Nvidia (NVDA) CPUs to Power Supercomputers at the Los Alamos National Laboratory - Yahoo Finance

NVIDIA has announced that its new 'Vera Rubin' platform will be used in supercomputers at the Los Alamos National Laboratory (LANL), underscoring its growing influence in high-performance computing an... Read original →

NVIDIA’s Vera Rubin platform at Los Alamos marks its strategic shift from AI accelerator vendor to foundational HPC infrastructure provider. Technically, integrating double-precision FP64 with CUDA-X forces EDA, compilers, and storage stacks to co-evolve around NVIDIA’s full-stack dominance, raising software barriers. Geopolitically, U.S. export controls on advanced compute increase overseas deployment costs but cement NVIDIA’s status as a ‘trusted’ domestic supplier for national labs. Facing AMD’s MI300 and Intel’s Gaudi 3, NVIDIA counters by embedding safety standards via Halos for Robotics—turning system-level integration into a defensible moat. Over the next 18 months, its GPU-InfiniBand-AI stack will dictate global supercomputing procurement, compelling research institutions to architect data pipelines around its ecosystem, deepening lock-in.