The CPU Engine of Agentic AI: Why Nutanix and NVIDIA Vera Are Redefining Enterprise Intelligence
As more Agentic AI workflows move into production, the infrastructure requirements to support the workloads will change. Large language models still mostly depend on GPUs for inference. But agents do much more than consume tokens. Agentic AI workloads autonomously execute multi-step business workflows, compile dynamic code, query enterprise databases, and directly interact with line-of-business applications.
All of these activities run on CPUs, making CPU performance an increasingly important part of the AI infrastructure equation.
Agents are changing the role of the CPU
Early generative AI deployments focused primarily on GPU capacity. Organizations evaluated how quickly they could train or serve models, how much GPU memory they needed, and how to optimize inference.
Agentic AI introduces a broader set of requirements. An agent may need to:
- Execute code in a controlled environment
- Invoke APIs and enterprise tools
- Retrieve information from databases and business systems
- Coordinate multiple steps across a workflow and apply policies and safety controls
- Communicate with other agents or services
- Validate results before taking action
These activities create a continuous execution loop around the model. The GPU may generate the response, but the CPU coordinates much of what happens before, between, and after inference.
An AI agent's inner reasoning loop requires far more than an inference solution to run safely at scale:
- Agentic sandboxing and tool execution: When an agent formulates an action plan, it compiles generated code, executes tools, runs evaluation loops, and validates software output - tasks that run natively on CPUs.
- Integration with enterprise business systems: Agents do not exist in isolation. They retrieve context and execute operations across ERPs, CRMs, databases, and analytics pipelines that run on enterprise CPU infrastructure.
- Orchestration: Coordinating data pipelines, model routing, retrieval-augmented generation (RAG) context retrieval, and safety guardrails relies heavily on control-plane performance - that runs on CPU.
As enterprises scale from 10’s to 10's of thousands of concurrent agent workflows, CPU performance becomes a critical factor in overall AI-factory efficiency. If the CPU layer cannot keep pace, the infrastructure surrounding the GPU becomes a bottleneck.
NVIDIA Vera is purpose-built for the agentic era
NVIDIA Vera is designed to address this changing workload profile, with a focus on high-performance CPU execution, power efficiency, and the demands of agentic AI and data processing. For enterprises, the opportunity is not simply to add another CPU option. It is to create a balanced infrastructure foundation for AI workloads in which CPU and GPU resources work together across the full agent execution path.
Nutanix is working with NVIDIA to support Vera workloads, beginning with Nutanix Kubernetes Platform (NKP) Bare Metal for highly performance-sensitive AI-factory environments. Nutanix is also evaluating support for virtualized Nutanix environments.
Rather than building fragmented bare-metal AI clusters that increase management overhead, Nutanix is working toward enabling IT teams to deploy, operate, and govern Vera CPU systems using the same unified management platform, storage fabric, and hypervisor stack used across their hybrid multicloud footprint.
Bringing enterprise operations to agent execution
Specialized hardware is only one part of the solution. To move Agentic AI into production, organizations also need a way to operate these workloads securely and consistently.
Nutanix plans to extend its enterprise cloud operating model to support Vera based environments, helping customers manage agentic workloads alongside their existing infrastructure. The objective is to avoid any need to create fragmented AI environments that require separate tools, processes, and operational expertise.
The design goal is a predictable foundation for running advanced agents across the infrastructure environments where enterprise data and applications already reside, providing the following benefits:
- VM-Level sandboxing for untrusted code execution: Agents routinely generate and execute code dynamically. Running untrusted agentic code on bare-metal containers without a sandbox solution presents serious security risks. Nutanix Cloud Platform can run Kubernetes® on virtual machines, providing a secure agent sandbox using Nutanix AHV hypervisor, Flow virtual networking and full multi-tenant aware resource isolation.
- Sandboxing for containerized environments: For customers that prefer NKP Bare Metal, Nutanix plans to integrate with NVIDIA OpenShell TM to provide secure, private agent sandboxing with the isolation, control, and operational consistency required for enterprise use cases. This design allows for safer, more manageable deployment of agentic workloads at scale without treating each agent as a separate infrastructure project.
- DPU-Accelerated Zero Trust networking: Offloading Nutanix Flow virtual networking and micro-segmentation to NVIDIA BlueField TM DPUs removes network security processing overhead from the CPU. This design preserves the benefits of Vera CPU’s performance for agent execution and tool calling.
- Centralized AI Control Plane and Agent Gateway: Nutanix Enterprise AI (NAI) features the Nutanix Agent Gateway, giving IT teams centralized token cost visibility, rate limiting, access control for Model Context Protocol (MCP) servers, and native integration with NVIDIA Nemo TM Guardrails designed to help detect and mitigate prompt injection attacks.
A consistent cloud operating model for Agentic AI
The shift to agentic AI is expanding the infrastructure conversation beyond GPU acceleration. Enterprises need to consider the complete execution path: models, CPUs, GPUs, storage, networking, runtime security, governance, and operations.
NVIDIA Vera CPUs will bring purpose-built CPU capabilities to this emerging workload category. Pairing NVIDIA’s high-performance CPU architecture with Nutanix Cloud Platform is intended to support predictable performance and low-latency foundation for running advanced AI agents.
The future of agentic AI will depend on the infrastructure that enables those models to act securely, efficiently, and at scale. Together, Nutanix and NVIDIA are helping organizations prepare for this next phase of Enterprise AI.
©2026 Nutanix, Inc. All rights reserved. Nutanix, the Nutanix logo and all Nutanix product and service names mentioned are registered trademarks or trademarks of Nutanix, Inc. in the United States and other countries. Kubernetes is a registered trademark of The Linux Foundation in the United States and other countries. All other brand names mentioned are for identification purposes only and may be the trademarks of their respective holder(s).
This content may contain express and implied forward-looking statements, including but not limited to statements regarding our plans and expectations relating to new product features and technology under development, the capabilities of such product features and technology, and our plans to release product features and technology. Such statements are not historical facts and are instead based on our current expectations, estimates and beliefs, including statements about planned support for NVIDIA Vera based workloads and environments. The accuracy of such statements involves risks and uncertainties and depends upon future events, including those that may be beyond our control, and actual results may differ materially and adversely from those anticipated or implied by such statements, including, among others: failure to develop, or unexpected difficulties, delays or disruptions in developing, releasing or distributing, new products, services, product features or technology in a timely or cost-effective basis. Any forward-looking statements included speak only as of the date hereof and, except as required by law, we assume no obligation to update or otherwise revise any such forward-looking statements to reflect subsequent events or circumstances. Certain products and features or functionalities described herein remain in varying stages of development and will be offered on a when-and-if-available basis. The development, release, and timing of any such products, features or functionalities are subject to change. Nutanix will not have any liability for any failure to deliver or delay in the delivery of any such products, features or functionalities. Any future product or product feature information is intended to outline general product directions, and is not a commitment, promise, or legal obligation for Nutanix to deliver any functionality. This information should not be used when making a purchasing decision.