Clusters deployed by k0rdent are now Certified Kubernetes AI Conformant
)
)
Summarize with AI
){kind=icon}
Claude){kind=icon}
ChatGPT){kind=icon}
Gemini){kind=icon}
Perplexity){kind=icon}
GrokRecently, clusters deployed and lifecycle-managed by k0rdent became Certified Kubernetes at v1.35. k0rdent has now cleared the next bar: the CNCF Certified Kubernetes AI Conformance program. The submission, k0rdent v1.9.0 provisioning Kubernetes v1.35 clusters, is public and reproducible in the cncf/k8s-ai-conformance repository.
k0rdent is an open-source platform for deploying and managing fleets of Kubernetes clusters. In plain terms, this certification is independent confirmation that every cluster k0rdent provisions is a reliable place to run modern AI workloads: model training, inference serving, and AI agents.
What this means
The Certified Kubernetes AI Conformance program, launched at KubeCon North America in November 2025, defines the baseline capabilities a platform needs to run AI and machine learning workloads, and asks vendors to prove them with public, re-runnable evidence instead of marketing claims.
There was a prerequisite to clear first. A platform has to be a Certified Kubernetes distribution before it can pursue AI conformance, and k0rdent earned that earlier this year. With that in place, the remaining work was layering the AI capabilities onto a cluster shape we had already proven out.
We ran the proof on a real, k0rdent-provisioned cluster on Azure with NVIDIA T4 GPUs. It showed that the clusters k0rdent deploys can:
Hand GPUs to workloads safely, giving each container access to exactly the accelerators it requests and managing the drivers automatically.
Route model and inference traffic through a standard Gateway API, the same one k0rdent itself uses internally.
Schedule large distributed jobs with all-or-nothing ("gang") scheduling, and scale clusters and pods up and down as GPU demand changes.
Show what the hardware is doing, feeding per-GPU utilization, memory, temperature, and power into standard monitoring alongside workload metrics.
Run complex AI tooling like KubeRay, reliably and with automatic recovery.
Where k0rdent's submission stands out is scope. Because it is a full platform rather than a bare distribution, it went past the mandatory requirements and covered the optional, nice-to-have capabilities too, such as sharing a single GPU across several workloads. The one exception is virtualized GPUs, which depend on licensing and hardware outside this submission's test cluster.
Two things are distinctly k0rdent's:
Autoscaling that comes for free. k0rdent builds clusters using Cluster API, the Kubernetes standard for cluster lifecycle, so the upstream autoscaler plugs in without any custom glue, on any cloud or on bare metal.
A clear production path for monitoring. The test cluster used the popular open-source monitoring stack, but k0rdent's own observability product, KOF, speaks the same language and is the recommended choice for production fleets.
The full requirement-by-requirement breakdown lives on the k0rdent documentation site.
Why it matters
The most important thing this certification demonstrates is architectural: k0rdent needed no custom abstractions to satisfy AI conformance. The clusters it provisions are standard Kubernetes, and the rest of the ecosystem, from autoscaling to monitoring to traffic routing, already knows how to operate them.
That is the payoff of building on standard foundations. Independent verification of it, by a CNCF reviewer rerunning public documentation, is worth far more than a vendor datasheet.
There is a fleet dimension too. k0rdent's job is provisioning and managing many clusters at once. This certification confirms that every cluster it stamps out is a verified, standardized place to run AI workloads, not a bespoke environment each team has to relearn.
Resources
k0rdent AI Conformance submission: the public submission and per-requirement evidence (v1.35/k0rdent).
k0rdent Certified Kubernetes entry: the prerequisite K8s conformance listing (PR #4167).
k0rdent documentation: the AI conformance evidence page, with per-requirement detail and reproduction steps.
k0rdent on GitHub: source, releases, issues.
CNCF Certified Kubernetes AI Conformance: the program, instructions, and the full list of certified platforms.
)
)
)
)
)
