NVIDIA and SK hynix Enter Multiyear Memory Partnership

NVIDIA and SK hynix entered a multiyear technology partnership focused on next-generation memory codevelopment. The companies linked the work to NVIDIA’s AI infrastructure roadmap and to semiconductor design and manufacturing timelines.

The agreement addressed supply planning for advanced memory with extended development cycles, including advanced fabrication and capital investments. NVIDIA said it supported memory supply to keep pace with its infrastructure roadmap as AI factory buildout continued.

SK hynix said it would use NVIDIA CUDA-X libraries and NVIDIA PhysicsNeMo to speed semiconductor simulation, including technology computer-aided design and computational lithography workflows, as well as acceleration across in-house simulation codes and AI physics workflows. The collaboration also involved fab digital twins using NVIDIA Omniverse and OpenUSD, along with NVIDIA cuOpt for decision optimization and NVIDIA Metropolis for operational optimization, including movement of autonomous mobile robots and other fab assets.

NVIDIA and SK hynix said they planned to codevelop the next generation of memory and to apply AI to semiconductor chip design and manufacturing. Jensen Huang, founder and CEO of NVIDIA, said, “AI factories are the engines of the next industrial revolution, and advanced memory is essential to their performance,” and that “Together, we will codevelop the next generation of memory for AI factories and support the accelerating global expansion of AI infrastructure — from frontier model training to agentic and physical AI.” Chey Tae-won, Chairman of SK Group, said, “Together, we are codeveloping the next generation of memory for AI factories and applying AI to how we design and manufacture semiconductors — work that will shape the future of AI infrastructure.” The companies also described exploring ways to connect digital twins with existing legacy software and agentic AI workflows.