NextFin News - In a move that signals a tightening of the global artificial intelligence supply chain, SK Group Chairman Chey Tae-won met with Nvidia CEO Jensen Huang in Silicon Valley last week to discuss a multi-layered expansion of their strategic partnership. According to The Korea Times, the meeting took place on February 5, 2026, at a Korean-style restaurant in California, continuing a trend of informal but high-stakes "fried chicken diplomacy" between Huang and South Korean tech leaders. The discussion, which lasted approximately two hours, was attended by key executives including SK Biopharmaceuticals Vice President Chey Yoon-chung and Nvidia Robotics Division Senior Director Madison Huang, highlighting the cross-generational and cross-sector nature of the alliance.
The primary objective of the meeting was to secure the supply roadmap for Nvidia’s next-generation high-bandwidth memory, HBM4. This sixth-generation memory technology is critical for Nvidia’s upcoming Vera Rubin AI accelerator platform, which is expected to define the next frontier of large-scale model training. Beyond hardware procurement, the leaders explored the development of South Korea’s AI ecosystem and potential synergies in the biotechnology sector. This high-level engagement follows Chey’s extended stay in the United States since early February, during which he has been coordinating with major American technology firms to align SK Group’s semiconductor and AI solution strategies with global demand.
The deepening relationship between SK and Nvidia is not merely a transactional buyer-supplier arrangement but a deep integration of infrastructure and software. According to Asia Business Outlook, the two companies are currently developing a massive AI factory in South Korea equipped with over 50,000 Nvidia GPUs. The first phase of this facility is slated for completion by late 2027 and is designed to support SK affiliates and external clients through a GPU-as-a-Service (GaaS) model. Furthermore, SK hynix is utilizing Nvidia’s Omniverse and CUDA-X libraries to create digital twins of its semiconductor fabrication plants, optimizing manufacturing efficiency through virtual simulation. This level of technical integration suggests that SK Group is moving up the value chain, evolving from a component provider to a foundational partner in Nvidia’s global AI infrastructure.
From a market perspective, the collaboration highlights a significant power shift within the AI hardware ecosystem. As AI models grow in complexity, the bottleneck has shifted from raw processing power to memory bandwidth. Industry data indicates that SK hynix currently commands approximately 70% of the HBM4 market share for 2026. This dominance grants SK significant bargaining leverage, as Nvidia is reportedly set to source nearly two-thirds of its HBM4 requirements for the Vera Rubin platform from the Korean firm. The technical demands of HBM4 are also reshaping the broader memory market; the production of HBM4 requires roughly three times the wafer space of standard DRAM. This shift is expected to tighten global DRAM supply, potentially driving up prices for conventional computer memory throughout 2026 and 2027.
The inclusion of biotechnology in the discussions points toward a long-term trend where AI-driven drug discovery becomes a core growth engine for diversified conglomerates. By involving Chey Yoon-chung of SK Biopharmaceuticals, SK Group is positioning itself to leverage Nvidia’s BioNeMo platform and generative AI capabilities to accelerate its pharmaceutical pipeline. This diversification strategy is mirrored in SK’s recent rebranding of its U.S. NAND flash subsidiary, Solidigm, as an "AI Company," signaling a total organizational pivot toward the intelligence economy.
Looking forward, the SK-Nvidia alliance faces competitive pressure from Samsung Electronics, which has also begun mass production of HBM4 and is aggressively courting Nvidia’s business. However, the established infrastructure projects, such as the Korean AI factory and the deep software integration at the fab level, provide SK with a formidable first-mover advantage. As U.S. President Trump continues to emphasize domestic technology leadership and secure supply chains, the collaboration between a top-tier American chip designer and a leading South Korean memory manufacturer serves as a critical pillar of the Western AI industrial complex. The success of this partnership will likely dictate the pace of AI hardware evolution for the remainder of the decade, as the industry moves toward more specialized, high-density memory architectures.
Explore more exclusive insights at nextfin.ai.
