SK Group announced a massive collaboration in AI semiconductors and data center infrastructure with NVIDIA, Microsoft, and Anthropic at the ‘K-AI Summit’ held in San Francisco on the 24th (local time). This agreement targets an investment of over $500 billion and aims to solidify the foundation for three mega-projects intended to propel South Korea into a global AI innovation hub.

SK Telecom is partnering with NVIDIA to construct the largest AI factory in South Korea, boasting a 2GW (gigawatt) capacity. Based on NVIDIA’s full-stack AI platform ‘DSX’, the facility will incorporate the ‘Vera Rubin’ next-generation AI computing system equipped with SK Hynix’s HBM4 memory, with phased operations beginning in 2027. The two companies will also co-develop and optimize next-generation AI memory capable of supporting tasks ranging from large language model (LLM) training to Agentic AI (autonomous AI systems that make independent decisions and execute tasks).

SK Hynix is pursuing a long-term supply agreement with Microsoft for server memory optimized for AI workloads. Furthermore, SK Telecom signed a memorandum of understanding (MOU) with Anthropic to build a GW-class AI data center in Korea, while agreeing with AWS to expand AI computing infrastructure across domestic regions beyond their existing collaboration in Ulsan. SK Group Chairman Chey Tae-won also met with Sam Altman (CEO of OpenAI) to discuss strengthening long-term partnerships, including memory supply.

Chairman Chey emphasized that through this cooperation, South Korea must move beyond being a mere consumer of AI to become a global AI infrastructure hub. He presented three core tasks: realizing an ‘AI-native nation’ where every citizen uses one AI agent, reducing token costs through increased memory production and expanded data centers, and establishing AI governance to minimize side effects while ensuring environmental preservation. SK intends to secure AI technological leadership and accelerate its global ecosystem expansion through these partnerships.

SK Group announced a massive collaboration in AI semiconductors and data center infrastructure with NVIDIA, Microsoft, and Anthropic at the ‘K-AI Summit’ held in San Francisco on the 24th (local time). This agreement targets an investment of over $500 billion and aims to solidify the foundation for three mega-projects intended to propel South Korea into a global AI innovation hub.

SK Telecom is partnering with NVIDIA to construct the largest AI factory in South Korea, boasting a 2GW (gigawatt) capacity. Based on NVIDIA’s full-stack AI platform ‘DSX’, the facility will incorporate the ‘Vera Rubin’ next-generation AI computing system equipped with SK Hynix’s HBM4 memory, with phased operations beginning in 2027. The two companies will also co-develop and optimize next-generation AI memory capable of supporting tasks ranging from large language model (LLM) training to Agentic AI (autonomous AI systems that make independent decisions and execute tasks).

SK Hynix is pursuing a long-term supply agreement with Microsoft for server memory optimized for AI workloads. Furthermore, SK Telecom signed a memorandum of understanding (MOU) with Anthropic to build a GW-class AI data center in Korea, while agreeing with AWS to expand AI computing infrastructure across domestic regions beyond their existing collaboration in Ulsan. SK Group Chairman Chey Tae-won also met with Sam Altman (CEO of OpenAI) to discuss strengthening long-term partnerships, including memory supply.

Chairman Chey emphasized that through this cooperation, South Korea must move beyond being a mere consumer of AI to become a global AI infrastructure hub. He presented three core tasks: realizing an ‘AI-native nation’ where every citizen uses one AI agent, reducing token costs through increased memory production and expanded data centers, and establishing AI governance to minimize side effects while ensuring environmental preservation. SK intends to secure AI technological leadership and accelerate its global ecosystem expansion through these partnerships.