Leaving VMware is a project, not a switch — but a well-run migration is far less daunting than an open-ended Broadcom renewal. This guide sets out the phases a successful VMware exit follows, what each delivers, and how to de-risk the move, then points you at the tools to size and cost it. Start by sizing the target cluster on our VMware alternatives calculator.
Start with an honest assessment
Before choosing a platform, inventory what you actually run: VM counts and their vCPU, RAM and storage, the dependencies between systems, and everything wrapped around vSphere — backup, monitoring, automation. This inventory sizes the whole project, and it almost always reveals over-provisioning that shrinks the replacement cluster. A tool like RVTools exports it quickly; feed the totals into the calculator to see the hosts and cost.
The assessment is also where you pick the target — matched to your team and estate, not the loudest marketing. Proxmox suits cost-driven, open-source-comfortable teams; Nutanix is the closest turnkey HCI replacement; Azure Local fits Microsoft-centric shops. Our guide to the best VMware alternative walks the choice.
Design the target, then prove it with a pilot
Design the cluster — host spec, storage model, networking and backup — and prove it with a small pilot before committing the estate. Migrate a handful of non-critical VMs, rehearse the cutover, and validate backup, monitoring and performance on the new platform. A pilot surfaces awkward dependencies and day-two gaps while they are still cheap to fix.
This is also when you retrain the team and rebuild automation. The hypervisor is rarely the hard part; the surrounding operational muscle memory is, so budget real time for it rather than assuming a like-for-like swap.
Migrate in waves, not a big bang
Move workloads in waves grouped by risk and dependency — least-critical first, tightly-coupled systems together. Free tooling does most of the lifting: Nutanix Move, Veeam or StarWind for cross-hypervisor conversion, or storage-level replication. Each wave rehearses the next, so speed and confidence build as you go.
Keep VMware running alongside until each wave is verified. A phased migration means you never bet the whole estate on one weekend, and you can pause a wave if something needs attention without derailing the programme.
Cut over, then retire the licences
Once a wave's VMs are validated on the new platform, cut over production and decommission the VMware capacity — the moment the Broadcom saving actually starts. You only stop paying for capacity you have vacated, so sequence the waves to retire whole hosts (and their billed cores) rather than leaving half-empty clusters licensed.
After the estate has moved, optimise: right-size VMs, tune the storage and fold the platform into standard operations and support. The 72-core minimum means a tidy, fully-retired host is worth more in saving than a lightly-used one.
How Servnet runs it
We supply the hardware across Dell, HPE, Lenovo and Nutanix, arrange finance to spread it, and run the migration end to end — assessment, design, wave cutovers and hypercare. See our VMware-to-Nutanix, VMware-to-Azure-Local and VMware-to-Proxmox programmes, and start by sizing the target on the calculator.