Mirantis

Mirantis is a founding member of the Open Secure AI Alliance, now hosted by the Linux Foundation (@linuxfoundation).

An AI system runs on a full stack of model, harness, runtime, and infrastructure. It is only as secure and as open as the layers underneath.

We are advancing that work from the infrastructure upward, grounded in what we know best. That means turning hard-won operational experience into open artifacts that others can deploy, test, and trust, instead of keeping it in private runbooks.

For the operators who run this infrastructure, you can reproduce a deployment in an environment you control and inspect the evidence behind every claim.

See what open AI infrastructure really means. buff.ly/AfSo2tD

#Mirantis #k0rdent #OpenSource

A layer you can’t inspect is a layer you can’t predict, automate or recover when it fails.

That’s why full-stack control matters for anyone operating an AI factory.

Mirantis k0rdent AI delivers it through a composable platform that integrates components of NVIDIA DSX OS™, the open, modular software portfolio of the NVIDIA DSX AI Factory Platform, into an interoperable stack operators can inspect, extend and adapt.

NVIDIA Run:ai provides AI workload and GPU orchestration within NVIDIA DSX OS - and it is now open source. By extending its integration to include NVIDIA Run:ai, Mirantis gives operators a consistent operational model from bare-metal provisioning and lifecycle management to workload orchestration and GPU allocation.

See what full-stack control looks like. buff.ly/phVSI8a

#Mirantis #k0rdentAI #NVIDIA #Runai #OpenSource @nvidia

We are excited to announce that Mirantis k0rdent AI has been validated in the NVIDIA-Certified Hypervisors program.

The k0rdent AI virtualization stack has been certified through the NVIDIA-Certified Hypervisors program on the NVIDIA GB200 NVL72 validation track. GPU virtualization can help reduce the traditional performance trade-off.

What you get:
• Predictable, near bare metal performance
• GPU allocation matched to each workload, full node or sub-node
• Strong tenant isolation with the VM as the tenancy boundary
• One control plane for VMs and containers

Learn more about running virtualized GPU workloads on k0rdent AI. buff.ly/dAEofSx

#Mirantis #k0rdentAI #NVIDIA #GPUVirtualization @nvidia

k0s is now Certified Kubernetes AI Conformant. 🎉

Big milestone for the project: the CNCF has certified k0s under its Certified Kubernetes AI Conformance program at Kubernetes v1.35. It's a more demanding bar than standard Kubernetes conformance, and it gives independent, public verification that the AI and ML workloads organizations run today (model training, inference serving, and AI agents) operate reliably on k0s using standard, upstream tooling.

The certification was validated on a live cluster: k0s running on NVIDIA T4 GPUs on Azure. k0s met every mandatory capability, along with optional ones such as GPU time-slicing.

A lean, single-binary distribution often raises the question of what has been sacrificed to keep it small. For AI workloads, the answer is nothing of consequence. Any project that runs on upstream Kubernetes runs on k0s, with no proprietary components to adopt.

For teams evaluating where to run training and inference, this offers independent, reproducible verification rather than a vendor claim.

Read the full breakdown: buff.ly/nfJNAcs

Clusters deployed by k0rdent are now Certified Kubernetes AI Conformant. 🎉

Here's the question every platform team asks. If we standardize our clusters on one tool, are we quietly giving something up when it's time to run AI?

The CNCF just answered it for k0rdent. Its clusters passed Certified Kubernetes AI Conformance at v1.35, validated on a live cluster with NVIDIA T4 GPUs. And it's not a datasheet claim, but independent, public proof that anyone can rerun.

What it means if you run AI on k0rdent:
• No lock-in. The clusters are standard Kubernetes, so your existing tooling for GPU scheduling, traffic routing, and monitoring just works. Nothing proprietary to adopt.
• Every cluster, verified. k0rdent's job is managing many clusters at once. Now each one it provisions is a verified, standardized place to run AI, so your teams don't relearn a bespoke environment every time.

k0rdent covered all mandatory capabilities plus optional ones like sharing a single GPU across workloads.

Read the full breakdown: buff.ly/LMstn14