June 8, 2026 /SemiMedia/ — Nvidia and SK hynix announced a multi-year technology partnership to develop next-generation memory for global AI factories while expanding cooperation in semiconductor design, manufacturing simulation and fab automation.

The agreement builds on years of joint development between the two companies. SK hynix is one of Nvidia’s key suppliers of high-bandwidth memory, which has become a critical component in AI accelerators and large-scale data center platforms.

Nvidia Chief Executive Officer Jensen Huang said AI factories are becoming the engines of a new industrial revolution, with advanced memory playing a central role in system performance. He said the two companies will work together on next-generation memory products to support large language model training, agentic AI and physical AI applications.

SK Group Chairman Chey Tae-won said the partnership reflects a deepening relationship between SK hynix and Nvidia. He said the companies will jointly develop memory for AI factories while bringing AI technologies into semiconductor design and manufacturing workflows.

Under the partnership, SK hynix will develop customized memory technologies for Nvidia’s future platforms, including the Vera Rubin AI supercomputing platform, Vera CPUs, RTX Spark-powered PCs and Jetson Thor robotics computing systems.

Industry analysts said the agreement could further strengthen SK hynix’s position inside Nvidia’s AI ecosystem and expand its opportunities across AI infrastructure, personal AI and robotics computing markets.

Beyond memory supply, SK hynix is using Nvidia CUDA-X libraries and AI technologies to accelerate semiconductor simulation, technology computer-aided design and computational lithography. The company is also using Nvidia PhysicsNeMo and GPU acceleration to improve in-house simulation code and AI-based physics workloads.

The collaboration is also expected to extend into the semiconductor EDA and simulation ecosystem, potentially supporting closer cooperation among chipmakers, Nvidia and EDA software providers.

On the manufacturing side, SK hynix is building a fab digital twin system as a foundation for more autonomous factory operations. The system uses Nvidia Omniverse libraries, OpenUSD workflows and scene optimization technologies to create three-dimensional fab environments for visualization, simulation and production optimization.

SK hynix also plans to use Nvidia cuOpt, a GPU-accelerated decision optimization engine, and the Nvidia Metropolis vision AI platform to improve scheduling for autonomous mobile robots, production equipment and factory logistics.

The companies are also exploring ways to connect digital twin systems with legacy software and agentic AI workflows, enabling AI systems to analyze factory data and execute certain operational tasks more autonomously.

Analysts said the partnership shows AI moving deeper into semiconductor design and manufacturing, not just chip demand. As AI platforms require higher memory bandwidth, greater capacity and better energy efficiency, joint development between memory suppliers and AI platform companies is becoming increasingly important.