Nvidia CEO Jensen Huang announced on June 8, 2026, that SK Hynix will remain Nvidia’s largest memory partner under a multi-year agreement lasting more than two years with an option to extend [1, 2]. Huang said, "SK Hynix has been Nvidia’s largest memory partner. SK Hynix will continue to be Nvidia’s largest memory partner" [2].
The deal covers chip design and manufacturing cooperation focused on advanced memory technologies, including the next-generation high-bandwidth memory standard HBM4 [1, 2]. Huang added, "Together, we will co-develop the next generation of memory for AI factories and support the accelerating global expansion of AI infrastructure – from frontier model training to agentic and physical AI" [1].
SK Hynix’s parent company, SK Group, also signed agreements with SK Telecom, Naver, and Doosan, all of which will deploy Nvidia technology to build AI data centers [2]. SK Telecom will construct a gigawatt-scale AI cloud in South Korea using Nvidia hardware. The first AI data center from this project is expected to come online in 2027 [2].
Nvidia has been purchasing billions of dollars’ worth of memory components from SK Hynix annually and expects this volume to grow substantially going forward, Huang said [2].
The partnership supports Nvidia’s expanding AI infrastructure portfolio, including the Vera Rubin AI accelerator system expected to begin shipping in the third quarter of 2026. This system integrates HBM4 memory supplied by SK Hynix [1].
The first shipments of the Vera Rubin AI accelerators are scheduled for Q3 2026, marking a near-term milestone for the cooperation [1]. SK Telecom’s AI data center project will follow in 2027, reinforcing Nvidia’s growing presence in South Korea’s AI ecosystem [2].
