NVIDIA and SK Hynix and have signed a multi-year memory partnership

SK Hynix and NVIDIA have signed a multi-year memory partnership to co-develop the chips that will power the next generation of AI factories. The collaboration pulls SK Hynix deeper into AI servers, personal AI, and physical AI. The company will jointly design memory for NVIDIA’s Vera Rubin AI supercomputers, Vera CPUs, RTX Spark-powered PCs, and Jetson Thor platforms. It will also leverage NVIDIA’s CUDA-X, PhysicsNeMo, Omniverse, OpenUSD, and cuOpt tools to accelerate chip design, semiconductor simulations, and autonomous fab digital twins. Advanced DRAM and high-bandwidth memory can no longer be designed after the GPU is finished — they require years of tight co-design, fabrication planning, and massive capital commitment. The partnership also brings AI directly into chipmaking itself, using simulation and digital twin technology to optimize semiconductor physics, lithography, and internal engineering processes. SK Hynix plans to create full digital copies of its factories, allowing robots, tools, wafers, and logistics to be tested virtually before any changes reach the real production line. This deal further solidifies SK Hynix’s position as a critical supplier in the AI hardware supply chain.