture; SK Telecom managing infrastructure and deployment for a massive 2-gigawatt AI Cloud; and SK hynix locking in long-term supply and co-development for advanced HBM4 memory substrates.

The 2-Gigawatt Build: Powered by NVIDIA Vera Rubin and DSX

The infrastructure cornerstone of the deal expands on SK Telecom’s initiative to construct a 2-gigawatt-scale AI Cloud facility in South Korea, with the first operational AI factory slated to go live in 2027.

The facility will be engineered using the NVIDIA DSX full-stack AI factory architecture. Succeeding legacy Hopper and Blackwell DGX iterations, the DSX platform integrates NVIDIA accelerated hardware, custom networking fabrics, enterprise system software, and partner cooling technologies to deliver thelowest token processing cost at maximum energy efficiency.

By deploying this computing density locally, SK Telecom and NVIDIA aim to broaden regional customer access to capital-intensive infrastructure. The facility will serve as a computing engine for regional enterprises, government bodies, and international tech hubs deploying:

• Sovereign AI: Localised foundation models compliant with regional data residency and governance rules.

• Agentic AI: Autonomous multi-step software agents executing complex business workflows.

• Physical AI: Real-time multimodal reasoning engines orchestrating industrial robotics, digital twins, and autonomous vehicles on shop floors.

Securing the Semiconductor Substrate: SK hynix HBM4 Co-Development

While GPU compute logic handles mathematical processing, memory bandwidth remains the primary bottleneck in scaling large language model (LLM) training and real-time inference. As part of the strategic agreement, SK hynix, which currently holds approximately 60% of the global HBM market share, has entered into an expanded, multi-year AI memory partnership with NVIDIA.

The long-term agreement guarantees NVIDIA a stable, priority supply of next-generation HBM4 memory modules. In return, SK hynix secures guaranteed volume commitments to offset the capital intensity of building next-generation memory fabrication facilities. The two engineering teams will co-design and optimise custom memory stacks directly alongside NVIDIA’s Vera Rubin silicon, optimising physical layer interfaces to accelerate memory bandwidth and thermal dissipation.

Executive Perspectives on Global Compute Expansion

SK Group Chairman Chey Tae-won underscored that future economic competitiveness relies directly on intelligence generation capacity:

“In the AI era, competitiveness depends not just on how effectively AI is utilized, but on how much intelligence we can produce. By leveraging SK hynix’s AI memory and SK Telecom’s AI infrastructure capabilities, SK will collaborate with NVIDIA to build a world-class AI factory, helping Korea transcend its role as a leading adopter of AI and become a global hub that drives AI innovation.”

Jensen Huang, founder and CEO of NVIDIA, highlighted South Korea's industrial readiness to host frontier-scale compute infrastructure:

“South Korea has all the ingredients to become a global AI powerhouse — world-class networks and data centers, leadership in chip technology and vast industrial scale. Together with SK Telecom and SK hynix, we are building a new generation of AI factories that will power Korea’s next wave of growth.”

By coupling 2 gigawatts of accelerated power infrastructure with guaranteed next-generation HBM4 memory production, SK Group and NVIDIA establish an integrated compute fabric. This symbiotic alliance ensures that as global demand for artificial intelligence transitions from software experiments to real-world physical deployment, the underlying silicon, memory, and data centre infrastructure is positioned to scale efficiently.

Read More:

From ERP to ERX: How AI Is Turning Planning into Intelligent Execution

Why AI security is becoming the biggest growth opportunity for channel partners

Why Dell believes AI PCs need consultative selling, not hardware transactions

How cloud transformation is becoming a business strategy conversation