Nutanix expands Kubernetes & AI push for neoclouds

Nutanix has introduced NKP Metal for running Kubernetes on bare-metal infrastructure and outlined plans to expand its Agentic AI offering for a new class of AI cloud providers known as neoclouds.

The announcements extend Nutanix's reach beyond virtualised environments into physical infrastructure and hosted AI services, two areas where demand is rising as companies build AI systems and edge deployments.

Platform expansion

NKP Metal adds bare-metal support to the Nutanix Kubernetes Platform, allowing organisations to run container workloads directly on physical servers. It uses the same operating model as virtualised environments, including automation, lifecycle management, networking and data services.

The product targets customers that want Kubernetes on physical hardware for workloads such as AI training and edge computing, where GPU density and lower overhead can matter. Nutanix argues that these environments often create operational complexity because IT teams must manage servers, firmware, storage and networking outside the standard tools used for virtual machines.

NKP Metal supports what Nutanix describes as a dual-native architecture, in which containers and virtual machines operate side by side under one operating model. Customers can use Nutanix storage through a container storage interface or Cloud Native AOS for bare-metal Kubernetes deployments, alongside Nutanix Data Services for Kubernetes-native data services.

"Running Kubernetes on bare metal has traditionally meant sacrificing the operational simplicity of virtualised environments," said Dan Ciruli, Vice President and General Manager, Cloud Native, Nutanix. "With NKP Metal, we're extending the Nutanix operating model to bare-metal Kubernetes, combining automated lifecycle management with integrated Cloud Native AOS data services to deliver the simplicity, consistency, and enterprise storage capabilities customers need on their physical infrastructure."

The product uses Nutanix Foundation for automated node deployment and Lifecycle Manager for operating system and firmware updates. The aim is to let organisations provision, patch and scale bare-metal Kubernetes clusters using processes closer to those already used in virtual infrastructure estates.

Michael Warrilow, Chief Researcher at Virtified, said companies increasingly run containers across physical, virtual and cloud environments and need consistency in how those clusters are created and managed.

"Today, organisations run their containerised workloads across physical and virtual environments that could be in a data centre, at the edge, or in the cloud. Nutanix calls this a dual-native architecture. I call it a necessity that requires consistency in creating and managing Kubernetes clusters wherever they are. The key to success is finding a trusted and proven platform capable of meeting these needs with operational consistency-be it virtual, bare metal, or cloud," Warrilow said.

Neocloud push

Alongside the Kubernetes update, Nutanix will add new functions to its Agentic AI software aimed at neocloud providers, a term used for AI cloud operators that rent out GPU infrastructure and related services. The changes are intended to help those providers move beyond raw GPU access and offer a broader set of managed AI services.

That catalogue is set to include GPU-as-a-service, Kubernetes-as-a-service and an AI platform service built on Nutanix Agentic AI. The software stack is designed to support the deployment of agentic AI applications while giving customers more control over data, infrastructure and operations.

A central part of the update is a multitenant, multiservice portal delivered through Nutanix Service Provider Central. The framework is designed to let neocloud operators run shared AI infrastructure securely across multiple enterprise customers with tenant isolation and resource controls.

Providers will be able to allocate GPU and compute resources across tenants, apply tenant-specific networking and security policies, and create separate AI environments for each customer. The service catalogue can also extend to VM-aaS, Notebooks-aaS, VectorDB-aaS and Models-aaS.

Enhancements to Nutanix Cloud Manager are also intended to support the commercial side of hosted AI services. The software can monitor AI infrastructure and provide usage-based metering, allowing providers to track and bill customers based on GPU use, API calls or model consumption.

"Demand for sovereign and specialised AI clouds is accelerating as organisations look for ways to access AI while maintaining control over their data," said Thomas Cornely, Executive Vice President, Product Management at Nutanix. "The Nutanix Agentic AI solution, with its secure multitenant and AI management portal, is designed to enable neocloud providers to rapidly deliver advanced high value AI services to enterprises and public sector organisations looking for powerful AI capabilities from trusted regional providers."

The neocloud strategy reflects a broader shift in the AI market, where providers are trying to capture more value than simple infrastructure rental. As enterprises move AI projects from pilot stages into production, they are seeking hosted environments with clearer governance, isolation and billing models, particularly for regulated or sovereign workloads.

Scott Sinclair, Practise Director, Infrastructure, Cloud, DevOps, and Networking at Omdia, said the spread of autonomous agents is creating new demands around security, governance and predictable performance.

"The deployment of autonomous agents is rapidly becoming the next frontier in enterprise AI, but this rise is introducing significant new risks related to data security, governance, and unpredictable performance. Organisations cannot manage this transformation on legacy infrastructure. Given these demands, Nutanix's focus on strong governance, security, performance, tenant isolation, and predictable resource management in its purpose-built Agentic AI solution provides a welcome option for CIOs as they seek to deploy an enterprise-grade foundation for their AI agent strategy," Sinclair said.

Both product lines are being offered first through early access programmes, with broader release planned in the second half of 2026. The staged rollout suggests Nutanix is testing demand in two adjacent markets at once: organisations that want Kubernetes directly on hardware, and service providers looking to package AI infrastructure into managed products.

For Nutanix, the combined message is that customers want one operating approach across virtual machines, containers, physical servers and AI infrastructure, even as those workloads spread across data centres, edge sites and specialist cloud operators.

"As we integrate AI into our core business processes and customer applications, security and control are non-negotiable. The Nutanix Agentic AI solution provides the secure, governed infrastructure foundation that allows our teams the self-service capability they need to accelerate the development and deployment of AI innovation across our production environments," said Jasim Adbul Rahman, Group Chief Information Officer, Power International Holding.