SK Hynix and NVIDIA are taking their high-bandwidth memory (HBM) alliance to the next level, focusing on artificial intelligence (AI) factories. The two companies have agreed to collaborate on creating AI factories for the global market, jointly developing next-generation memory tailored for AI factory infrastructure.
As part of this initiative, they will develop memory for NVIDIA’s newly launched Vera Rubin AI supercomputer and other projects, aiming to enter new markets encompassing AI infrastructure, personal AI, and physical AI.
This partnership combines NVIDIA’s AI technology with SK Hynix’s semiconductor manufacturing expertise, positioning them to spearhead global innovation in realizing Chief Executive Officer (CEO) Jensen Huang’s vision of AI factories. SK Hynix will supply specialized memory crucial for AI factories, such as HBM.
Jensen Huang: AI Factory, Driving the Next Industrial Revolution; Chey Tae-won Commits to Developing and Supplying Customized Memory
On Monday, SK Group Chairman Chey Tae-won and NVIDIA CEO Jensen Huang met at the SK Serin Building in Seoul. They agreed that SK Hynix and NVIDIA would jointly develop next-generation memory for AI factories and integrate AI into semiconductor design and manufacturing processes.
During a press briefing, Chairman Chey stated that while SK Hynix and NVIDIA have collaborated extensively on memory in the past, it’s now elevating the partnership to a broader scope involving SK Group and NVIDIA. It’ll be working together to build the future AI factory.
This development marks a significant evolution in the alliance between the two companies. As Chairman Chey noted, their previous collaboration centered on memory supply. SK Hynix has been the exclusive supplier of NVIDIA’s fourth-generation HBM3 and has maintained its competitive edge by being the first to mass-produce HBM3E. SK Hynix’s rapid growth has been bolstered by its symbiotic relationship with NVIDIA, a frontrunner in the AI era.
Through this meeting, the companies have decided to move beyond memory semiconductor supply, jointly creating the infrastructure needed for AI factories, including the development of next-generation memory.
SK Hynix will develop and supply customized memory aligned with NVIDIA’s infrastructure roadmap and global AI infrastructure demands. NVIDIA has already introduced key components for the AI factory, including the Vera Rubin AI supercomputer, Vera Central Processing Unit (CPU), RTX Spark Personal Computer (PC), and Jetson Thor robotic computing platform.
SK Hynix will also venture into new markets in AI infrastructure, personal AI, and physical AI alongside NVIDIA.
Collaboration Expands to Semiconductor Development and Manufacturing
A notable aspect of this partnership is its expansion into semiconductor development and manufacturing. This move bridges NVIDIA’s strengths in semiconductor design and development with SK Hynix’s core capabilities in production and manufacturing.
The companies are collaborating to enhance simulation technology crucial for semiconductor development. SK Hynix is leveraging NVIDIA’s CUDA-X library and PhysicsNeMo framework to improve the speed and efficiency of simulations in semiconductor design and manufacturing. This includes Technology Computer-Aided Design (TCAD) for analyzing semiconductor process characteristics and computational lithography for precise circuit implementation.
Simultaneously, they’re driving innovation in semiconductor development environments, agreeing to integrate digital twin technology with existing manufacturing systems and AI-based workflows.
These plans appear to have crystallized through several recent meetings between Chairman Chey and CEO Huang. They’ve met eight times in the past seven months, including three meetings during Huang’s recent visit to Korea. SK Hynix executives, including Chairman Chey, also attended NVIDIA’s annual AI conference, GTC Taipei, in Taiwan on May 1-2.
During his Taiwan visit, CEO Huang expressed plans to focus on building hardware-centric supply chains with domestic companies, including HBM, while also aiming to create a next-generation AI service hub integrating infrastructure and software.
At Computex 2026, Chairman Chey emphasized the AI Factory as SK Hynix’s strategic goal. He stated that its currently produce memory chips for AI, but the future ambition is to produce AI factories themselves. He envisioned AI factories capable of generating more Intel Corporationligence in the future, signaling SK Hynix’s intention to evolve into a comprehensive AI infrastructure company. This shared vision of the AI Factory aligns with concepts CEO Huang highlighted at Computex in 2024 and 2025.