9.5 Nutanix Kubernetes Platform, Enterprise AI, and End-User Computing

9.5 Nutanix Kubernetes Platform, Enterprise AI, and End-User Computing

These three items sit in Objective 1.2 next to NUS, NDB, and NCM. They are the portfolio's newest members, which is exactly why they are worth studying: material written for NCA 6.5 or 6.10 either omits them or uses retired product names.

Nutanix Kubernetes Platform (NKP)

NKP is the Nutanix enterprise Kubernetes offering. It supersedes the earlier Nutanix Kubernetes Engine, which itself was previously branded Karbon — if you see "Karbon" or "NKE" in a study resource, that resource is out of date.

What NKP does:

• Automates cluster deployment. Kubernetes clusters are provisioned to a consistent standard instead of being hand-built.
• Manages the full lifecycle. Upgrades, scaling, and node pool management are driven from the platform rather than assembled from scripts.
• Runs consistently across locations. The same platform operates in the datacenter, at the edge, and on NC2, which is the point of a "platform engineering" product: one operating model everywhere.
• Ships a curated application catalog of the observability, ingress, and security components a production cluster needs.

NKP is licensed in Starter, Pro, and Ultimate tiers, matching the pattern used by NCI and NCM.

Nutanix Data Services for Kubernetes (NDK)

Containers are stateless; the applications inside them frequently are not. NDK connects Kubernetes workloads to the Distributed Storage Fabric so that persistent volumes, snapshots, and data protection work the same way for containers as for VMs. It is listed as its own reference in the blueprint ("Nutanix Data Services for Kubernetes Purpose"), so recognize the name and its job: stateful data services for containerized applications.

Konnector and NKP

Recall from section 9.2 that the Nutanix Konnector service in Prism Central and Nutanix Central automatically registers NKP clusters as they are created, so they appear in the management console without manual onboarding. That is the specific link between NKP and the management plane.

Nutanix Enterprise AI (NAI)

Nutanix Enterprise AI lets an organization run AI inference on infrastructure it controls. The use case is governance-driven: many organizations want to use large language models against their own data but cannot send that data to a public inference API for regulatory, contractual, or competitive reasons.

What to recognize:

• It deploys and serves models on your own hardware, including GPU-equipped Nutanix nodes, and exposes them through standard inference endpoints so existing applications can call them.
• It keeps data and models inside your security boundary, which is the whole point — the alternative is sending sensitive data to a third-party endpoint.
• It runs alongside traditional workloads on the same platform rather than requiring a separate AI-only stack.

[!TIP] The NCA-level question here is almost always a use-case match: "an organization wants to use an LLM against confidential internal documents without sending data to an external service." The answer is Nutanix Enterprise AI.

End-User Computing and NCI-VDI

The final Objective 1.2 item is "identify End-User Computing use cases and features." Section 1.3 introduced the branding; here is what to recognize.

Licensing. End-user computing is sold as NCI-VDI, metered per user rather than per core. That matters because desktop projects are budgeted by headcount, and per-user licensing lets desktop growth be costed directly.

Broker support. Nutanix does not replace the connection broker. It runs the ones customers already use:

• Citrix Virtual Apps and Desktops
• Omnissa Horizon (formerly VMware Horizon)
• Nutanix Frame, a desktop-as-a-service offering delivering desktops and applications through a browser

Platform features that matter to VDI:

Note how NUS reappears here: a VDI deployment usually needs both NCI-VDI for the desktops and Nutanix Files for the profile and home directory data. Objective 2.4 asks you to identify the correct license by use case, and this pairing is a realistic version of that question.

Portfolio Recognition Summary