Mirantis k0s and k0rdent Achieve CNCF Certified Kubernetes AI Conformance, Validating Enterprise-Ready AI Infrastructure from Cluster to Fleet

Independent certification confirms Mirantis delivers AI-ready Kubernetes infrastructure for training, inference, GPU orchestration, and distributed AI workloads at both the cluster and platform level

CAMPBELL, Calif.--(BUSINESS WIRE)--Mirantis, delivering Kubernetes-native infrastructure for AI, today announced that both k0s, its lightweight single-binary Kubernetes distribution, and k0rdent, its open-source distributed container management platform, have achieved CNCF Certified Kubernetes AI Conformance at Kubernetes v1.35.

The certifications independently validate that organizations can run modern artificial intelligence and machine learning workloads on Mirantis infrastructure software using standard upstream Kubernetes – from individual clusters to globally distributed AI fleets – without proprietary dependencies or operational complexity.

The public certification submissions are fully reproducible and available on the k0s documentation site and the k0rdent documentation site.

As organizations standardize AI development, model training, inference serving, and agentic applications on Kubernetes, the CNCF AI Conformance program establishes an objective benchmark for determining whether a platform can support production AI workloads. Unlike vendor claims, certified platforms provide public evidence that can be independently verified and reproduced.

“Organizations investing in AI infrastructure need confidence that their Kubernetes foundation can support production AI workloads without locking them into proprietary platforms or creating operational complexity,” said Randy Bias, vice president, Open Source Strategy and Technology, Mirantis. “Achieving CNCF Certified Kubernetes AI Conformance for both k0s and k0rdent validates Mirantis’ full-stack AI infrastructure strategy – from the individual cluster to fleet-scale operations – built entirely on open, upstream Kubernetes.”

"As AI workloads scale in production, platform teams require a consistent, vendor neutral infrastructure layer to eliminate the guesswork of managing distributed GPU workloads,” said Chris Aniszczyk, chief technology officer, CNCF. “The Kubernetes AI Conformance program relies on community reproducible verification so that lightweight distributions like k0s – and the management platforms like k0rdent that orchestrate them – prove they support upstream Kubernetes components out-of-the-box. This objective validation confirms that community-driven standards remain the definitive foundation for responsible scaling in the agentic era."

Independent Validation for Real-World AI Workloads

The CNCF Certified Kubernetes AI Conformance program defines a baseline set of capabilities required to run AI and machine learning workloads on Kubernetes. Vendors must prove these capabilities through public, repeatable testing.

The certifications confirm that k0s and k0rdent-managed clusters can:

• Securely allocate and manage GPU resources for AI workloads;
• Automatically install and manage required GPU drivers;
• Route inference and model-serving traffic using the Kubernetes Gateway API;
• Support distributed AI training through gang scheduling;
• Dynamically scale clusters and workloads based on GPU demand;
• Deliver comprehensive GPU observability, including utilization, memory consumption, temperature, and power metrics;
• Run advanced AI frameworks and operators, including KubeRay, with Kubernetes self-healing capabilities;
• Support GPU sharing through time-slicing to improve infrastructure utilization.

The k0s submission validated all mandatory AI workload requirements defined by CNCF. The k0rdent submission extended beyond the baseline to cover additional optional AI capabilities, demonstrating its ability to operationalize AI infrastructure at larger production scale.

From Lightweight Kubernetes to Fleet-Scale AI Operations

For platform teams, the certifications prove a critical architectural point: simplicity does not require compromise.

k0s packages the entire Kubernetes control plane into a single binary, reducing operational overhead while preserving full compatibility with upstream Kubernetes. The AI conformance results demonstrate that all core technologies required for AI infrastructure – including GPU scheduling, traffic routing, distributed job orchestration, observability, and AI operators – run on k0s without proprietary extensions.

k0rdent builds on that foundation by provisioning and managing fleets of standardized k0s clusters using the open source Kubernetes Cluster API. Because k0rdent-managed clusters inherit the same AI-ready foundation, organizations can scale from a single validated AI cluster to globally distributed AI factories using the same operational model.

Unlike many AI infrastructure platforms, k0rdent requires no custom abstractions to satisfy AI conformance. Standard Kubernetes autoscaling, monitoring, and networking tools can be used, which simplifies production deployment across cloud, edge, and bare metal.

Building the Foundation for AI Infrastructure at Scale

The certifications validate Mirantis’ architectural approach: open, composable, upstream compatible Kubernetes infrastructure that can support the full AI lifecycle – from experimentation and training to inference and multi-cluster AI operations.

For organizations building AI factories, the result is a standardized path to deploy, operate, and scale GPU-powered AI infrastructure with independent CNCF validation at every layer of the stack.

About Mirantis

Mirantis delivers the fastest path to scalable, efficient GPU cloud infrastructure for neoclouds and enterprise AI factories, with full-stack AI infrastructure technology that removes complexity and streamlines operations across the AI lifecycle, from Metal-to-Model. Through k0rdent AI and strategic partnerships, Mirantis helps organizations maximize GPU utilization, improve cloud economics, and accelerate time to revenue with production-grade multi-tenancy, intelligent workload orchestration, and automated operations.

Mirantis serves many of the world’s leading enterprises, including Adobe, Ericsson, Inmarsat, MetLife, PayPal, and Societe Generale. Learn more at www.mirantis.com.

Mirantis is a registered trademark of Mirantis, Inc. Metal-to-Model is a trademark of Mirantis, Inc. All other trademarks are the property of their respective owners.