NVIDIA and SK Hynix Announce Multi-Year Partnership to Build Next-Generation Memory for AI Infrastructure
NVIDIA has announced a major multi-year technology partnership with SK Hynix to co-develop advanced memory solutions designed for the next wave of AI infrastructure. The agreement strengthens an already important relationship between the two companies at a time when demand for high-performance memory is rising sharply across the global AI industry.
After a highly active visit to Computex in Taipei, NVIDIA CEO Jensen Huang traveled to Korea to meet with SK Hynix, one of NVIDIA’s most important memory partners. The visit highlighted how central memory technology has become to NVIDIA’s AI roadmap, especially as AI data centers, often described as “AI factories,” continue to expand worldwide.
The new partnership will focus on next-generation memory technologies that align with NVIDIA’s future AI computing platforms. These memory solutions will support large-scale AI training, agentic AI systems, personal AI devices, robotics, and physical AI applications.
A key goal of the agreement is to ensure that memory supply can keep pace with the rapid buildout of AI infrastructure. Advanced AI systems require enormous memory bandwidth and capacity, and development cycles for cutting-edge memory are becoming longer and more complex. By working closely together over multiple years, NVIDIA and SK Hynix aim to improve planning, production, and technology development for future AI platforms.
The partnership covers several major areas. SK Hynix will help develop memory technologies for NVIDIA’s upcoming Vera Rubin AI supercomputers, Vera CPUs, RTX Spark-powered PCs, and Jetson Thor robotic computing platforms. This signals that the collaboration is not limited to massive data centers, but also extends into personal AI computers and robotics systems.
NVIDIA and SK Hynix also plan to apply artificial intelligence directly to semiconductor design and manufacturing. SK Hynix will use NVIDIA CUDA-X libraries and NVIDIA PhysicsNeMo to speed up semiconductor simulations, technology computer-aided design workflows, and internal engineering processes. These tools are expected to help improve chip development efficiency and accelerate production innovation.
Another major part of the collaboration involves factory digital twins. SK Hynix plans to use NVIDIA Omniverse, OpenUSD scene optimization, and cuOpt to advance digital replicas of semiconductor fabrication facilities. These virtual environments could help move chip factories closer to fully autonomous operations by improving planning, optimization, and real-time decision-making.
Jensen Huang described AI factories as the engines of the next industrial revolution, emphasizing that advanced memory is essential to their performance. He also praised SK Hynix as a central partner in delivering advanced memory technologies for NVIDIA’s AI computing platforms. According to Huang, the two companies will work together to create the next generation of memory needed for frontier AI model training, agentic AI, and physical AI.
The timing of the partnership is especially important because AI growth has placed intense pressure on the global memory supply chain. High-bandwidth memory, or HBM, has become one of the most critical components in AI accelerators. As demand for AI chips continues to surge, companies such as SK Hynix are being pushed to expand production and accelerate the development of newer memory standards.
SK Hynix is already supplying NVIDIA with advanced HBM technologies and is expected to play a key role in future platforms such as Vera Rubin. The company is also working on newer memory solutions, including HBM4 and future HBM4E technologies, which are expected to support even more powerful AI systems in the coming years.
This partnership shows how closely the future of AI computing is tied to memory innovation. GPUs and AI accelerators often receive the most attention, but without faster, denser, and more efficient memory, the next generation of AI infrastructure would face serious performance limits.
By deepening its collaboration with SK Hynix, NVIDIA is strengthening one of the most important parts of its AI ecosystem. The agreement also gives SK Hynix a broader role in emerging markets such as personal AI, robotics, and autonomous manufacturing.
As AI infrastructure expands across the world, partnerships like this will be critical. NVIDIA needs reliable access to advanced memory, while SK Hynix gains a stronger position in the fast-growing AI hardware market. Together, the two companies are preparing for a future where AI factories, intelligent robots, and personal AI systems require memory technologies far beyond what today’s platforms can deliver.
