The VMware Cloud Foundation Licensing Brief For Your 2026 Broadcom Renewal

VMware Cloud Foundation lists at 350 dollars per core per year, bills a 16 core per CPU floor, and cannot be ordered below 72 cores. A workload mapping moves 30 to 60 percent of most estates off the full bundle.

Prepared by Redress Compliance · June 2026 · Representative VMware estate scenario (benchmark scenario, not a quote)

Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025. List prices from Broadcom and VMware published materials, retrieved June 2026.

VCF is priced per core on an annual subscription, with a 16 core per CPU floor and a 72 core minimum order quantity. The whole quote is built on those three mechanics, and each one inflates the count above what the deployment actually runs.

Broadcom documents the product on the Broadcom VMware Cloud Foundation page and the feature scope in the VCF 9.1 comparison document. The pricing levers below are where the proposal moves.

The first non obvious mechanic is the 72 core minimum order. Effective April 10, 2025 Broadcom raised the smallest orderable quantity from 16 cores to 72 cores per subscription. A two node edge site running 48 cores now buys 72, a structural overbuy the buyer cannot remove without consolidating sites.

How the per core list compares to vSphere Foundation

Normalize both SKUs to a single 64 core host, two 32 core CPUs, and the spread is wide. The chart prices one host on each SKU at list and at a realized 30 percent off. Every figure matches the section 3 table.

VCF prices the full software defined data center as one per core rate. The buyer licenses vSphere, vSAN, NSX, and the Aria and VCF management plane together, whether or not the deployment uses every component. The audit question is simple: what inside the bundle does this estate actually run.

The second non obvious mechanic lives in the vSAN entitlement. VCF includes roughly 1 TiB of vSAN capacity per core, while vSphere Foundation includes only 0.25 TiB, and capacity above the bundled allowance is a separate add on meter. The bundle is not flat: it carries a capacity ceiling that triggers a second charge.

• NSX usage: confirm which clusters actually run NSX networking and security, not which clusters could.
• vSAN capacity: measure consumed TiB per core against the bundled allowance to find the second meter exposure.
• Management plane: identify whether the full Aria suite is operational or whether the operations subset would cover the estate.

Map every workload to the lightest SKU that covers it. VCF carries the full stack, but most estates run a large share of standard virtualization that vSphere Foundation covers at 135 dollars per core rather than 350. The mapping is the single most valuable hour in the renewal.

How to run the workload mapping

1. Inventory clusters: capture the deployed core count and the components each cluster actually uses.
2. Match the SKU: assign VCF only where NSX, full vSAN, or the full management plane is in production use.
3. Quantify the gap: total the cores a VCF default over licenses against a mapped position.

On a representative estate, a regional health insurer running 160 dual socket hosts and 8,000 billed cores on VCF, the mapping moved 35 percent of the estate to vSphere Foundation without losing any capability. The table prices three renewal paths. Every figure is internally consistent and labeled as a benchmark scenario, not a quote.

The arithmetic checks. Default is 8,000 cores at 308 dollars (350 less 12 percent); leverage is 8,000 at 245 (350 less 30 percent); mapped is 5,200 at 245 plus 2,800 at 95 (135 less roughly 30 percent). The mapped position saves 924,000 dollars a year against the default.

Where the common advice on VMware Cloud Foundation licensing is wrong

The standard advice since the Broadcom acquisition is that VCF is the only viable path. We disagree. Across roughly 25 to 40 VMware situations Redress Compliance benchmarked in 2024 to 2025, a workload mapping moved 30 to 60 percent of the estate to vSphere Foundation or substitution with no loss of capability.

Accepting the VCF default is simply how the unused stack gets paid for. The buyer side move is to map the workloads, document the deployed core profile, and price the conversion line by line. A mapped estate beats a default VCF quote by a wide margin every time.

Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025.

The 16 core per CPU floor bills every processor at a minimum of 16 cores, regardless of how many cores the CPU physically carries. On modern dense CPUs the floor is harmless. On older small CPUs it is the single most expensive line in the quote.

Take a host with two 10 core CPUs. It deploys 20 cores but bills 32, two CPUs at the 16 core floor, a 60 percent inflation on that host alone. At 350 dollars per core the 12 phantom cores cost 4,200 dollars a year for capacity that does not exist.

• Document the profile: capture cores per CPU for every host so the floor exposure is visible, not buried in a total.
• Consolidate small CPUs: retiring sub 16 core hosts onto dense CPUs removes the floor tax at source.
• Mind the 72 core order floor: small remote sites cannot buy below 72 cores, so consolidation also fixes the minimum order overbuy.

Break the conversion into its components and price each line. Broadcom presents the move from perpetual licenses to the VCF subscription as one number, which hides where the exposure sits and removes the buyer leverage that lives in the detail.

The third non obvious mechanic is the support entitlement transfer. Unexpired perpetual support and subscription carries no automatic credit into the new subscription rate. The conversion can quietly charge for support the customer has already paid through the end of the perpetual term.

A VCF quote priced on the full stack, a 16 core floor, and a single conversion number is an opening position, not the licensing your workloads require.

Six clauses carry the value on a VCF renewal. Each is a contract position, so it must be documented and placed before the renewal date closes the window. We have negotiated every one of these into live Broadcom enterprise contracts.

The fourth non obvious mechanic is the term price hold. Without an explicit renewal cap, the locked per core rate applies only inside the current term, and at renewal it resets to then current list, which has risen every year since the acquisition. The hold is what makes a multi year commit safe.

Align VCF, vSphere Foundation, and any substitution path into one plan that runs across renewal cycles. A portfolio view keeps the Broadcom commitment matched to the workloads the estate actually runs, term over term, rather than locking the whole estate to the full stack in one signature.

Track the SKU scope on the VMware Cloud Foundation page and plan the estate against it before every renewal. The portfolio position is what keeps each renewal a decision rather than a default.