News

Nutanix Expands Cloud Platform for Production Agentic AI

• By David Ramel
• 08/26/2026

Nutanix is expanding its Nutanix Cloud Platform around production agentic AI, releasing Nutanix Enterprise AI 2.8 while preparing Nutanix Kubernetes Platform 2.19 for general availability. The updates are intended to let organizations run AI alongside existing virtualized and containerized applications while applying common management and governance.

The company announced the platform enhancements Aug. 26. NAI 2.8 is generally available now, while NKP 2.19 is described only as "available soon." Nutanix also announced general availability of Service Provider Central, or SP Central, a multitenant control plane for infrastructure, application, cloud-native and AI services.

The broader architecture treats virtual machines and containers as first-class infrastructure, with Nutanix describing the approach as "dual-native." Nutanix said the objective is to let AI workloads run near existing applications and data without requiring organizations to move those workloads into a single deployment model. "Enterprise AI should not require customers to rebuild the systems that already run their business," Thomas Cornely, executive vice president of product management at Nutanix, said in the announcement.

NAI 2.8 Adds MCP Governance
For developers working with agentic systems, one of the main additions in Nutanix Enterprise AI 2.8 is generally available MCP server management in Nutanix Agent Gateway. Agent Gateway serves as a centralized connection point between AI agents and MCP servers, including servers deployed locally within a Nutanix Enterprise AI environment or remotely. Administrators can assign tool permissions to specific users or API keys, while locally deployed MCP servers support rolling updates.

NAI 2.8 also extends Nutanix Private Inference. Nutanix lists fine-tuning for models with fewer than 8 billion parameters and support for deploying NVIDIA NIM microservices in air-gapped environments. Multi-node and multi-GPU inference, which is intended to support models with more than 100 billion parameters, is in tech preview. KV cache offloading from GPU memory to CPU host memory is also in tech preview. Nutanix specifically warns that tech preview features should not be used in production environments.

The MCP governance work follows Nutanix's Aug. 10 release of its open-source MCP server for NCP. That software uses the Nutanix Prism v4 API to let AI agents and developer tools interact with the platform. Nutanix identified GitHub Copilot, Claude Code and Cursor as examples of AI assistants that can use the server. The Prism v4 API Gateway handles execution, governance and security controls, including role-based access, throttling, metering and auditing.

The combination gives Nutanix two MCP-related components with different roles: the NCP MCP server exposes Nutanix infrastructure operations to compatible AI tools, while Agent Gateway provides centralized governance for agent access to MCP servers and the tools and data behind them.

NKP 2.19 Targets AI and Bare-Metal Kubernetes
NKP 2.19, which Nutanix said will be available soon, is expected to expand container management across virtualized and bare-metal environments. The planned release includes NKP Metal, NKP running on the AHV hypervisor, and an AI Applications Catalog for deploying validated AI and machine learning software. Nutanix named Kubeflow, Milvus and Slurm among the software available through that catalog.

NKP Metal is intended to extend the Nutanix operating model directly to physical servers. Nutanix previously said the technology uses Nutanix Foundation and Lifecycle Manager capabilities to automate node deployment and operating system and firmware lifecycle management. NKP on AHV, meanwhile, can be combined with Nutanix Flow for network-level isolation of AI agents. Nutanix also said NKP has received CNCF Kubernetes AI Conformant Platform certification.