A Kubernetes factory, or a platform for your applications
Kubermatic Kubernetes Platform (KKP) is well engineered for one job, which is producing and hosting Kubernetes clusters for many tenants. Its seed architecture runs each user cluster’s control plane as pods, which keeps the cost per cluster low. Projects, quotas, metering and white-labelling make it a product for service providers, telcos and central IT teams. Kubermatic is based in Hamburg and appears in the 2026 Gartner Magic Quadrant for Container Management.
Ankra serves the team that consumes the cluster. It builds or imports Kubernetes, then turns repositories into running services, keeps the platform add-ons current, ranks vulnerabilities, estimates cost, backs up data and operates it all with AI.
If your product is Kubernetes itself, Kubermatic fits that shape. If your product is software that runs on Kubernetes, Ankra covers more of the work.
Building and running the clusters
KKP supports a long list of providers. Its top support tier covers private clouds such as OpenStack, VMware Cloud Director, Nutanix, vSphere and KubeVirt alongside AWS, Azure and GCP, with Hetzner and DigitalOcean in lower tiers. Bare metal and edge are listed as experimental. KKP creates EKS, AKS and GKE clusters as external clusters, while other clusters connected by kubeconfig are view-only. You run the master and seed clusters yourself, unless you buy the managed subscription.
Ankra builds kubeadm or k3s clusters on Hetzner, OVHcloud, UpCloud, DigitalOcean, AWS EC2 and Proxmox VE. It creates six provider-run services in general availability (EKS, GKE, AKS, DOKS, OVHcloud MKS and UpCloud UKS), plus Scaleway Kapsule and Ankra Cloud Kubernetes in closed beta. It imports any conformant cluster, including clusters KKP made. On clusters it builds, Ankra handles upgrades, node groups and power schedules, and on the five public clouds node autoscaling and rolling OS security patching, one host at a time.
Its control plane is a hosted service in the EU, so there is no management cluster for your team to run. Enterprise customers who need it inside their own perimeter can run Ankra self-hosted, including air-gapped, as an Enterprise add-on.
What happens after the cluster exists
Creating a cluster is the first hour of its life. The years after are add-on upgrades, releases, incidents and security fixes.
KKP’s Enterprise application catalogue installs tools such as Argo CD, Flux, Trivy and Falco. The Kubermatic Developer Platform adds self-service APIs for internal services. Kubermatic is open about where it stops. In its Gartner follow-up it writes that “KKP does not provide native container image management, image building, container-native storage, or CVE scanning”, calls that a deliberate boundary, and notes that Kubernetes expertise is still required. Delivery, CI and incident handling run in whichever tools you deploy and connect.
Ankra builds that work into the platform.
- Applications turn a GitHub repository into a running service and deploy every successful build. If the repository already has CI, Ankra keeps it and deploys its image.
- Ankra Pipelines run tests, rootless builds and Semgrep, Checkov and Trivy scans in your cluster, and publish only the image that passed the gate
- stacks order Helm charts and manifests by dependency, and stack profiles version them across environments
- the Security Center ranks running-workload CVEs by CISA KEV and EPSS, with optional SBOMs
- Cloud Cost estimates spend from list prices and reprices clusters on other providers
- backups write volumes and database dumps to a bucket you own and restore them in place. Databases run by the CloudNativePG or Percona operators are captured, but cannot be restored in place yet.
Some documented pipeline stage kinds do not run yet. Pull-request previews exist for Applications on GitHub, but Ankra’s own docs say not to rely on them yet while a known bug is open.
Tenants and teams
KKP’s multi-tenancy is a core strength. Projects, resource quotas, metering and group bindings let a provider or central IT team hand clusters to many tenants and charge for them. Several of those features are part of the Enterprise Edition.
Ankra added projects to its new portal in October 2026. They are rooms for teams inside one organisation, holding their own clusters, applications, pipelines and cost view, with project-only members. They are not Kubernetes as a service: there are no per-tenant quotas or metering for reselling clusters. If that is your business, KKP is the closer fit.
AI on each side
Kubermatic gives your own AI tools access rather than shipping an AI operator. K8sGPT and the Kubernetes MCP server are Enterprise catalogue apps, and KDP 1.2 exposes a remote MCP endpoint in every instance. Kubermatic AI is a separate product for pooling GPUs across teams and serving models.
Ankra’s AI is part of the platform. It drafts stacks from a description, analyses firing alerts with AI Insights, reviews pull requests on GitHub, GitLab and Bitbucket Cloud and opens fix pull requests on GitHub through the AI Gateway. Its writes wait for a person by default, under autonomy controls you set. Ankra also has an MCP server for your own tools. Its writes are audited, and your MCP client decides how to confirm them.
Virtual machines and private clouds
Kubermatic’s portfolio reaches beyond containers. Kubermatic Virtualization 1.3, released in October 2026, adds a VMware migration wizard and live VM backup to S3. For a team leaving VMware with virtual machines to move, that is a real offer.
Ankra manages Kubernetes clusters, not virtual machines as workloads. It can build clusters on Proxmox VE next to VMs you already run there.
Two kinds of openness
Kubermatic is open core. The Community Edition is Apache 2.0, while enterprise features such as multiple seeds, the application catalogue, cluster backups, quotas and policy are source-visible under a read-only licence, and using them needs an Enterprise licence agreement.
Ankra’s platform is commercial. Its CLI is Apache-2.0 and its Terraform provider is MPL-2.0. What Ankra manages stays standard Helm charts and manifests, and with a Git repository connected they live in your repository. Leaving means keeping working configuration, not exporting it.
When to choose Ankra, and when to choose Kubermatic
Choose Ankra when your team ships software on Kubernetes and wants clusters, delivery, security ranking, cost and AI operations without running a management plane.
Choose Kubermatic when
- you offer Kubernetes as a service to many tenants or business units
- control-plane density and per-tenant metering drive your costs
- you run OpenStack, VMware Cloud Director, Nutanix or vSphere private clouds
- you are moving virtual machines off VMware onto Kubernetes
- a management plane with an open source core is required
Compare the work that remains
- Create the foundation. Record who runs the management plane and the work outside the product.
- Ship a release. Include build, tests, scans and a rollback.
- Upgrade an add-on. Note who checks dependencies and how a failed upgrade is undone.
- Investigate a vulnerability. Find every affected workload and release the fix.
- Compare total cost. Include licences, infrastructure and the people who operate the platform.
Ankra’s pricing has a free tier with 30 worker vCPU. Beyond that, Startup costs €0.01 per worker vCPU-hour (about €7.20 a month), and your provider bills the servers. Kubermatic Enterprise is priced on request.
Reviewed 9 November 2026 against Kubermatic’s documentation and pricing. Product scope and plans can change.