The enterprise infrastructure landscape has been in flux since Broadcom changed the rules on VMware customers, and many companies are looking for cost-effective, safe exit options for different VMware workloads. Enter Mirantis with their latest announcement of k0rdent Enterprise and k0rdent Virtualization —a comprehensive solution that promises to unify AI, container, and VM-based workloads under a single, open source platform.

In an exclusive interview with TFIR, Dominic Wilde, SVP of Marketing at Mirantis, outlined how their Kubernetes-native approach addresses the growing enterprise demand for VMware alternatives while positioning organizations for the AI-driven future of computing.

The VMware Exodus: A Market Opportunity

The timing of Mirantis’ announcement couldn’t be more strategic. “The world is looking for alternatives to VMware,” and the company is positioning itself as the bridge between traditional virtualization and modern cloud-native infrastructure.

“We wanted to ensure that all the years of training and the experiences people have had while working on VMware—and the workflows they’ve worked so hard to build—are not wasted,” Wilde explained. “We’re not asking people to tear everything down and start from scratch.”

This approach addresses one of the biggest concerns enterprises face when evaluating alternatives: the fear of losing existing investments in training, processes, and workflows. Mirantis has created what Wilde calls “a very viable alternative for those who are feeling that they are now priced out of VMware.” The Broadcom acquisition of VMware has created significant disruption in the enterprise virtualization market.

Unification Strategy: Beyond Just Virtualization

What sets k0rdent apart from other VMware alternatives is its unified approach to infrastructure management. Rather than simply replacing VMware with another virtualization platform, Mirantis is addressing the broader challenge of managing diverse workload types across hybrid environments. The shift mirrors broader industry trends, as evidenced by discussions around OpenShift vs. VMware alternatives.

“Over time, we’ve seen a fragmentation of toolchains, toolsets, and management platforms available to organizations,” Wilde noted. “k0rdent was essentially born out of the belief that AI is the next killer app for the cloud.”

The platform leverages Kubernetes-native functions and capabilities, particularly the Cluster API (CAPI), to deliver a common control plane that allows organizations to treat all their cloud real estate as a single entity. This is particularly relevant as Gartner research shows that more than 95% of new digital workloads will be deployed on cloud-native infrastructure by 2025.

AI Infrastructure: The Driving Force

The complexity of AI workloads is becoming a significant driver for infrastructure unification. As Wilde explained, AI inference workloads are highly distributed, pushing out to the edge while requiring orchestration at incredible scale. This creates challenges around data sovereignty, GPU utilization, and multi-tenancy that traditional virtualization platforms weren’t designed to handle. The intersection of AI/ML and Kubernetes platforms is becoming increasingly important for enterprise infrastructure strategies.

“Many service providers and large enterprises are deeply concerned about what we call hard multi-tenancy—true isolation of workloads,” Wilde said. “In the Kubernetes world, a single cluster with namespaces no longer suffices. You actually need a cluster per application, or an application per cluster.”

This requirement for massive multi-cluster deployment aligns with current Kubernetes trends showing enterprises will begin deploying AI agents across departments as large language models become faster, cheaper and more efficient. The approach also connects to broader discussions about AI/ML scaling through Kubernetes.

Open Source: The Strategic Differentiator

Perhaps most importantly, Mirantis is betting on open source as the key differentiator in a market where vendor lock-in has become a primary concern. The company has contributed mature technologies like K0s and k0smotron to open source, integrating them into the k0rdentplatform. This builds on Mirantis’ broader open source strategy, including their recent k0rdent project that addresses Kubernetes and AI sprawl.

“We strongly believe that the community should drive technology solutions, rather than vendors dictating the toolsets, toolchains, and management interfaces,” Wilde emphasized. “Customers want to know they have an on-ramp—but also an off-ramp, should they choose to take that path.”

This approach directly addresses the trust issues that have emerged in the wake of the VMware-Broadcom acquisition. “There’s been a breakdown of trust with vendors,” Wilde observed. “One of the first questions we get is, ‘If I choose you and, in three years, I want to leave, how does that work?'”

Market Positioning and Go-to-Market

Mirantis is targeting a broad range of enterprises, with initial traction in SaaS companies, financial services, cloud service providers, and telecommunications companies. The company’s strategy is built around enabling customers to “start simple and then scale fast,” with typical deployments happening over weeks for smaller scales and months for massive scale migrations.

The platform’s composability allows organizations to maintain existing VMware environments while gradually migrating workloads, providing the flexibility that enterprises need during transition periods.

Looking Ahead: The Kubernetes-Native Future

As the infrastructure industry continues its shift toward cloud-native architectures, Mirantis is positioning k0rdent as more than just a VMware replacement—it’s a platform for the AI-driven future of computing. The company’s strong belief that “infrastructure, particularly AI infrastructure, should be Kubernetes native” aligns with broader industry trends showing Kubernetes becoming the standard orchestrator for diverse workload types. This evolution builds on fundamental concepts explored in the art of Kubernetes blueprints and recent developments like Kubernetes 1.33’s native sidecars.

With the enterprise infrastructure market in transition and AI workloads driving new requirements, Mirantis’ timing appears well-calibrated. Whether k0rdent can deliver on its ambitious promise of unifying traditional and modern workloads while maintaining the familiar experience VMware users expect will ultimately determine its success in this competitive landscape.

The question for enterprises isn’t whether they need to evolve their infrastructure strategy—it’s whether they can afford not to, especially as the demands of AI workloads continue to reshape the requirements for scalable, efficient, and cost-effective infrastructure management.

Related TFIR Articles:

• OpenShift vs. VMware – The Enterprise Shift and the Road Ahead
• Mirantis k0rdent delivers Kubernetes-native platform engineering across any infrastructure
• AI/ML Scaling Made Simple: Harnessing Kubernetes
• Kubernetes 1.33: Native Sidecars are here, plus Big Security Boost

Interview Transcript

Swapnil Bhartiya: Welcome to a brand new episode of KubeStruck. Mirantis has announced k0rdent Virtualization, which is seen as a major step towards unifying infrastructure for AI, containers, and VM-based workloads. In today’s discussion, we are going to break down what this means for enterprise IT. What does this mean for those businesses who have been longtime VMware customers who are looking at alternatives after the Broadcom acquisition? We will also talk about whether this offering by Mirantis is a legitimate VMware alternative or VMware killer.

We’ll also talk about why Kubernetes-native infrastructure can be the foundation for the next wave of cloud-based innovation.

To go deeper into this topic, today we have with us once again, Dominic Wilde, SVP of Marketing at Mirantis. Dominic, it’s great to have you back on the show.

Dominic Wilde: It’s great to be here. Thanks, Swapnil. It’s my pleasure.

Swapnil Bhartiya: Today, you folks are announcing k0rdent Virtualization. If I’m not wrong, it’s based on k0rdent Enterprise. First of all, let’s just give folks a bit of background about k0rdent, k0rdent Enterprise, and what led to this announcement today.

Dominic Wilde: I think you’ll remember, we announced k0rdent as an open source project back in February, and k0rdent provides the ability for platform engineers to do large-scale multi-cluster management. We also then followed quickly with another announcement, articulating that we were really focusing on k0rdent in supporting AI infrastructure rollout, and we’d added some functionality around AI inferencing. The whole platform went GA back in February.

So fast forward to today: k0rdent has, from the very beginning, been able to help with managing and deploying virtualization workloads, as well as containerized workloads, bare metal, etc. But today we’re following up with an announcement around further integration around virtualization. Most specifically, what we’ve done is created a more familiar user experience and workflow for those platform engineers and admins who have been working with VMware in the past. We’ve made it a more familiar experience, and that’s basically the basis of the announcement today—creating that user interface experience.

Swapnil Bhartiya: When I was reading the press release, I see that it’s unifying infrastructure for modern workloads—AI, containerized, and also VM-based. What is the driver behind this unification? Why is it so important to bring all these infrastructures together?

Dominic Wilde: It’s not really just for AI, but overall, what we’ve seen happen over the course of time is a sort of fragmentation of the tool chains, the tool sets, and the management platforms available to organizations. Up to this point, there has been a hard choice: if you choose to be in a hybrid environment, you may find yourself working with tools and interfaces that are helping you manage the public cloud and a different set of tools that are helping you manage on-premise cloud and workloads.

k0rdent was birthed out of a belief that AI is the next big killer app for cloud, and because it adds complexity in terms of pushing a lot of the inferencing workloads to the edge—and now there’s greater focus on things like data sovereignty—now is the right time to be focused on creating a unified management experience for the hybrid cloud experience, public cloud experience, and private cloud experience.

What customers need in terms of operationalizing AI today is a centralized platform that delivers a common control plane that allows them to treat all of their cloud real estate as a single entity. k0rdent does that in a very non-invasive, standardized way. We’ve leveraged Kubernetes-native functions and capabilities, particularly the Cluster API (CAPI). We introduced k0rdent back in February as an open source initiative, and the whole point behind that was we strongly believe that the community should drive a technology solution, rather than vendors deciding on the tool sets, tool chains, and management interfaces.

Swapnil Bhartiya: From a business perspective, when they look at this unified approach, how do they see this solving their problems? What problems is it solving for them?

Dominic Wilde: From a business perspective, there are several things that, if I’m sitting in the C-suite or boardroom, I’m worried about. I’m worried about growing the top line and new streams of revenue. That speaks directly to my organization’s ability to innovate and innovate fast. In a world where AI is coming up very quickly and creating disruption but also tremendous opportunity in the market, every organization needs to be able to innovate rapidly.

k0rdent was designed to solve underlying infrastructure problems and abstract the complexity of infrastructure, allowing developers to move much faster in terms of their innovation agenda. Secondly, you want to reduce the bottom line as well—take out as much expense and create as much efficiency as possible.

As you start to look at how to scale out these distributed AI services and broader legacy services, you start to realize this creates a scale-out problem in terms of complexity of the infrastructure and how to manage it. Because we take a declarative approach and a templatized approach, we enable you to create a library of operational services that can quickly be scaled out in a repetitive, reliable, and robust manner. We enable you to do more with less.

The other big challenge from a business perspective is mitigating risk. Every C-suite, every boardroom, every business owner is worried about risk today—not just from a security perspective, but increasingly from a regulatory perspective. We’re living in a fairly complex world in terms of geopolitics, and we need to make sure we’re regulatory compliant and that as we scale and innovate fast, we’re not breaking any rules.

Our approach with k0rdent is to deliver a very robust, reliable, and low-risk capability to scale out businesses rapidly and accelerate innovation.

Swapnil Bhartiya: These days, when we talk about virtualization or VMware workloads, one word that pops up in everybody’s head is VMware, especially after the whole Broadcom acquisition. Because of the change in pricing and licensing concerns, folks are looking at alternatives. This is also good timing for you folks because people looking at alternatives can look at k0rdent and say, “Hey, this is what we can rely on.” Is that correct?

Dominic Wilde: It’s absolutely spot on and in the right space. In delivering k0rdent Virtualization, we’ve created a user experience that is very akin to what a broad swath of customers have become familiar with in terms of their VMware operationalization.

We wanted to ensure that all the years of training and things that people have gone through as they’ve worked on VMware, and their workflows that they’ve worked hard on, are not wasted. We’re not asking people to rip things up and start all over again. Indeed, we’re coming in with a very viable alternative for those who are feeling that they are now priced out of VMware. Let’s be frank about it—that’s what the concern is.

We’ve now created an alternative for them that is based purely on open source, is fully composable, gives them choice, allows them to keep parts of the environment that they currently use (including VMware), but allows them a smooth migration path and a familiar workflow and familiar interface as well.

We’re strong believers that businesses very rarely rip things up and start all over overnight. This will be a transition over time. There are not just VMware challenges happening, but also new challenges with AI. As people start to look at re-architecting or re-looking at how they’re doing things, we’re offering this open source, composable approach that’s very familiar to them.

Swapnil Bhartiya: When we look at VMware alternatives, folks are still trying to figure out what’s there, how viable it is, how mature it is, because VMware solutions have been mature. If you look at this offering, would you call it a VMware killer? Folks also look at Kubernetes and Red Hat OpenShift as well. How do you position it?

Dominic Wilde: As I said, customers have had a wake-up call. The whole VMware debacle has been a wake-up call, and fundamentally, there’s been a break of trust with vendors. One of the first questions we get is, “Well, if I choose you, and in three years I want to get off you, how does that work?” Customers want to know that they have an on-ramp, but also an off-ramp should they choose that path.

Because we’ve taken an open source and standardized approach, and our belief is that customers should be following community-driven technology evolution rather than vendor-opinionated vendor stacks, we’ve taken the approach of using standardized pieces. The virtualization that we’re delivering is based on KubeVirt, which is open source. We’ve integrated that into the k0rdent environment to add additional value around scalability, rollout, deployment, and day-zero, day-one, day-two capabilities.

We believe that the path forward is giving customers choice rather than locking them in. There are many other vendors who send an open source message and say they pull from upstream, but they pull from upstream and then basically closed-source things. Sure, there are advantages for specific use cases, but if you want a solution that fits your business rather than trying to get your business to fit a solution, that is what Mirantis is delivering.

We are delivering the capability and a platform that is flexible enough to fit your business, to allow you to have choice without adopting unnecessary risk, and also being affordable. At the end of the day, if you need that off-ramp in three or five years, or whatever it is, that’s available to you because this is all really based on open source.

Swapnil Bhartiya: When we talk about businesses, what kind of businesses or use cases do you feel will be served best with this offering?

Dominic Wilde: This is broadly applicable to all enterprises. We’re getting initial traction with SaaS companies, financial services, CSPs, and telcos. It’s a broad range of different customers who are using our solutions.

But in general, our approach is that most customers really want to start simple and then scale fast. That’s our mantra. We’ve designed k0rdent and these virtualization capabilities being packaged on top so that customers can try this out at a smaller scale, be fully in control, and get themselves comfortable.

In general, we would see a deployment and a full migration happening, on average, over the course of a number of weeks. For a much larger scale, maybe a couple of months, because we also offer migration services for a massive scale. We’ve worked on that with customers as well. But in general, most customers want to come in, try it out, and then once they get comfortable and understand how this serves their business, they want to be able to scale fast. That’s the part that k0rdent adds—that infinite scalability.

Swapnil Bhartiya: Why is it so important to enable both traditional and modern workloads? We’re looking at AI workloads, we’re looking at VMware-based workloads…

Dominic Wilde: Ultimately, the value is efficiency. When I boil it down to the most basic value add, it is really about being able to do more with less. A single platform, a single set of tools, and a single operational model is always more efficient and more cost-effective. It gives you a platform to build upon and innovate on.

Now, are there going to be examples where customers need to keep multiple tools or multiple platforms, or have different stacks supporting different capabilities? Yeah, of course, that’s going to happen, and customers should not be forced in that instance—if it serves their business, they should not be forced to move to a single model.

Because k0rdent is built on open source and we’re using open standards, interfaces, and APIs, we can very easily integrate with existing management platforms and stacks as well and add value. This is one of the things we actually do with the public cloud, which is a very opinionated and somewhat closed system. We will integrate, for instance, with the Kubernetes provided by Amazon, Google, or Azure, but equally, if you find yourself frustrated or wanting to move away from using the Kubernetes distributions provided by those public cloud providers, we can deploy your Kubernetes of choice using k0rdent as an overlay onto the public cloud and run that way, and normalize a single operational model as well.

This hybrid example is a good example of what you’re talking about. A lot of customers will be very happy using the tools provided by Amazon. If they aren’t, we provide an alternative as well. We’ll integrate with Amazon—we’ve done all that integration work—but if you want to move to a single operational model, then that is available to you too.

Swapnil Bhartiya: I want to also talk about AI-specific workloads, because conversations about AI are easy, but actually the workloads are complex and resource-intensive. How is your Kubernetes and cloud-native approach going to further simplify these modern use cases?

Dominic Wilde: As you say, we were the original private cloud pioneers. Back in the day, we’ve got many years of experience and thousands of installations of experience servicing customers. We’ve learned a lot over the years.

One of the things we’ve clearly seen when it comes to AI workloads is there are many different factors and complexities. One of the key things is that the overall model and overall architecture is very distributed, particularly when you start to talk about inference. Training is more centralized, but inference starts to push out and be very distributed.

When you put that in the context of a regulated, highly regulated environment and this geopolitical environment we live in, it causes a lot of complexity. How do you ensure data sovereignty? How are you ensuring that you’re using the right model at the right place, that you’re running on the right GPU and using that GPU efficiently, that you’re pulling data from the right place? All of this has to be orchestrated and coordinated at an incredible scale. The number of workloads running out there is growing exponentially.

One of the use cases we’ve seen is that service providers and big enterprises are very concerned about what we call hard multi-tenancy—true isolation of workloads. Where we’ve done that before in the Kubernetes world, in a single cluster with namespace, that doesn’t cut it anymore. You actually need a cluster per application, or application per cluster.

Once you start to do that, you then move into a world of massive complexity and scale. You need to be able to get your arms around that and not have that be the second-order problem that kills you. k0rdent was designed for that and designed for this AI use case, in terms of being able to scale massively on the multi-cluster side, do that in a very declarative way, be composable, and push out as far to the edge as you need.

Also be able to take, for instance, an inference workload, and say, “Hey, I want to be able to pre-define the models I’m using, the GPUs I’m using, and get the greatest utilization I can out of the underlying infrastructure and hardware while making sure that it’s all orchestrated properly.”

So lots of big, complex problems, but we feel we’re providing a great start for organizations to start practically addressing those in the real world.

Swapnil Bhartiya: If you look at Mirantis’ family of Kubernetes-related projects—K0s, k0smotron, k0rdent, and of course Lens—how does all of it fit together to help enterprises in their journey?

Dominic Wilde: Actually, when we introduced k0rdent back at the beginning of the year and launched and went GA with k0rdent, we had taken mature technology that we have used and matured over many years, including K0s and k0smotron, and we contributed those to open source. So those are now sandbox projects, because we feel it’s really important to innovate in the open, and we’ve integrated those into the k0rdent platform.

So k0rdent is actually built of building blocks that are mature technologies that we’ve nurtured and matured over the years, that we’re now contributing to open source. All of these building blocks, all of these pieces, we’re driving towards a world where our strong belief is that infrastructure, particularly AI infrastructure, should be Kubernetes-native, that Kubernetes is the common denominator and the common orchestrator and language for AI infrastructure of the future.

This is why, in building an AI platform and in building a multi-cluster platform like k0rdent, we’ve ensured that this and all of the other pieces that we’re working on and that we’ve contributed are all founded in a Kubernetes-native paradigm.

Swapnil Bhartiya: One last question: you folks made the announcement today. Talk a bit about how mature it is for businesses to start using it.

Dominic Wilde: This is actually quite mature. As I said, we went GA a while back, and we’ve been working with customers. This announcement is really based on the virtualization capabilities—we’ve now created a user interface and a user experience for the virtualization piece of k0rdent that enables it to be familiar to VMware users.

So this is quite mature. This is out there, we have customers using it. The underlying technology was GA months back, but this is more about the user experience. This is more about now being able to provide UIs and things that make it a familiar operational experience.

Swapnil Bhartiya: Dominic, thank you so much for joining me today and giving us an update on k0rdent. Thank you so much, and I look forward to chatting with you again.

Dominic Wilde: Thanks, Swapnil. Appreciate it.