NVIDIA's Vera CPU Begins Shipping to AWS and Leading AI Labs

NVIDIA's Ian Buck hand-delivered the first Vera CPU server and Vera Rubin GPU to AWS, marking the start of Vera's deployment across the AI ecosystem. This significant milestone follows deliveries to Oracle Cloud Infrastructure and three leading AI labs: Anthropic, OpenAI, and SpaceXAI.

"Agentic AI is creating a new CPU moment in the AI factory — as models move from answering to acting, Vera is purpose-built to keep that work moving at scale," Buck said.

Vera, NVIDIA's first custom CPU, is designed specifically for agentic AI, which demands more from infrastructure than traditional workloads. It features 88 custom Olympus cores, 1.2TB/s of memory bandwidth, and delivers up to 1.8x faster per-core performance on agentic AI workloads. This makes Vera ideal for tasks such as orchestration, tool-calling, reinforcement learning (RL) workloads, data analytics, agent sandboxing, and long-context state management.

The Vera CPU system is part of NVIDIA's broader extreme codesign strategy, which includes the Vera Rubin GPU, BlueField-4 DPU, Spectrum-X, and MGX rack architecture. In these systems, Vera and Rubin share a unified memory architecture, ensuring accelerated compute remains highly utilized.

AWS received its first Vera CPU server and Vera Rubin GPU in Seattle, handed over by Buck to Willem Visser, vice president of Amazon EC2 at AWS, and Supreeth Sheshadri, vice president of Amazon EC2 hardware engineering. This delivery marks the beginning of Vera's integration into AWS's infrastructure, further expanding NVIDIA's collaboration with AWS and its commitment to advancing agentic AI.

Source: nvidia-blog