Mirantis

AI infrastructure margins are won or lost at the utilization layer. GPU cloud capacity is growing fast. But more GPUs do not automatically mean more revenue. In multi-tenant AI environments, fragmented scheduling, siloed clusters, and limited workload visibility create hidden margin leakage. Join us on April 8 to explore how AI infrastructure teams improve utilization and turn GPU capacity into a profitable platform. We’ll cover: • Where margin erosion happens in multi-tenant GPU clouds • Why GPU-as-a-Service alone cannot guarantee monetization • How GPU-aware orchestration improves utilization across tenants • Why a centralized platform layer is required to coordinate clusters and workloads Register now: buff.ly/2hzcDvU

AI Observability: Tools and Best Practices via @MirantisIT mirantis.com/blog/ai-observ…

What a week it’s been! ✨Wrapping up NVIDIA GTC 2026. Thank you to everyone who stopped by Booth 102 and shared how you’re building and scaling AI infrastructure. This week reinforced what we see across the market: running GPUs at scale takes more than hardware. It requires disciplined cluster automation, multi-tenant control, and operational consistency. That’s where Mirantis helps teams turn GPU infrastructure into repeatable, profitable services. We’re looking forward to continuing these conversations beyond GTC. If we didn’t get a chance to connect, reach out here: buff.ly/DScji47

Day 3 at NVIDIA GTC 2026 and the momentum hasn’t slowed down. 🙌 We’ve been having strong conversations with teams building AI platforms and scaling GPU environments in production. If you haven’t stopped by yet, now’s the time. Join us for a live demo, pick up a limited-edition Mirantis t-shirt, and enter the raffle for a custom skateboard. Visit Booth 102 before the swag runs out!

KubeCon + CloudNativeCon Europe 2026 is almost here! 🌷 If you haven't registered for a demo at our Booth 492 yet, now's the time. We are running live sessions every hour, just 5–10 minutes, featuring k0rdent, k0s, and k0rdent AI. Seats fill up fast at 15 per session, and registered attendees get first pick on exclusive t-shirts. 👕 Plus, every demo session is a chance to win a limited-edition skateboard deck. And if you're just passing through, stop by anyway for a collectible k0-b robot key ring. 🤖 See you in Amsterdam! Register for a demo: buff.ly/3R8vCI1 #KubeCon #CloudNative #k0rdent #k0s #Kubernetes

Supercharging Supermicro Bare Metal with Mirantis k0rdent AI for Sovereign AI and Hybrid GPU Clouds via @MirantisIT mirantis.com/blog/superchar…

Day 2 at @nvidia GTC 2026 and the conversations are getting deeper. 🔥 We’ve been discussing multi-tenant isolation, faster provisioning, and how to turn AI infrastructure into repeatable service delivery models. If you’re at GTC, stop by Booth 102. Let’s talk about what’s working, what’s slowing you down, and how to move from experimentation to disciplined operations that make your GPU investment financially sustainable. Prefer to lock in time? Book a meeting at your convenience: buff.ly/yD2QI36

What’s got people buzzing at NVIDIA GTC 2026 this year? Bruno Teodoro, VP of EMEA Business Development at Mirantis, shares what he’s seeing across the show floor. AI infrastructure is everywhere. The bigger question is how organizations modernize the infrastructure behind it. At NVIDIA GTC 2026, Mirantis is showing how teams build multi-tenant AI platforms, isolate workloads, and improve GPU utilization on Kubernetes. Stop by Booth 102 to learn how Mirantis can help you run AI infrastructure more efficiently. buff.ly/Mvs3lEu

Building a GPU cloud is the easy part. Making it a profitable business is the hard part. Gartner projects $2.5 trillion in worldwide AI spending by 2026, and demand is accelerating. Yet many neocloud providers and NVIDIA Cloud Partners remain constrained by manual provisioning, single-tenant architectures, and fragmented tooling. Teams spend months engineering the platform before generating the first dollar of cloud revenue. Today Mirantis announced it has joined the NVIDIA AI Cloud Ready Initiative as a founding ISV partner. It is the only open, composable infrastructure platform in the founding group. Mirantis k0rdent AI is a turnkey Metal-to-Model™ platform built for the neocloud lifecycle. Powered by NVIDIA NCX Infra Controller, it enables providers to move from zero to a production-ready AI cloud in days: • Automated bare metal provisioning and hardware lifecycle management • Multi-tenant GPU clouds with virtualization and hardware-enforced isolation • Self-service AI platforms with a composable, extensible service catalog • Validated on NVIDIA Hopper, Blackwell, and Grace Blackwell NVL72 Whether starting at Stage 1 or scaling to a full self-service AI cloud, Mirantis k0rdent AI provides a clear path to higher GPU utilization, stronger margins, and faster time to revenue. 📍 Join us at NVIDIA GTC 2026 at Mirantis Booth 102 Learn more: buff.ly/tuOIcZn #AI #Kubernetes #K8s #AIInfrastructure #k0rdentAI #MLOps #AIStack #NVIDIA #NVIDIAGTC

GPU Cloud investment is surging. But are providers truly innovating, or just struggling to balance costs? Mirantis CTO Shaun O'Meara takes the Main Stage at the Xcelerated Compute Show in New York to break down the state of the GPU Cloud market and the critical shift from commodity Bare Metal services to high-value, specialized offerings. Unleashing GPU Cloud Economics 📅 Monday, March 23 | 9:30 am - 10:00 am | Main Stage Catch Shaun at our booth right after. Come with your questions! Or book a meeting in advance: buff.ly/fncbDM8

Are you enjoying your time at @nvidia GTC 2026? How would you describe the vibe at the show? Ian Martinez, VP of Alliances at Mirantis, shares what he’s seeing at NVIDIA GTC 2026. The entire AI ecosystem is here. Hardware vendors, OEMs, orchestration platforms, networking, security, and data infrastructure all in one place, building the systems that will power the next wave of AI. If you’re exploring how to run AI workloads at scale, stop by Mirantis at Booth 102. buff.ly/ArMCtHc #AI #Kubernetes #K8s #AIInfrastructure #k0rdentAI #MLOps #AIStack #NVIDIA #NVIDIAGTC

In an interview with PangeaGlobe, @MirantisIT Co-founder and CEO Alex Freedland discusses Mirantis’ partnership with #neoclouds, a new class of AI infrastructure providers powering the next wave of #AIScaling. Read here: pangeaglobe.com/articles/miran…

NVIDIA GTC 2026 is live and the conversations at Booth 102 are already in full swing! 🚀 We’re speaking with teams building AI platforms and operating large GPU environments. The focus is shifting from GPU access to utilization, service delivery, and long-term profitability. If you’re at GTC, stop by for a live demo and a practical discussion on improving GPU utilization and delivering infrastructure as a service. Or schedule time for a personalized consultation focused on your specific GPU and AI infrastructure goals: buff.ly/GNXlsQ4 See you at Booth 102!

Today we’re announcing that Mirantis has joined the @nvidia AI Cloud Ready Initiative as a founding ISV partner. The goal is to solve AI infrastructure complexity at the platform level. Mirantis k0rdent AI now uses NVIDIA NCX Infra Controller, the open source lifecycle automation technology used inside NVIDIA to operate large GPU fleets. It provides a single control plane that orchestrates the stack across the NVIDIA reference architecture portfolio including Ampere, Hopper, Blackwell, DGX, HGX, MGX, and Grace Blackwell NVL72. What operators get out of the box: ✔ Zero-trust hardware lifecycle automation with NVIDIA NCX Infra Controller ✔ Hardware-enforced tenant isolation using NVIDIA BlueField DPUs at the silicon level ✔ Unified Ethernet, InfiniBand, and NVLink fabric orchestration configured automatically ✔ Full Metal-to-Model operations from OS provisioning to NVIDIA NIM inference services As part of the NVIDIA AI Cloud Ready validation, Miranits k0rdent AI was tested on the Grace Blackwell NVL72 reference architecture. 📍 See it live at NVIDIA GTC 2026 at Mirantis Booth 102 Deep dive: buff.ly/nAgOuQQ

AI isn't a model. It's a stack. And if you want to run AI workloads in production, you need every layer from the hardware up. The "five-layer cake 🍰" has become the go-to framework for understanding AI infrastructure. We fired up some GPUs and visualized what that stack actually looks like when you're building it for real: • Bare Metal: the physical compute foundation • K8s: Kubernetes clusters orchestrating at scale • AI Services: cloud-native inference and model serving • Services & Applications: where value reaches the user Every layer pulls on the one beneath it. Skip a layer and the whole thing collapses. You can't serve the top of the cake without baking everything underneath. Whether you are a platform engineer, an AI builder, or exploring how AI systems move from GPU to production, this is the stack. If you are at NVIDIA GTC 2026, stop by Booth 102 to see how this stack runs in practice with Miranits k0rdent AI. 📆 For a deeper technical discussion tailored to your environment, book a meeting here: buff.ly/MxkLxg7 #AI #Kubernetes #K8s #AIInfrastructure #BareMetal #CloudNative #FiveLayerCake #k0rdent #MLOps #AIStack #NVIDIA #NVIDIAGTC2026

Mirantis, providing Kubernetes-native AI infrastructure, revealed that Supermicro's modular server architecture has been validated with Mirantis k0rdent, which automates processes for cloud environments with hybrid GPUs and sovereign AI. @MirantisIT HostingJournalist.com/tech-wire/mira…

By combining Supermicro hardware with @AMD GPU operator-driven acceleration and @MirantisIT k0rdent automation, organizations can consolidate their legacy workloads and modern AI applications onto a unified bare-metal platform. Learn more: hubs.la/Q046C0k00 PR: hubs.la/Q046BQ-_0 Blog: hubs.la/Q046C2Jy0