Estates licensed the cores they owned rather than the cores they ran, and the gap was the overspend
VMware cost in 2026 is the SKU multiplied by the licensed core count, and buyers negotiate the rate because the rate is the number on the quote. Broadcom sets the rate. You set the SKU and the core count, and those two move the bill much further.
Prepared by Redress Compliance · August 16, 2026 · Broadcom VMware advisory. 30 to 40 estates benchmarked, 2024 to 2025.
Executive summary
SKU choice, not the per core rate, was the largest controllable driver. In 19 of the 30 plus estates benchmarked, the biggest cost came from landing in the premium VCF bundle when vSphere Foundation covered the actual workload.
Defaulting into VCF cost 20 to 45 percent over the bundle the workload needed, because VCF folds in vSAN, NSX, and Aria whether or not you deploy them.
The 16 core per CPU floor added 10 to 25 percent on low density hardware, so two estates running identical workloads pay very differently based purely on core density.
First renewals off perpetual multiplied annual cost 2 to 3 times, because a one time license plus maintenance became a recurring per core subscription. That is a structural change, not only a price rise.
Three SKUs, and what each one actually covers
Broadcom consolidated VMware into a short list. Each is per core, so your bill is the per core rate multiplied by licensed cores, and the SKU choice and the core count together set it.
| SKU | Scope | Best for | Where the overspend starts |
|---|---|---|---|
| VMware Cloud Foundation | vSphere, vSAN, NSX, Aria | Full private cloud estates | Paying for components you never deploy |
| vSphere Foundation | Compute plus core management | Virtualization focused estates | Under scoping where vSAN is genuinely in use |
| vSphere Standard | Basic virtualization | Small, simple estates | Feature caps discovered after signature |
| Add ons | Extra vSAN capacity, advanced services | Storage heavy estates | Modeled late, can rival the base subscription |
Match the SKU to the components you actually run. Estates that only virtualize compute rarely need the full VCF bundle, and defaulting into it is the most common 2026 overspend we see, worth 20 to 45 percent against the bundle the workload needed. The bundle question is decided before the quote arrives, by an inventory of what is deployed, not by a discount conversation afterwards. Extra vSAN capacity beyond the included entitlement and advanced services bill on top of the base SKU, so model those in the same sheet.
Owned cores, running cores, and the floor between them
- Count physical cores per CPU and apply the 16 core floor. Every current SKU licenses physical cores with a 16 core per CPU minimum, so a CPU below that count carries a permanent penalty on every renewal.
- License running cores, not owned cores. This is the gap that compounds every other error, because an inventory of what you own is not an inventory of what is carrying workload.
- Consolidate onto dense hosts. Denser hardware means fewer total licensed cores and a lower bill, and it is the lever most buyers underuse because it sits with infrastructure rather than procurement.
- Refresh low density hardware before the renewal, not after. CPUs below 16 cores add 10 to 25 percent through the floor alone, and the refresh business case is strongest when the subscription quote is still open.
- Compare total subscription cost against prior license plus support, not against list price. The 2 to 3 times jump on first renewals is a change of structure from perpetual to recurring, and comparing to list hides that.
- Use multi year terms to hold the per core rate where the estate is stable, and pair the term with a cap so the rate you hold does not reset in year three.
The VMware alternatives brief
The exit architecture, the tier by tier migration math, and the costed alternative that resets a Broadcom renewal.
Get the brief →The rate is Broadcom's. The SKU and the core count are yours.
The standard line from resellers is that the 2026 increase is unavoidable and the only remaining choice is which VCF term to sign. We disagree, and the benchmarking says something specific about where the money actually moves. In roughly 19 of the 30 plus estates reviewed, the largest cost driver was not the per core rate at all. It was landing in the premium VCF bundle when vSphere Foundation covered the actual workload, compounded by licensing the cores an organization owned rather than the cores it was running.
Those two errors compound rather than add. A wrong bundle applies a premium rate to every licensed core, and a wrong core count inflates the number of cores that premium applies to. An estate that defaults into VCF and licenses its full owned inventory pays the higher rate on hardware that is idle, decommissioned, or carrying nothing that needs the bundle. Neither error appears anywhere on the quote, because the quote is denominated in a rate, and a rate is the one variable in the equation that the buyer cannot change.
The 16 core per CPU floor deserves separate attention because it punishes a decision most estates made for good reasons years ago. Buying more, smaller hosts was sound practice under socket based licensing and under perpetual terms. Under a per core subscription with a floor, every CPU below sixteen cores pays for capacity that does not exist, permanently, on every renewal. Two estates running identical workloads can therefore pay very differently on core density alone, which makes consolidation a licensing decision rather than only an infrastructure one.
The practical sequence is unglamorous and it happens before the quote. Inventory every host by CPU and physical cores per CPU, apply the floor, map which VCF components you genuinely deploy, choose the lowest SKU that covers the real workload, and model vSAN overage and advanced services in the same sheet. Then benchmark, and bring a costed alternative you would actually act on. The rate is set by Broadcom. Everything the rate multiplies is set by you. The bundle comparison sits in VCF against vSphere Foundation, the per core arithmetic in VCF pricing 2026, and the wider library in the VMware practice.
- RVTools in, core sizing out: three pricing scenarios across discount bands
- Your quote benchmarked against real closed Broadcom deals, not survey data
- Exit and alternative scenarios priced so you negotiate with a credible walkaway
What the cost engagements showed, 2024 to 2025
Across roughly 30 to 40 Broadcom VMware estates benchmarked, landing in the wrong SKU was the most expensive single error:
Overpay by estates that defaulted into VCF when vSphere Foundation covered their deployed components.
What the 16 core per CPU minimum adds on low density hardware, permanently and on every renewal.
The third pattern is the structural one. First renewals off perpetual multiplied annual VMware cost by 2 to 3 times, because a one time license plus maintenance converted into a recurring per core subscription. Compare the total subscription cost against your prior license plus support, not against list, or the comparison flatters the quote.
Perpetual licences can continue to run, but new purchases and renewals are subscription only, which is precisely why the renewal is negotiable: the vendor needs it to land. That leverage is priced in the Broadcom VMware pillar.
Watch the briefing · 4:44The VMware VCF Renewal: How to Prepare Before Broadcom Names the PriceThe core inventory and bundle mapping that has to happen before the quote arrives.
Your first five moves
- List every VMware host by CPU and physical cores per CPU, then apply the 16 core per CPU floor to compute the licensed core count.
- Separate owned cores from running cores and license only what carries workload, retiring or consolidating the rest before the quote is drawn.
- Map which VCF components you actually deploy and choose the lowest SKU that covers the real workload, rather than the bundle that covers every possibility.
- Model vSAN overage and advanced services in the same sheet as the base SKU, because on storage heavy estates they can rival the base subscription.
- Benchmark the renewal and prepare a costed migration option you would genuinely act on. The Broadcom practice runs the core and SKU work with you.
Frequently asked questions
What are the main VMware SKUs in 2026?
Broadcom consolidated VMware into VMware Cloud Foundation, vSphere Foundation, and vSphere Standard. VCF is the full private cloud bundle including vSAN, NSX, and Aria. vSphere Foundation is compute focused, and vSphere Standard suits small, simple estates with capped features.
How is VMware licensed now?
All current SKUs license per physical core on subscription, with a 16 core per CPU minimum, on annual or multi year terms. Your cost is the per core rate multiplied by total licensed cores across hosts, which is why the core count matters as much as the rate.
Why did our VMware bill jump in 2026?
Broadcom retired perpetual licensing, converting a one time license plus maintenance into a recurring per core subscription. That structural change, often combined with defaulting into the VCF bundle, is why first renewals commonly multiplied cost two to three times.
What does licensing running cores rather than owned cores mean?
It means licensing the cores that actually carry workload rather than every core in your hardware inventory. The two diverge through idle hosts, decommissioned kit, and capacity held for growth, and the gap is where a wrong core count multiplies a wrong bundle rate.
Do we need VMware Cloud Foundation?
Only if you run its components. VCF bundles vSAN, NSX, and Aria. Estates that only virtualize compute usually fit vSphere Foundation, and defaulting into VCF is the most common 2026 overspend we see, worth 20 to 45 percent against the bundle the workload needed.
How does core density affect VMware cost?
Heavily, because of the 16 core per CPU floor. Any CPU below sixteen cores pays for capacity that does not exist, permanently and on every renewal. Two estates running identical workloads can pay very differently on density alone, so consolidation is a licensing decision as much as an infrastructure one.
What add ons increase the bill?
Extra vSAN capacity beyond the included entitlement, and advanced services, both billing on top of the base SKU. Model these before signing, because on storage heavy estates they can rival the base subscription rather than sitting as a rounding line.
What leverage is left on the 2026 renewal?
Right size the SKU to deployed components, license running cores on consolidated dense hosts, commit multi year with a cap to hold the rate, and bring a tested migration plan. The rate belongs to Broadcom, but the SKU and the core count belong to you.
The VMware Estate After the Repackaging
Part 2 of the Negotiating Broadcom series. Two bundles, per core with a sixteen core floor, three year terms paid up front, and a support horizon in October 2027 that decides your timing more than your renewal date does. What the estate actually looks like now, and which numbers are real.
