VMware VVF vs VCF in 2026: What You Get, What to Do

The question we actually get

"Our VMware renewal quote came back as VVF, and the partner also offered VCF. What is the difference, do we need the bigger one, and why is the core count on the quote higher than the cores we actually have?"

We usually hear it from an IT manager with two or three hosts and a renewal date that is too close. This guide explains the two VMware by Broadcom bundles as they stand in late 2026, what each one contains, how they compare with the main alternatives, and what we would do in a few common Lao situations. It does not quote prices: those depend on your core count, term and sales channel, and they change. For the wider question of leaving VMware, read our guide after Broadcom bought VMware.

How VMware is sold now

Since Broadcom took over VMware at the end of 2023, the product line has been cut down to a few bundles sold only as subscriptions. Perpetual purchases are gone. The two names you will see on almost every quote are VMware vSphere Foundation (VVF) and VMware Cloud Foundation (VCF).

Three rules drive the number on the quote:

• You pay per physical CPU core, and every core in the host counts, whether VMs use it or not.
• Each CPU is counted as at least 16 cores. A CPU with 8 or 12 cores is billed as 16.
• The subscription runs for a fixed term. When it ends, you lose access to patches, upgrades and support.

Broadcom has also set minimum order quantities at times and has adjusted bundle contents more than once since 2024. Treat everything below as the shape of the offer, and check the current Broadcom terms against a written quote before you decide.

What is inside VVF and VCF

VVF is the compute bundle. It gives you vSphere (ESXi hosts with the full feature set: HA, vMotion, DRS, distributed switches), vCenter to manage the cluster, an allowance of vSAN capacity per subscribed core, operations tooling for monitoring and capacity planning (formerly Aria Operations, now under the VCF Operations name), and a way to run Kubernetes on vSphere (the feature formerly branded Tanzu). For most Lao companies running Windows and Linux VMs on a small cluster, VVF covers the technical need.

VCF is the full private-cloud stack. On top of what VVF has, it adds NSX for software-defined networking (overlay networks, routing and load balancing inside the cluster), a much larger vSAN allowance per core, lifecycle tooling that deploys and patches whole environments together, and a wider set of automation and operations features. It is built for organisations that run many clusters or segment their network in software rather than on physical firewalls.

Some features are sold as add-ons on top of either bundle, for example vSAN capacity beyond the allowance and some advanced security features. The vSAN allowance per core has already been changed once, so check the current figure for each bundle instead of trusting an old datasheet.

VVF, VCF and the alternatives side by side

Our server virtualization and hyperconverged HCI pages explain how we build each option.

Count your cores before anything else

The bill follows physical cores, so work from the hardware, not from the VM list. Two examples:

• Three hosts, each with two 8-core CPUs. Physical cores: 3 × 2 × 8 = 48. Billed cores: 3 × 2 × 16 = 96, because each CPU is raised to the 16-core minimum. Half of what you pay for does not exist.
• Two hosts, each with one 32-core CPU. Physical and billed cores: 2 × 1 × 32 = 64.

The second cluster has more real cores than the first (64 against 48) and is billed for fewer (64 against 96). This is why older dual-socket servers with low core counts became expensive to keep on VMware, and why a hardware refresh and a renewal should be planned together. Before you consolidate, check that the smaller cluster still carries every VM with one host down: with two hosts, each one must be able to run the full load alone.

Which one we would pick for…

A 2–3 host SME cluster. VCF is almost never the right answer here. If you stay on VMware, VVF covers HA, vMotion and central management, and you will not use NSX on three hosts. Ask whether a smaller vSphere edition is still offered through your channel; availability has shifted since 2024. If the quote still looks heavy, this is exactly the size where Proxmox VE with a support subscription, or Hyper-V when you already own Windows Server Datacenter, deserves a proof of concept.

A branch office. One or two small hosts in a provincial branch rarely justify their own subscription. What works: run branch VMs from the head-office cluster over a dependable link, use a single small host with good local backup, or a two-node cluster with a witness. Whatever you choose, branches fail more often from power than from software: brownouts around 17:00–19:00 and rainy-season lightning. Put the host and its switch on a UPS sized for a clean shutdown.

Data that has to stay in the country. Some businesses keep key data on servers in Laos for engineering reasons. The internet link to a foreign cloud is the weakest part of the chain, latency to the region adds delay to every database call, and restoring a large dataset over a thin link takes days. VVF, VCF and every alternative here run fully on your own hardware, so the bundle is not the deciding factor. Restore speed and local skills are. Keep a second copy at a separate site in Laos and test restores; our backup and disaster recovery page covers the design.

A multi-site estate already on NSX. If you already segment your network with NSX and your team knows it, VCF is the natural renewal. Leaving means rebuilding that network design elsewhere, so budget real time for it.

A mine or remote site on a generator. Keep it simple: a small cluster with local storage, a hypervisor your on-site staff can restart, and spare disks and power supplies on site. A team that can recover a host at 2 a.m. matters more than advanced features.

What to do at renewal time

1. Inventory cores, not VMs. For every host, write down CPU model, sockets, cores per socket and the billed figure after the 16-core minimum. Record raw vSAN capacity too if you use it.
2. List what you actually use: HA, vMotion, DRS, vSAN, NSX, Kubernetes. If NSX is not on the list, you probably do not need VCF.
3. Get the quote in writing, by bundle and term, and ask how the next renewal will be calculated.
4. Test a migration on one host. Take one host out of the cluster (or use a spare server), install the alternative and move two or three non-critical VMs, including one Windows VM. Test backup and a full restore, not only the move. Our backup restore testing guide lists what to check.
5. Decide with numbers: renewal cost, migration labour, training and risk for each option over three to five years.

Start six to twelve months before the subscription ends. With two months left, renew short and test.

Mistakes we see on Lao sites

• Counting VMs or vCPUs instead of physical cores, so the quote never matches the plan.
• Assuming VCF is simply "better" and buying it for three hosts.
• Forgetting Windows Server. It is billed per core on every hypervisor, so it belongs in every column of the comparison or in none.
• Moving to a new platform without checking that the backup software supports it.

Next step

Send us on WhatsApp a screenshot of your vCenter host list (CPU model, sockets, cores per socket) and the page of your renewal quote that shows the bundle and term. We will count your billed cores, tell you whether VVF or VCF matches what you actually use and, if it makes sense, plan a one-host migration test.