Nvidia-Backed CoreWeave to Offer First AI Agent CPU, NVIDIA Vera

CoreWeave, an Nvidia-backed cloud provider, is set to offer NVIDIA Vera, what Nvidia calls the first CPU built specifically for AI agents. This marks a shift in how cloud providers think about the 'boring' half of AI infrastructure: the CPU cores that manage sandboxes, tool calls, and orchestration around the GPUs.

Nvidia designed Vera around workloads barely existing three years ago: autonomous AI agents that plan, call tools, spin up sandboxed environments, and loop through reasoning steps thousands of times a minute. CoreWeave is betting agent-heavy customers will pay for CPU capacity engineered the same way its GPU capacity is: for a specific job, not a general one.

Vera doesn't ship alone; it's part of Nvidia's Vera Rubin NVL72 platform, which pairs 72 Nvidia Rubin GPUs with 36 Vera CPUs in a single integrated unit. A full rack packs 128 Vera CPUs and 11,264 CPU cores, supporting over 11,000 concurrent execution environments.

Nvidia claims the platform delivers up to 10 times higher inference performance per watt than its previous-generation Blackwell NVL72 system, and can cut cost per million tokens to one-tenth of Blackwell's. However, these figures are tied to specific workloads and benchmark conditions rather than independently verified performance numbers.