Executive Summary
Cloud-native infrastructure stopped being a set of projects and became a set of commitments. The orchestrator is settled. The migration off traditional virtualization is not. Storage and security kept arriving as afterthoughts on both, and that is where most of the cost hides.
This is the map of those commitments, with the posts that go deep on each one. Read it as a starting point rather than a verdict. The board has been moving for three years and it has not stopped.
The cloud-native conversation has narrowed. A few years ago it was a list of projects competing for attention. Now it is a short list of decisions that most platform teams have already made, plus two or three they keep deferring.
Kubernetes won the orchestration argument. That is no longer interesting. What is interesting is what winning changed, which is that every other layer now has to justify itself against a default that already works.
The decisions that are actually open
Three of them. Where workloads run, now that leaving VMware is a live option for almost everyone. How storage and identity follow those workloads, which nobody scoped properly at the start. And whether the platform is something you build or something you buy.
Each one is bigger than the tool choice it looks like. The migration decision sets your cost curve for years. The storage and identity decision decides whether the migration holds. The build-or-buy decision decides how fast your team can move on everything after it.
Where to start depending on where you are
If you are mid-migration, the cost models are the first thing to fix. Most of them count the license and miss the years of process debt behind it. We walked that math in the full VMware exit cost.
If sovereignty is on your desk, the useful distinction is not residency. It is who can be compelled to hand data over, which data residency does not answer. The wider sovereignty picture is worth reading alongside it.
If you run Kubernetes at fleet scale, the runtime question is settled and containerd and CRI-O both won. The open question is fleet management, which is where the operating cost actually sits.
If AI is landing on your platform, start with the mismatch. Kubernetes for AI is a different job than Kubernetes for apps, and the defaults fight the workload shape.
If supply chain is the pressure, an SBOM is the beginning of the conversation rather than the end of it. Auditors want proof, not a longer list.
What the data says about the shape of the market
The 2026 State of Enterprise Infrastructure report pulls the public numbers together. 86 percent of organisations are actively reducing their VMware footprint and only 4 percent have finished, which tells you the migration is a multi-year unwind rather than an event. Data center power is the other constraint, because AI capacity competes for the same megawatts as everything else.
Those two facts set most of the agenda. Budget follows the exit, and power follows the budget.
How to use this
Pick the decision that is actually open on your desk and follow the link for it. The rest can wait, and most of it should. A platform team that tries to settle all three at once usually settles none of them well.
Related reading. Lock-In Escape Plans Work, If You Build Them First. Cloud Native Transformation Is an Operating Model Change, Not a Lift and Shift. The Container Runtime Wars Are Over. containerd and CRI-O Both Won.. The 2026 State of Enterprise Infrastructure.
Get the next one before it is old news
Independent analysis of cloud-native infrastructure, Kubernetes and data center economics. No vendor spin.
[…] reading. The VMware Exit Has a Shortlist. Nutanix Just Took a Front Seat.. The Cloud-Native Landscape in 2026, and the Three Decisions Still Open. A Full VMware Exit Costs More Than Three Years of Broadcom’s Bill. The 2026 State of Enterprise […]