Nutanix Agentic AI drives neoclouds, NKP Metal forges bare-metal Kubernetes
Nutanix has had plenty to say at its .NEXT 2026 conference, held in Chicago this April.
The company will now introduce new capabilities for its Nutanix Agentic AI service that are designed to help a new generation of AI cloud providers, known as neoclouds, to deliver secure, scalable AI services to AI engineers and Agentic AI developers.
What is a neocloud?
We can define a neocloud as a specialised, on-demand cloud service provider (not necessarily a hyperscaler, although the majors have a finger in this pie) that is specifically optimised for AI use cases with its fast access channels to GPU clusters and software stacks dedicated to providing AI training and inference.
Nutanix says the AI era has fueled the emergence of neocloud providers, but while demand has largely been driven by AI training workloads from a small number of large customers, the next phase of AI will centre on scaling inference and running agentic AI applications in production for a large number of enterprise customers.
“As organisations deploy and scale these agentic AI applications, they increasingly require platforms that deliver enterprise-grade security, performance, control, and self-service capabilities for developers while reducing the cost per token for AI services,” noted Nutanix, in a technical product statement.
To meet these demands, neocloud providers are evolving from GPU infrastructure providers into full AI service platforms. Nutanix will enable neoclouds to deliver a broader catalogue of AI services, including GPU-as-a-service, Kubernetes-as-a-service, and an enterprise-ready AI platform service powered by Nutanix Agentic AI.
According to Thomas Cornely, executive VP for product management at Nutanix, the Nutanix Agentic AI solution is a software stack “purposely designed to help customers accelerate adoption of agentic AI” by reducing complexity and offering lower and more predictable token costs.
Multi-cloud, multi-tenant, multi-service
The addition of a multi-tenant, multi-service portal enables neocloud providers to deliver high-value AI services on their GPU infrastructure and support sovereign AI deployments, giving enterprise users control over their data, infrastructure and AI operations.
“Demand for sovereign and specialized AI clouds is accelerating as organisations look for ways to access AI while maintaining control over their data,” said Cornely. “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.”
Nutanix Agentic AI updates will include the next generation of Nutanix’s multitenancy framework, delivered through Nutanix Service Provider Central, which is designed to help neocloud providers securely operate shared AI infrastructure at scale.
Data isolation situation
The framework introduces strong tenant isolation (which ensures security, privacy, and performance across multiple users) and granular resource management (which optimises efficiency by precisely allocating compute and storage), allowing providers to host multiple enterprises on the same physical GPU infrastructure while maintaining predictable performance, security… and, logically then, data isolation.
With these capabilities, neocloud builders will be able to allocate GPU and compute resources dynamically across tenants, enforce tenant-specific security and networking policies and enable independent AI environments for each customer with a catalogue of GPU-aaS, K8S-aaS, VM-aaS, Notebooks-aaS, VectorDB-aaS and Models-aaS.
McKinsey analyst Scott Sinclair thinks that 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,” said Sinclair, who is practice director for infrastructure, cloud, DevOps and networking.
Complementing the new multitenancy capabilities, enhancements to Nutanix Cloud Manager (NCM) help service providers operate and monetize AI infrastructure as a service. NCM offers monitoring of AI infrastructure and adds usage-based metering, enabling providers to track and bill customers based on GPU usage, API calls, or model consumption.
Together, capabilities here are promised to enable providers to manage capacity, monitor tenant usage, and operate distributed AI infrastructure through a unified management interface, helping neocloud builders deliver scalable AI services while maintaining operational control.
NKP Metal
Nutanix also announced NKP Metal, which extends the Nutanix operating model and Nutanix Kubernetes Platform (NKP) solution to support Kubernetes deployments directly on bare-metal infrastructure.
The company explains that running Kubernetes on bare metal can deliver the performance and flexibility many modern workloads require, particularly for edge environments and AI training workloads that rely on dense GPU infrastructure.
But operating these environments at scale often introduces new complexity, from provisioning physical servers to managing firmware updates and integrating storage and networking services. As a result, many organisations end up having to build a highly specialised and siloed team for managing bare-metal Kubernetes deployments.
Dual-native architecture
Unlike solutions which are strictly hypervisor-based or Kubernetes-based, NKP Metal supports a “dual-native architecture” here, in which containers and virtual machines operate as first-class infrastructure under a unified operating model, including for AI and other performance-intensive workloads that often run directly on bare-metal infrastructure.
NKP Metal represents an extension of the Nutanix operating model and HCI stack to bare-metal Kubernetes environments, enabling organisations to run containers directly on physical infrastructure while maintaining a consistent level of automation, lifecycle management, networking, and enterprise data services they rely on in virtualised environments.
As part of this approach, customers can choose to consume Nutanix storage through a container storage interface or use Cloud Native AOS as a purpose-built storage option for true bare-metal Kubernetes deployments while leveraging Nutanix Data Services for Kubernetes-native data services, extending the Nutanix experience end to end while keeping storage closer to Kubernetes workloads.
“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.”
With NKP Metal, organisations can deploy and manage containerised workloads on physical servers while maintaining the operational simplicity, automation, and enterprise services of the Nutanix Cloud Platform solution.
NKP Metal will also simplify the lifecycle management of physical infrastructure. Leveraging capabilities such as automated node deployment with Nutanix Foundation and Operating System and Firmware lifecycle management through Lifecycle Manager, organisations will be able to provision, scale, patch, and update bare-metal Kubernetes environments while retaining operational consistency used for virtualised workloads.
Key takeaways
We’re clearly seeing a shift happening in the market i.e. enterprise organisations are moving away from AI training and moving towards inference and agentic AI applications in production. With neocloud popularisation rising, Nutanix wants to make sure we think of its Agentic AI stack as suitable for neocloud providers to evolve beyond raw GPU infrastructure into full AI service platforms. The multi-tenant, multi-service portal (Nutanix didn’t say multi-cloud for once, but we can take that as read) will be a welcome addition, especially for customers who seek to build sovereign AI deployments.
Plus, hey, DevOps teams always appreciate usage-based billing and capacity management tools, so the enhanced tooling in Nutanix Cloud Manager will appeal, especially for providers who want to monetize AI services effectively while (and there’s that term again) maintain strict data sovereignty. With air-gapped and other deployments always in need of bare metal services (think about warships, nuclear power stations, highly-secured medical research facilities and so on) NKP Metal, which extends the Nutanix Kubernetes toolkit and platform to bare-metal infrastructure will be welcome to those who want edge and GPU-intensive AI workloads to run with operational simplicity.
The service’s dual-native architecture is also interesting (where Kubernetes and virtual machines run as first-class infrastructure) as its another step towards infrastructure automation and the management of physical servers, which modern platform teams are certainly interested in.
