Why Kubernetes Licensing Is Mostly Hidden Cost
Kubernetes licensing is a phrase that confuses procurement, because Kubernetes is open-source and free to licence. The cost is everywhere else — the enterprise distribution, the management plane, and above all the infrastructure the clusters consume. The managed control plane on Amazon EKS or Google GKE costs about $0.10 per cluster per hour, roughly $73 a month; Azure AKS gives the standard control plane away. Those figures look harmless, and that is the trap: the control plane is less than 5% of total Kubernetes spend, while the other 95% — compute nodes, persistent storage, load balancers, inter-zone traffic, registry pulls and logging — is where the bill actually lives.
This is the classic emerging-tech pricing pathology our emerging technology contracts guide tracks: a trivial headline unit hiding an enormous aggregate. Containers are also the substrate your API gateways, observability agents and DevOps tooling all run on, so Kubernetes spend compounds with every adjacent platform metered per node or per container. Treat it as a portfolio cost to be consolidated and committed, not a free open-source freebie.
OpenShift, Tanzu and Managed Kubernetes: 2026 Pricing
Enterprises mostly choose between a paid distribution and a cloud-managed service. Red Hat OpenShift — an IBM property — sells self-managed subscriptions at roughly EUR 12,000-30,000 per year for six core-pairs with a Premium 24/7 SLA, with most enterprise pricing custom-quoted; the managed ROSA service on AWS charges about $0.171 per hour per 4 vCPU of worker nodes plus a $0.25 per hour hosted-control-plane fee, and 1-year or 3-year commitments cut the worker-node fee by 33-55%. VMware Tanzu, now under Broadcom, licenses Tanzu Mission Control per managed cluster — around GBP 40-80 per cluster per month — or per CPU core under management, with a free Starter tier for attaching clusters.
The cloud-managed services from AWS, Microsoft and Google Cloud price the control plane flat and the nodes on consumption. The Broadcom shift matters most here: components are being folded into VMware Cloud Foundation and re-sold as add-ons above the base VCF subscription, a bundling move we cover in detail in the VMware Tanzu licensing guide.
| Platform | 2026 Pricing Reference | Negotiation Lever |
|---|---|---|
| EKS / GKE control plane | $0.10 / cluster / hour (~$73/mo) | Consolidate clusters; reserved nodes |
| Azure AKS control plane | Free (standard tier) | Pay only for nodes — right-size them |
| OpenShift self-managed (Premium) | EUR 12K-30K / yr per six core-pairs | Core-pair count; custom quote benchmark |
| ROSA (managed) | $0.171/hr per 4 vCPU + $0.25/hr HCP | 1-3 yr commit (33-55% off worker fee) |
| Tanzu Mission Control | GBP 40-80 / cluster / month | Unbundle from VCF; per-core audit |
The licence is the visible 5%; the nodes are the hidden 95%. A Kubernetes negotiation that fixates on the distribution subscription and ignores node sprawl, reserved-capacity terms and egress is optimising the wrong number.
The Control-Plane Illusion and Node Sprawl
The most expensive mistake in container economics is anchoring on the control-plane fee. At $73 a month per cluster it looks immaterial — until an enterprise runs forty clusters across teams and regions and the per-cluster fees, each with their own idle worker nodes, compound. One widely cited example shows an EKS footprint reaching $438 per month before a single line of business workload runs. Multiply unmanaged sprawl across dozens of clusters and the gap between a consolidated estate and a fragmented one runs into seven figures a year.
Node sprawl is the real leak: over-provisioned worker nodes, reserved capacity bought and never used, and inter-zone traffic that no one budgeted. The Broadcom-era Tanzu repricing adds a second trap — buyers who accept Tanzu inside a VMware Cloud Foundation renewal pay for management capability they may already get free from their cloud provider. Both leaks are addressable, and both belong in the contract, not just the FinOps dashboard. The Cloud Contract Framework sets out the commitment-and-exit structure these deals should inherit.
Negotiation Levers for Kubernetes Licensing
Four levers move a containerization and Kubernetes licensing position. First, consolidate clusters — fewer, larger clusters cut per-cluster control-plane fees and reduce idle node overhead, often recovering 20-40% of the run rate before any vendor conversation. Second, commit the nodes: 1-year and 3-year reserved instances or savings plans, and ROSA service-fee commitments at 33-55% off, convert predictable compute into discounted compute. Third, unbundle the distribution: benchmark OpenShift core-pair pricing and Tanzu per-cluster pricing against the cloud-managed alternative, and refuse to let Tanzu ride into a VCF renewal unbenchmarked.
Fourth, use the open-source floor — upstream Kubernetes and a cloud-managed control plane are a credible alternative to a paid distribution, and a documented migration analysis disciplines an OpenShift or Tanzu quote the same way our low-code and data fabric guides use credible alternatives as leverage. Because containers are bought platform-by-platform and grow silently with the node count, most enterprises are over-provisioned on compute and over-bundled on the distribution. If your organisation is renewing OpenShift, Tanzu or a managed Kubernetes commitment, request a confidential briefing and our cloud contract negotiation team will benchmark the distribution, commit the nodes, and consolidate the estate.