en.Wedoany.com Reported - On June 8, NVIDIA's new Vera CPU will use DRAM memory from South Korea's SK Hynix. During a visit to South Korea, NVIDIA founder and CEO Jensen Huang stated that the Vera CPU, a new processor designed for data centers, will also utilize SK Hynix memory. This statement echoes the multi-year technology partnership the two companies just announced, further embedding advanced memory supply into NVIDIA's AI factory roadmap.
The Vera CPU is a key central processing unit platform from NVIDIA targeting AI data centers. It will serve large-scale training, inference, and agent workloads alongside the upcoming Vera Rubin AI supercomputing system. Compared to discussions of AI computing power centered solely on GPUs, the significance of the Vera CPU lies in complementing the general-purpose computing, data scheduling, memory access, and system-level coordination capabilities within AI factories. As AI cluster scales expand, the efficiency of data flow between CPUs, GPUs, memory, networks, and storage directly impacts training throughput and inference costs. The inclusion of SK Hynix DRAM in the Vera CPU supply system means their collaboration is no longer limited to pairing HBM and other high-bandwidth memory with GPU platforms. Instead, it extends into NVIDIA's data center CPU and overall system architecture, providing a more stable foundation for memory coordination in future AI factory construction.
On the same day, NVIDIA and SK Hynix announced a multi-year technology partnership. The two companies will jointly develop next-generation memory for AI factories and support the Vera Rubin AI supercomputer, Vera CPU, RTX Spark PC, and Jetson Thor robotics computing platform.
The industry signal from this collaboration is clear: AI infrastructure is transitioning from a "graphics card procurement" phase to a "system-level supply chain binding" stage. With the Vera CPU adopting SK Hynix DRAM, memory manufacturers need to participate earlier in processor platform definition, memory specification coordination, power control, signal integrity, and mass production scheduling. AI factories differ from ordinary data centers; they feature high load density, strong task continuity, and rapid cluster expansion, demanding greater stability in memory supply and longer development cycles. Advanced memory requires a lengthy period from R&D and validation to mass production. Insufficient coordination between chip platforms and the memory supply chain can impact system delivery and the pace of AI cloud expansion. NVIDIA's choice to pursue a multi-year partnership with SK Hynix not only locks in a high-performance memory roadmap for future AI factories but also secures clearer memory support for personal AI, physical AI, and robotics computing platforms.
The collaboration also extends to semiconductor manufacturing itself. SK Hynix will use NVIDIA's CUDA-X libraries and PhysicsNeMo framework to accelerate semiconductor simulation, technology computer-aided design, and lithography computing processes. Additionally, it will leverage Omniverse, OpenUSD, and cuOpt to build digital twins of wafer fabs for simulating and optimizing complex manufacturing environments. For SK Hynix, NVIDIA is not just a memory procurement and platform partner but is also becoming a technology supplier for chip design, factory scheduling, and manufacturing automation. The Vera CPU's adoption of SK Hynix DRAM is just one specific outcome of the collaboration. The broader change behind it is the formation of a tighter cycle among AI chip platforms, advanced memory, and semiconductor manufacturing processes: AI computing power demands stronger memory, and memory manufacturing, in turn, leverages AI tools to enhance R&D and production efficiency.
As the Vera Rubin system moves into subsequent mass production preparation stages, SK Hynix's role in NVIDIA's AI infrastructure will continue to evolve from a memory supplier to a joint engineering development partner. The next phase of industry observation will focus on the mass production pace of the Vera CPU platform, SK Hynix's supply capacity for DRAM and high-bandwidth memory, and the sustained pull of AI factory construction on advanced memory production capacity.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
