We have the same conversation at least once a month. A VMware renewal quote lands, someone does a double take, and the question becomes "what are our options?"
For a lot of organisations the answer is Hyper-V. Not because it's the cheapest option, although it often is, but because it's now a capable, supportable platform that most IT teams already half know.
Over the last two years we've migrated estates from VMware to Hyper-V ranging from 10 VMs to close to 1,000. What we've found is that the hypervisor is rarely the hard part. The planning, the dependencies and the backup strategy are where these projects succeed or fail.
What actually changed
Since Broadcom completed its acquisition of VMware in November 2023, the commercial model has been rebuilt from the ground up. In short:
- Perpetual licences have gone. Everything is now a subscription.
- The SMB bundles have gone. Essentials Plus, which suited a lot of two and three host clusters, has been withdrawn.
- vSphere Standard is at the end of the road. It stops at vSphere 8, which reaches end of general support on 11 October 2027. After that there are no new patches. To keep receiving security updates, customers need to move to a vSphere 9 subscription, and Broadcom's direction of travel is clearly VMware Cloud Foundation (VCF).
- Products are bundled. You buy per core, whether you need everything in the bundle or not.
- Late renewal costs more. Miss the renewal date and there's a reported 20% penalty.
The result is that renewals for smaller estates have commonly come in at several times their previous cost. Larger organisations have more negotiating room, but even they are being pushed towards bundles they didn't ask for.
That October 2027 date is the one to watch. VCF is built for large private clouds, with software-defined networking, storage and automation included. For a simple environment of a few hosts running file servers, line-of-business apps and a SQL box, it's overkill. You end up paying for a platform you'll use a fraction of, just to keep your hosts patched.
To be fair to VMware, vSphere is still an excellent platform. The issue isn't the technology. It's that the cost no longer matches what a lot of organisations actually use.
What we've learned from 10 to 1,000 VMs
A 10 VM migration and a 1,000 VM migration are different projects, but the same things tend to catch people out.
| Consideration | Why it matters |
|---|---|
| Know what you've got | Most estates have VMs nobody can explain. A proper discovery exercise usually finds things to retire, which shrinks the project before it starts. |
| Map the dependencies | Applications rarely live on one VM. Move the database without the app server and you find out quickly. We group migrations by application, not by host. |
| Check the guest OS | Older Windows and some Linux builds need drivers or integration services sorted before cutover, or they won't boot cleanly. |
| Check software licensing | Some vendors license by host, socket or hypervisor. Oracle, SQL Server and line-of-business apps all need checking before you move. |
| Plan the network | VLANs, port groups, static IPs and firewall rules all need translating. This is where small details cause big outages. |
| Decide on storage early | Cluster Shared Volumes, SAN or hyper-converged. The storage design shapes everything else, including performance and recovery. |
| Agree the downtime windows | Most VMs can be cut over in minutes, but someone needs to agree when and in what order. |
| Upskill the team | The day-to-day tooling is different. Failover Cluster Manager, Windows Admin Center and PowerShell replace vCenter. Budget time for this. |
One thing I've learned: the bigger the estate, the more the project becomes about scheduling and communication rather than technology. The conversion itself is the easy bit.
Why backup vendor support matters more than the hypervisor
If there's one thing to check before anything else, it's your backup platform.
Your backup is your safety net during the migration and your recovery plan afterwards. Change the hypervisor without thinking about it and you can end up with a platform that's cheaper to run but harder to recover. That's where the problem starts.
What we look for:
- Native Hyper-V support. Not agent-based workarounds. Image-level, application-aware backups with changed block tracking, just as you had on VMware.
- Cross-platform restore. Platforms such as Veeam can restore backups as Hyper-V VMs. That gives you a rollback route and, in some cases, a migration route in its own right.
- Feature parity. Instant recovery, granular file and application restores, immutability and ransomware detection should all work on the new platform. Check, don't assume.
- History. Your existing VMware backups still need to be restorable for as long as your retention policy says. Plan how you'll keep them.
- Licensing. Some backup products license per socket, some per workload. A hypervisor change can quietly change your backup bill too.
The good news is that the backup platforms we work with, including Veeam and Rubrik, support Hyper-V well. The risk is in assuming every feature you rely on today carries over. More often than not it does, but it's worth proving before cutover, not after an incident.
We always run a test restore on the new platform before we call a migration finished. It's the cheapest insurance in the whole project.
Hyper-V isn't what you remember
We see this regularly. Someone used Hyper-V on Windows Server 2008 or 2012, had a bad experience, and hasn't looked at it since. That's fair. Early versions were clunky, management was fiddly and clustering could be painful.
The truth is that Hyper-V has moved on a long way, and Windows Server 2025 is a big step forward:
| Area | What's changed |
|---|---|
| Scale | Up to 2,048 vCPUs and 240 TB of RAM per VM, and clusters of up to 64 nodes and 8,000 VMs. Scale isn't the limiting factor for anyone we work with. |
| GPU partitioning | One physical GPU can be shared between several VMs, and those VMs now support live migration and high availability. Useful for VDI, CAD and AI workloads. |
| Workgroup clusters | Clusters with live migration no longer need Active Directory, using certificates instead. Handy for branch sites and isolated environments. |
| Dynamic processor compatibility | Live migration between hosts with different CPU generations without switching off performance features across the board. Makes hardware refreshes much easier. |
| Security by default | Generation 2 VMs are the default, with Secure Boot and support for virtualisation-based security. |
| Migration tooling | Microsoft has released a free VM Conversion extension for Windows Admin Center that replicates VMs from vCenter with minimal downtime. It's still in preview, so we use it alongside proven methods rather than relying on it alone. |
There's also a commercial point people miss. If you already license Windows Server Datacenter for your hosts, Hyper-V is included, and so is unlimited Windows Server guest licensing on those hosts. For a lot of organisations, the hypervisor cost simply disappears.
In our experience, once IT teams spend a week or two with a well-built Hyper-V cluster, the preconceptions go quickly.
When Hyper-V isn't the answer
Hyper-V isn't right for everyone, and we'd rather say so up front.
- Heavy use of the wider VMware stack. If you rely on NSX networking, vSAN or VMware's automation tools, there's more to replace than the hypervisor.
- Applications certified only on VMware. Some vendors are still catching up. Check support statements before you commit.
- Very short timescales. If your renewal is weeks away, a rushed migration is a bigger risk than one more year of subscription. A short renewal while you plan properly is often the sensible move.
- Moving to the cloud anyway. If the plan is to move most workloads to Azure in the next couple of years, it may make more sense to go straight there.
Where to start
If your VMware renewal is in the next 12 months, start now. Not because migration is hard, but because good options disappear as the renewal date gets closer.
- Find your renewal date and what you're actually paying per core.
- Run a discovery of your VMs, dependencies and licensing.
- Check your backup platform supports Hyper-V with the features you rely on.
- Compare the full cost over three to five years: hypervisor, Windows Server, backup, hardware and project effort.
- Pilot a few non-critical workloads before committing the whole estate.
We're happy to have an honest conversation about whether Hyper-V makes sense for you, even if the answer is "not yet". Get in touch to talk it through.