TLDR: Samsung Electronics is in advanced discussions with Nvidia to supply its upcoming HBM4 (High Bandwidth Memory) chips, a critical component for advanced AI processors. This strategic move is complemented by Samsung’s plan to launch HBM4 chips in 2026 and the construction of a new AI-powered semiconductor factory, which will be equipped with 50,000 Nvidia GPUs to automate and optimize chip manufacturing processes.
Samsung Electronics has confirmed it is in ‘close discussion’ with Nvidia to become a key supplier for its next-generation HBM4 (High Bandwidth Memory) chips, marking a significant strategic maneuver in the escalating global competition for artificial intelligence infrastructure. This development was highlighted during the Asia-Pacific Economic Cooperation (APEC) CEO Summit in Gyeongju, where executives from both companies underscored their deepening technological collaboration.
HBM4 chips are indispensable for advanced AI processors, providing the ultra-fast memory bandwidth required for training large language models and powering generative AI applications. Samsung’s potential deal with Nvidia is seen as a crucial step for the South Korean tech giant to narrow the gap with its rival, SK Hynix, which currently holds a leading position in supplying HBM chips to Nvidia.
Samsung’s ambitious AI strategy extends beyond chip supply. The company plans to begin marketing its HBM4 chips in 2026, though specific shipping timelines are yet to be disclosed. Furthermore, Samsung is investing heavily in its manufacturing capabilities by constructing an AI-powered semiconductor factory. This state-of-the-art facility will be outfitted with an impressive cluster of 50,000 Nvidia graphics processing units (GPUs), solidifying Samsung’s commitment to AI-driven manufacturing and chip design.
This new AI factory is designed to automate and optimize chip manufacturing for various applications, including mobile devices and robotics. Samsung anticipates that integrating Nvidia’s AI-driven lithography and Omniverse simulation software will lead to a 20-fold performance improvement in computational lithography and technology computer-aided design simulation, significantly enhancing production efficiency and minimizing manual oversight.
Nvidia, a dominant force in AI computing, has also confirmed its ‘key supply collaboration for HBM3E and HBM4’ with South Korean firms, including Samsung, SK Group, and Hyundai Motor Group. During CEO Jensen Huang’s visit to South Korea, Nvidia announced plans to supply over 260,000 accelerated chips to support various AI projects across the nation. The company noted its long-standing alliance with Samsung, stating, ‘From our first collaboration using Samsung DRAM in NVIDIA’s first graphics card NV1, to the launch of the industry’s first commercial HBM chip, and now the key supply collaboration for HBM3E and HBM4, the solid alliance between our two companies over 25 years has laid the foundation for today’s AI transformation.’
In the competitive HBM market, SK Hynix, Nvidia’s current primary HBM supplier, has announced its own plans to begin shipping HBM4 chips in Q4 2025 and expand sales throughout 2026. Micron Technology also remains a significant competitor in this space. Samsung’s entry into the HBM4 supply chain is expected to reshape the competitive dynamics of the AI memory market, with analysts viewing the HBM4 rollout as a critical test of Samsung’s ability to regain its technological edge.
Also Read:
- AWS to Invest Over $5 Billion in South Korean AI Data Centers, Reinforcing Regional Tech Hub Status
- Semiconductor Leaders NXP and Amkor Thrive Amidst Surging AI Hardware Investment
Beyond Samsung, Nvidia is also collaborating with SK Group to build an AI factory featuring over 50,000 Nvidia GPUs, with the first phase expected by late 2027, and with Hyundai Motor Group to deploy 50,000 Blackwell GPUs for autonomous vehicles, robotics, and smart factories. These collaborations underscore a broader industry trend where AI systems are increasingly used to automate manufacturing, accelerate innovation cycles, and build robust AI infrastructure across South Korea.
