NVIDIA and SK Group Forge $500 Billion Alliance to Power Korea's AI Ambitions
News Summary
NVIDIA and SK Group have announced an expanded strategic partnership valued at more than $500 billion, covering the construction of large-scale AI factories and a long-term supply agreement for next-generation memory. The two companies signed letters of intent detailing plans that combine advanced GPU computing infrastructure with high-bandwidth memory development, positioning South Korea as a major hub for AI infrastructure in the Asia-Pacific region.
AI Factory Construction
At the center of the agreement is a plan for SK Telecom to build a 2-gigawatt AI factory using NVIDIA's DSX platform paired with the upcoming Vera Rubin accelerated computing architecture. The facility is expected to include more than 50,000 NVIDIA GPUs and will be powered in part by SK hynix HBM4 memory modules. The first phase is targeted for completion in 2027, Eastern Time, according to statements released alongside the announcement.
Once operational, the facility is expected to become one of the largest AI computing sites in Korea. It is designed to serve both internal SK Group subsidiaries — including SK hynix and SK Telecom — and external organizations through a GPU-as-a-service model, allowing businesses across manufacturing, telecommunications, and research to access advanced computing resources without building their own infrastructure.
Memory Supply and Joint Development
Beyond infrastructure construction, the partnership includes a long-term agreement between NVIDIA and SK hynix to secure next-generation memory supply. The companies will jointly develop high-bandwidth memory (HBM) technology intended to support AI training, AI agent workloads, and physical AI applications such as robotics and autonomous systems.
This memory collaboration is designed to help ensure a stable supply chain for the advanced components required by next-generation AI accelerators, while also advancing the underlying semiconductor technology used across data centers worldwide.
Regional and Industry Impact
The partnership is framed as part of a broader effort to meet surging global demand for AI compute capacity. By anchoring large-scale AI infrastructure in South Korea, the collaboration is expected to support digital transformation across manufacturing, telecommunications, and other industries in the region.
Executives from both companies described the agreement as a significant step in scaling accelerated computing and memory technology to meet the needs of AI research, enterprise adoption, and emerging application areas such as physical AI and robotics.
What Comes Next
The initial phase of the AI factory is slated for completion in late 2027, Eastern Time, with subsequent phases expected to expand capacity over time. Analysts have noted that while the scale of the announcement is substantial, the multi-year timeline means the full impact of the partnership — including its effect on global memory supply and AI infrastructure capacity — will become clearer as construction milestones are reached in the coming years.