Samsung wraps up executive meetings on supply of next-generation HBM - 매일경제

Samsung Electronics recently held executive meetings with NVIDIA regarding the supply of next-generation high bandwidth memory (HBM), signaling a deepening collaboration in the AI chip sector. The dis... Read original →

Samsung’s deepening collaboration with NVIDIA on HBM4E and 2nm foundry services is reshaping the AI hardware stack. Technically, Samsung’s lead in 12-Hi HBM4E samples accelerates bandwidth density demands for next-gen AI accelerators and pressures CoWoS packaging capacity. NVIDIA’s shift of Groq LP30 production to Samsung breaks TSMC’s (Taiwan, China) near-monopoly in high-end AI chip manufacturing. From a compliance standpoint, tightening U.S. export controls on advanced packaging compel NVIDIA to diversify supply chains, though Samsung’s Korean fabs face scrutiny over equipment licensing and node-level compliance. Competitively, SK Hynix may fast-track its HBM4E ramp, while TSMC could counter with flexible allocation of AI-dedicated capacity. Over the next 12–24 months, tight integration of HBM and leading-edge logic nodes will become the new design paradigm—Samsung stands to capture infrastructure-level pricing power if it aligns 2nm yield with memory-stack co-optimization.