On October 31, Samsung Semiconductor announced a collaboration with NVIDIA to build an AI Megafactory. By deploying more than 50,000 NVIDIA GPUs, Samsung will implement AI across the entire manufacturing process, accelerating the research, development, and production of next-generation semiconductors, mobile devices, and robots.
Samsung stated that the AI Factory will integrate all stages of semiconductor manufacturing—from design and process to equipment, operations, and quality control—into a single intelligent network, where AI will continuously analyze, predict, and optimize the production environment in real time. Samsung’s AI Factory will go beyond traditional automation, serving as a smart manufacturing platform capable of linking and interpreting the vast data generated during chip design, production, and equipment operation.
Samsung said that in addition to celebrating a partnership of more than 25 years, it is currently co-developing HBM4 with NVIDIA. Using Samsung’s sixth-generation 10-nanometer-class DRAM and a 4-nm logic base die, HBM4 can achieve processing speeds of up to 11 Gbps, significantly ahead of the JEDEC standard of 8 Gbps.
Samsung will continue to provide next-generation memory solutions—including HBM, GDDR, and SOCAMM—as well as foundry services, driving innovation and expansion across the global AI value chain.
Samsung noted that over the next few years it plans to adopt NVIDIA accelerated computing technologies to scale its AI Factory and use NVIDIA Omniverse libraries to accelerate digital-twin manufacturing. This initiative will be applied to one of the world’s most comprehensive chip manufacturing infrastructures, spanning memory, logic chips, foundry, and advanced packaging.
By leveraging NVIDIA cuLitho and CUDA-X libraries to enhance its optical proximity correction (OPC) process, Samsung has increased computational lithography throughput by 20×. As a key step in precise wafer patterning, the enhanced OPC enables AI to predict and correct circuit pattern errors faster and more accurately, thereby shortening development cycles. In electronic design automation (EDA), the two companies are working with EDA partners to develop next-generation GPU-accelerated EDA tools and design technologies.
Samsung plans to expand its AI Factory infrastructure to global manufacturing hubs, including Taylor, Texas, bringing greater intelligence and flexibility to its worldwide semiconductor operations. Meanwhile, Samsung is collaborating across multiple NVIDIA AI platforms to connect virtual simulations with real-world robotics data, enabling robots to understand environments, make decisions, and operate intelligently in real scenarios. Through the NVIDIA Jetson Thor module, Samsung has accelerated intelligent robots’ capabilities in real-time AI inference, task execution, and safety control.