SK Group committed more than $500 billion to a sweeping AI infrastructure partnership with NVIDIA, pledging to build one of the world’s largest AI factories and lock in next-generation memory supply through a multiyear HBM codevelopment deal. The announcement, made July 24 at NVIDIA’s AI Summit in San Francisco, includes separate letters of intent signed by SK Telecom, SK hynix, and the parent conglomerate.
SK Telecom will construct a 2-gigawatt AI data center powered by NVIDIA’s Vera Rubin DSX platform and equipped with SK hynix HBM4 memory, with the first phase targeted for 2027. The facility will serve sovereign AI, agentic AI, enterprise workloads, and physical AI across the Asia-Pacific region. SK hynix, meanwhile, formalized a long-term HBM supply arrangement that extends beyond HBM4 into future architectures aligned with NVIDIA’s GPU roadmap.
“Competitiveness depends not just on how effectively AI is utilized, but on how much intelligence we can produce,” said SK Group Chairman Chey Tae-won. NVIDIA CEO Jensen Huang called South Korea “a global AI powerhouse” with the ingredients to lead the next wave of AI infrastructure. The pact solidifies SK Group’s role as NVIDIA’s primary memory partner while giving the chip giant guaranteed access to the advanced HBM stacks its Rubin-class accelerators will require.
