Physical AI HBM Smart Factory SDV AIoT Power Semicon 특수 가스 정정·반론보도 모음 e4ds plus

SK Hynix and NVIDIA Jointly Develop Next-Generation Memory for AI Factories

Google 우선 소스 기사입력2026.06.08 11:50


SK Group Chairman Chey Tae-won and NVIDIA CEO Jensen Huang shake hands at the SK Seorin Building in Jongno-gu, Seoul. (Photo: SK Hynix)
Expanding long-term partnership based on HBM collaboration, extending cooperation to design, simulation, and manufacturing AI
SK Hynix is jointly developing next-generation memory for AI factories with NVIDIA. Building on their existing cooperation in High Bandwidth Memory (HBM) supply, the agreement expands collaboration in semiconductor design and manufacturing to develop memory tailored to NVIDIA's next-generation AI platform.

On the 8th, SK Hynix announced a long-term technology partnership with NVIDIA for the joint development of next-generation memory to build a global AI factory and to accelerate semiconductor design and manufacturing. The two companies agreed to strengthen the development and supply system of advanced memory in line with NVIDIA's AI infrastructure roadmap.

The collaboration is not limited to AI infrastructure for data centers. The two companies plan to jointly develop NVIDIA's Vera Rubin AI supercomputer, Vera CPU, RTX Spark PC, and memory for the Jetson Thor robotics computing platform. Through this, SK Hynix expands its applications not only to AI servers but also to the realms of personal AI and physical AI.

Collaboration to apply AI and GPU acceleration technologies to the semiconductor development process is also being pursued. SK Hynix is enhancing the processing efficiency of semiconductor design and manufacturing simulation tasks, such as TCAD and computational lithography, by utilizing NVIDIA's CUDA-X library and PhysicsNeMo framework. Both companies plan to expand the scope of cooperation to the entire field of EDA and simulation in the future.

In the manufacturing sector, the advancement of fab digital twins was presented as a major task. SK Hynix is implementing semiconductor factories in a 3D virtual space using NVIDIA Omniverse and OpenUSD, and is also pursuing the optimization of equipment and asset operations within the factories through cuOpt and the Metropolis platform.

This partnership is significant in that it expands the collaboration between the two companies beyond memory supply to the development of memory for next-generation AI platforms and the advancement of semiconductor development and manufacturing environments. Moving forward, the alignment between NVIDIA's AI infrastructure roadmap and SK Hynix's advanced memory development strategy is expected to grow.