SK Group and NVIDIA have detailed a comprehensive strategic partnership that they value at more than $500 billion, spanning AI factories, large-scale data centers and next-generation memory for AI workloads. The agreement is structured around Letters of Intent that formalize a group-wide alliance, taking the relationship beyond a traditional supplier–buyer model into joint planning across chips, cloud infrastructure and manufacturing.
At the center of the initiative is SK Telecom’s plan to build a 2‑gigawatt AI data center in Korea, described as an AI factory optimized for large-scale AI output rather than general-purpose computing. The facility will use NVIDIA’s Vera Rubin chips and the company’s DSX reference architecture, combined with SK hynix’s next‑generation high‑bandwidth memory, with the first phase scheduled to come online in 2027. SK Telecom and NVIDIA have said the infrastructure is intended to support applications ranging from large language model training to agentic and physical AI, with room to scale toward gigawatt‑class capacity across Asia
SK hynix, SK Group’s memory subsidiary, has signed a multiyear technology partnership with NVIDIA to co‑develop and supply next‑generation AI memory, including future HBM generations aligned to NVIDIA’s hardware roadmap. The scope covers memory for Vera Rubin AI supercomputers, Vera CPUs, RTX Spark personal AI systems and Jetson Thor robotics platforms, expanding SK hynix’s role across all four of NVIDIA’s upcoming product lines. The companies plan to apply NVIDIA’s CUDA‑X libraries and PhysicsNeMo framework to accelerate semiconductor design, simulations and technology computer‑aided design workflows, and to build digital twin fabs using NVIDIA Omniverse and related software to pursue more autonomous manufacturing operations.
SK Group and NVIDIA have also committed to jointly develop long‑term AI research and infrastructure roadmaps that encompass semiconductor fabs, data centers and broader AI infrastructure. SK Group Chairman Chey Tae‑won and NVIDIA CEO Jensen Huang have described the partnership pillars as building AI factories together, sharing and jointly developing long‑term AI R&D plans, and co‑developing next‑generation memory for AI factories and related systems. SK Group has said the alliance effectively gives it “full‑stack” capabilities across memory, compute and cloud infrastructure, while NVIDIA gains a stable, long‑term source of advanced memory and a major regional partner for its global AI factory buildout.
Read Article: HCLTech to Invest ₹14,257 Crore in First AI Data Centre at Odisha Sovereign AI Park
