Nvidia announced that major technology companies are planning to adopt Vera, its first CPU built for AI agents, as the chipmaker pushes deeper into the infrastructure powering the next phase of artificial intelligence.

Vera is now in full production and is designed to handle CPU intensive workloads behind AI agents, including tool use, code execution, reinforcement learning, data processing, and orchestration. Nvidia said the processor delivers 1.8x faster task completion compared with x86 CPUs.

The customer list includes NYSE, Anthropic, OpenAI, SpaceXAI, ByteDance, CoreWeave, Lambda, Nebius, Nscale, and Oracle Cloud Infrastructure. Nvidia said Vera is also being integrated into systems from Dell, HPE, Lenovo, Supermicro, and several Taiwan based manufacturers.

The launch builds on Nvidia’s Grace CPU business, which has shipped nearly 2.5 million units. Vera marks a broader push by Nvidia to make CPUs a more central part of the AI factory, where agents are moving from answering prompts to taking actions, running code, calling tools, and evaluating results.

NYSE said it plans to use Vera with Redpanda and HPE to scale capacity and improve latency across its market infrastructure. Anthropic is evaluating the CPU for agentic workloads, while Oracle Cloud Infrastructure plans to use Vera for high throughput reasoning and data processing.

The move expands Nvidia’s role beyond GPUs at a time when AI infrastructure demand is shifting toward full stack systems. As agents become heavier users of compute, Nvidia is positioning Vera as the CPU layer that keeps GPUs fed and data centers running more efficiently.