Nvidia and SK Hynix Formalize Multi-Year Memory Co-Development Pact

Nvidia and SK Hynix have established a multi-year collaboration to align memory roadmaps, focusing on HBM4, LPDDR5X, and 3D NAND integration for upcoming AI and robotic platforms.

Chips & Climate Desk
Updated Nov 6, 2026 2 min read
Nvidia and SK Hynix Formalize Multi-Year Memory Co-Development Pact

Nvidia and SK Hynix have formalized a multi-year collaboration agreement aimed at synchronizing their respective technology roadmaps. According to Tom's Hardware, the partnership focuses on co-developing next-generation memory technologies, specifically targeting HBM4, LPDDR5X, and 3D NAND. While the agreement ensures SK Hynix maintains a guaranteed role in supplying Nvidia's future systems, the companies intend to continue developing industry-standard solutions that are optimized for compatibility with Nvidia's processor architectures, rather than pursuing bespoke, non-standard memory designs.

The scope of this cooperation covers a broad range of Nvidia's upcoming hardware platforms. This includes the Vera Rubin AI systems, which will utilize HBM4, LPDDR5X, and 3D NAND, as well as standalone Vera processors, RTX Spark-powered personal computers, and Jetson Thor robotic computing systems, which will incorporate LPDDR5X and 3D NAND. By coordinating these roadmaps over multiple years, the companies aim to mitigate the extended lead times and significant capital expenditures currently required to bring advanced memory technologies to market.

Beyond supply chain alignment, the partnership emphasizes a deep integration of Nvidia's software stack into SK Hynix's semiconductor design and manufacturing workflows. SK Hynix is deploying Nvidia's CUDA-X libraries to accelerate complex development tasks, including technology computer-aided design (TCAD) and computational lithography (CuLitho). Furthermore, the memory manufacturer is adopting Nvidia's PhysicsNeMo to expedite proprietary simulation software and AI-driven physics models, with both companies exploring opportunities to expand these capabilities into broader electronic design automation (EDA) and simulation ecosystems.

Manufacturing efficiency is a central pillar of the agreement, with SK Hynix leveraging Nvidia's Omniverse and OpenUSD technologies to construct digital twins of its semiconductor fabrication facilities. These virtual environments allow engineering teams to model production lines and stress-test process adjustments before implementing them in physical fabs. The company also plans to integrate Nvidia's cuOpt and Metropolis platforms to optimize autonomous robotics and factory logistics, with future goals to connect these digital twins to existing manufacturing software for automated decision-making and data analysis.

Sources

  1. 01 Nvidia and SK hynix ink multi-year memory co-development and supply agreement — seeks to address extended development cycles — Tom's Hardware