🟡 🔧 Hardware Published: · 2 min read ·

NVIDIA: Vera CPU at Los Alamos — 7× faster agentic AI for nuclear science and 3 new supercomputers

Editorial illustration: NVIDIA Vera CPU procesorski čip i shema superračunala u istraživačkom laboratoriju

The NVIDIA Vera CPU with the Olympus core is coming to Los Alamos National Laboratory, delivering 7× better performance on URSA workloads and 3× on Monte Carlo heat-transfer simulations compared to the Crossroads x86 processor. Three new supercomputers — Mission, Vision, and Veritas — will be operational in 2027.

🤖

This article was generated using artificial intelligence from primary sources.

NVIDIA Vera CPU arrives at Los Alamos: what does it bring?

The NVIDIA Vera CPU, which uses the new Olympus core with LPDDR5 memory, achieves 7× better performance on URSA workloads and 3× better performance on Monte Carlo heat-transfer simulations compared to the Crossroads x86 processor currently used at Los Alamos National Laboratory. In addition, Vera delivers 4× more memory per core and 6× more memory per node compared to the previous generation — meaning dramatically larger simulation capacity at the same cost.

Three new supercomputers for 2027

Los Alamos National Laboratory will receive three new supercomputers: Mission, Vision, and Veritas. Each combines the NVIDIA Vera CPU with the Rubin GPU and a Quantum-X800 InfiniBand network interface. All three systems are planned for operational status in 2027 and are intended for the most demanding problems in nuclear physics, materials modeling, and complex multi-physics simulations.

What is agentic AI and how does URSA apply it?

Agentic AI refers to systems that autonomously plan and execute multi-step tasks without constant human supervision — as opposed to classical models that only respond to individual queries. Los Alamos has developed URSA (a modular AI framework) that operates in exactly such an agentic loop: the system independently formulates scientific hypotheses, plans experiments, runs simulations, and analyzes results.

The ultimate goal is for URSA agents to autonomously design materials and molecules for nuclear science — for example, optimizing the composition of fuel elements or predicting material behavior under extreme conditions. The NVIDIA Vera CPU, with its significantly larger memory capacity, is essential for running such URSA agentic loops on real nuclear workloads.

What this means for the market and AI hardware development

The NVIDIA–Los Alamos partnership signals that top-tier CPU design is no longer a secondary concern for AI hardware — alongside the GPU, memory density and CPU-side latency increasingly determine the overall performance of agentic systems. The Vera CPU is therefore not merely an upgrade of the Crossroads x86 architecture but a shift toward a vertically integrated NVIDIA hardware stack (CPU + GPU + InfiniBand) now being tested in one of the most demanding computing environments in the world.

Frequently Asked Questions

What is the NVIDIA Vera CPU and how does it differ from previous generations?
The Vera CPU uses the Olympus core with LPDDR5 memory and delivers 4× more memory per core and 6× more memory per node compared to the previous generation, making it suitable for complex nuclear simulations.
What is agentic AI and how does Los Alamos apply it to nuclear science?
Agentic AI refers to systems that autonomously plan and execute multi-step tasks without constant human supervision. Los Alamos uses the URSA framework in an agentic loop — from hypothesis formulation and experiment planning to simulation and result analysis — for autonomous materials and molecule design.