Los Alamos National Laboratory is building three new supercomputers named Mission, Vision and Veritas in partnership with HPE and NVIDIA. The systems will use the HPE Cray Supercomputing GX5000 architecture combined with NVIDIA Vera Rubin GPU nodes, NVIDIA Vera CPUs and NVIDIA Quantum-X800 InfiniBand networking.

Mission will include NVIDIA Vera Rubin GPU nodes and 2,300 standalone NVIDIA Vera CPUs using the HPE Cray Supercomputing GX240 blade. Veritas will feature approximately 1,150 standalone NVIDIA Vera CPUs to complement NVIDIA Vera Rubin nodes. Both Mission and Vision are expected to become operational in 2027. Mission will serve as the fifth Advanced Technology System in the National Nuclear Security Administration's Advanced Simulation and Computing program and will replace the Crossroads supercomputer for classified national security workloads. Vision will support fundamental science including materials and nuclear science, energy modeling, biomedical research and AI, allowing more scientists to test methods and train models. Veritas will serve the Laboratory Directed Research and Development program and help accelerate agentic AI for science while testing these technologies for use in larger systems at LANL.

The laboratory is deploying these systems to unlock agentic AI for science, with AI agents that can form hypotheses, choose tools, launch simulations, analyze outputs and refine next steps. LANL's public work on URSA, the Universal Research and Scientific Agent, demonstrates this approach. URSA is a modular, feedback-driven AI framework designed to help scientists brainstorm hypotheses, plan experiments, run simulations and analyze results, currently running on Venado and soon on Mission and Vision.

Testing has shown significant performance advantages for the Vera CPU. LANL demonstrated that the Vera CPU delivered 7x higher performance on URSA workloads than the CPUs in the Crossroads x86 supercomputer. In early testing of NVIDIA Vera CPUs on Branson, an open source Monte Carlo heat transfer simulation tool, Vera outperforms the Crossroads x86 CPUs by over 3x. A single Vera CPU outperforms a single socket x86-based CPU by more than 3x while providing more than 4x the memory per core and 6x the memory per node. These results were made possible by Vera's custom Olympus core, LPDDR5 memory and fast on-chip fabric.

All of LANL's supercomputers were codesigned by hardware architects, system software developers, domain scientists, computer scientists and applied mathematicians to ensure systems are shaped by real scientific workloads rather than abstract benchmarks alone. This represents more than a decade of deep collaboration between LANL and NVIDIA on CPUs, from Grace to Vera, using extreme codesign for LANL simulation workloads. The three new supercomputers build on Venado, the HPE Cray EX supercomputer installed at Los Alamos in 2024 with NVIDIA GH200 Grace Hopper Superchips and NVIDIA Grace CPU Superchips.