Nvidia and SK hynix have launched a long-term partnership to jointly develop cutting-edge memory technologies designed to support artificial intelligence (AI) systems, marking a new chapter in their collaboration. This alliance aims to push the boundaries of memory innovation, extending beyond conventional standards to meet the growing demand for AI infrastructure components.
The partnership includes co-creating technology roadmaps and accelerating development efforts for Nvidia’s upcoming AI platforms, notably the Vera Rubin supercomputers. Vera Rubin, a flagship AI system, relies heavily on advanced memory components such as HBM4, which SK hynix supplies alongside competitors Samsung Electronics and Micron Technology. Vera Rubin is currently in full production and expected to start deliveries later this year.
SK hynix anticipates that this collaboration will strengthen its foothold in emerging AI sectors, including personal AI technologies and physical AI applications, while addressing challenges related to the lengthy development cycles for advanced memory products. Nvidia CEO Jensen Huang emphasized that the AI revolution is still in its early stages and highlighted the significant potential in expanding their joint efforts.
During Huang's visit to South Korea, he met with industry leaders and researchers to broaden cooperation in AI and robotics, signaling Nvidia’s commitment to the Asian market. SK Group Chairman Chey Tae-won characterized the deal as an elevation of their existing memory-focused cooperation to a more strategic and expansive partnership.
Beyond enhancing memory performance, Nvidia intends to support SK hynix’s expansion into new AI infrastructure domains, including focusing on physical AI innovations. This integrated approach aligns with the growing complexity and scale of AI workloads, requiring a tighter integration between hardware memory solutions and AI platform capabilities.