After Broadcom Bought VMware: Renew, or Move to Proxmox VE, Nutanix, Hyper-V or XCP-ng?

What changed

Broadcom completed its acquisition of VMware on 22 November 2023. Within a few months the way VMware is sold had changed completely:

• Perpetual licences were discontinued. It's now subscription only, on one-year or multi-year terms.
• Products were folded into bundles, mainly VMware Cloud Foundation (VCF) and VMware vSphere Foundation (VVF), instead of being bought piece by piece.
• Pricing is per CPU core, counting every core in the host, with a minimum of 16 cores per CPU. An 8-core CPU is billed as 16.
• The partner programme and sales channels were reorganised.
• The vSphere Essentials Plus kit, the usual choice for small offices with two or three hosts, was discontinued in late 2024.
• Version 9 (VCF 9.0 from mid-2025) is delivered through the VCF and VVF bundles.
• Support dates matter: vSphere 7 general support ended in October 2025, and vSphere 8 general support ends on 11 October 2027.
• A free ESXi edition returned in April 2025 (8.0 Update 3e), but it has no vCenter management and no support. It is fine for a lab, not for a production cluster.
• The desktop virtualization business, including Horizon, was sold off and now operates separately as Omnissa.

Licensing terms, including minimums per order, keep being adjusted. Don't rely on articles (this one included); always check against a real quote.

Many organisations report higher annual costs, especially smaller ones that used the Essentials kits. The real impact depends on your core count and which bundle you need.

To be fair: VMware is still an excellent platform with the broadest ecosystem. For some organisations, renewing is still the sensible choice.

The main alternatives

A few current details. Proxmox VE 9 (2025) is built on Debian 13 and includes a wizard that imports VMs straight from an ESXi host. XCP-ng 8.3 is the current long-term-support release. Hyper-V in Windows Server 2025 follows normal Windows Server licensing: Standard edition covers two Windows VMs per fully licensed host, Datacenter covers unlimited Windows VMs on that host.

The hypervisor isn't the whole story

Changing hypervisor touches much more than the hypervisor. Start with backup: Veeam Backup & Replication supports Hyper-V natively, Nutanix AHV through a plug-in, and Proxmox VE through a plug-in since version 12.2 (2024). Proxmox also has its own Proxmox Backup Server. If you use other backup software, check its supported-platform list first.

Then: will your existing SAN work over iSCSI/NFS with the new platform? How much of your scripting, templates and alerting needs rewriting? Some ERP and core banking vendors only certify certain platforms. And training your team is a real cost, not an extra.

Which option fits which situation

Evaluation and migration

1. Inventory: every VM, CPU cores, RAM, storage, critical apps, and when your current subscription or support ends.
2. 3–5 year TCO: licences, support, hardware, training and migration labour.
3. Proof of concept: a 2–3 node test cluster, move non-critical VMs, test HA, backup and restore.
4. Phased migration: low-risk VMs first, critical systems last, each step with a rollback plan.
5. Post-migration review: performance, backups, monitoring and documentation.

Technical steps people skip

• Drivers before the move. Windows VMs going to a KVM-based platform (Proxmox VE, Nutanix AHV) need VirtIO drivers installed before migration, or the VM can fail to find its boot disk.
• Guest tools after the move. Remove VMware Tools and install the new agent: QEMU guest agent on Proxmox, Nutanix Guest Tools on AHV. Hyper-V integration services are built into current Windows.
• New network adapters. Windows sees a new NIC and may drop its static IP. Record every VM's IP settings before you start.
• Licences tied to hardware. Older applications licensed to a MAC address or a USB dongle can stop working. Ask the software vendor first.
• Cluster quorum. HA needs a majority. Use three nodes, or for a two-node Proxmox cluster add a small QDevice on a third machine.
• Copy time. Over 1 GbE a 500 GB disk takes well over an hour; 10 GbE cuts that sharply. Measure the speed on the first pilot VM and plan maintenance windows from that number, not from a guess.

Example: 3 hosts, 40 VMs

• 3 hosts × 2 CPUs × 16 cores = 96 cores to pay for on any per-core platform
• Most VMs run Windows Server → Windows Server Datacenter licensing belongs in every option (unless you already own it)
• Option A: renew VMware. Least change
• Option B: Proxmox VE with Ceph or the existing SAN, plus a support subscription. Low licence cost, but the team needs training
• Option C: a Hyper-V cluster, since Windows Server Datacenter is already licensed
• Compare all three on 5-year TCO and risk before deciding

The Windows licensing line

Windows Server is licensed per physical core whatever hypervisor you run, with a minimum of 8 cores per CPU and 16 per server. With 32 cores per host, each host needs 32 core licences. Datacenter on all three hosts (96 cores) covers unlimited Windows VMs, and with HA, where any VM can restart on any host, that is normally the sensible route for 40 VMs. Standard would mean licensing the same cores again for every two VMs on every host. This cost is the same under VMware, Proxmox or Hyper-V, so put it in every column of the comparison or in none.

Where migrations go wrong

1. Deciding on licence price alone and forgetting labour and training.
2. Migrating before testing a restore on the new platform.
3. Waiting until two months before renewal, leaving no time to trial anything.
4. Putting critical systems on open source with no support contract.

What changes in Lao conditions

• Skills. In Vientiane it is easier to find engineers who know VMware and Windows than people with real Linux, Ceph or ZFS experience. If you choose Proxmox, budget for training and a support subscription, and agree who you call at night.
• Power. Brownouts reboot switches before servers. If the cluster network drops, HA may restart VMs on the wrong host or fence a healthy node. Put the switches that carry cluster and storage traffic on UPS, not only the servers. Ceph needs at least three nodes and a dependable 10 GbE network.
• Hardware and warranty. Parts usually come through Thailand. Check that the server warranty actually covers on-site service in Laos, and keep one spare disk and power supply per model.
• Certified applications. Banking, ERP and hospital systems often list supported hypervisors. Get that list in writing before the proof of concept, not after the migration.

If you want our opinion

Start evaluating 6–12 months before your subscription or support runs out. For small and mid-sized organisations with one or two people comfortable on Linux, Proxmox VE with a support subscription is worth a proper trial. If you want simplicity and turnkey HCI, look at Nutanix. If you're almost entirely Windows, Hyper-V may be the best value. Whichever you pick, make sure someone can physically reach your servers within a day.

Send us a VM export from vCenter (or just host, CPU and VM counts) along with your renewal date. We'll put the three options side by side on TCO and, if you want, set up a proof of concept for you to test first.