The estates that followed the old best practice are penalised hardest by the new metric
Per core subscription with a per processor minimum replaced perpetual licensing, and that single change made your data centre layout a licensing decision retroactively. The estates built horizontally on many smaller hosts, which was sound engineering under the old model, are precisely the estates paying for cores their processors do not physically have.
Prepared by Redress Compliance · August 10, 2026 · Broadcom advisory. Based on 30 to 40 Broadcom VMware renewals benchmarked, 2024 to 2025.
Executive summary
Annual cost rose 2 to 5 times for estates that had only licensed the hypervisor.
Three changes stacked at once: the bundle widened scope so buyers who wanted compute now pay for storage, networking, and management; subscription replaced a one time purchase plus support; and the per processor core minimum inflated the billable count.
The steepest increases landed on the narrowest previous footprints, which is the opposite of how a volume based increase usually distributes.
Core minimums added 10 to 30 percent of phantom cores, and the penalty is structural rather than accidental. A host with two low core processors bills at the floor rather than at its real core count, which means an estate built on many small hosts pays for capacity that does not physically exist.
Horizontal build outs on modest hardware were sound engineering under a per processor perpetual model, so the estates being penalised most are the ones that followed the previous best practice most closely.
In 22 of the 35 renewals we reviewed, the discount was a distraction from the billable core count itself.
That is the core analytical point and it is worth stating as arithmetic: cutting billable cores by 20 percent beats a 20 percent discount, because the reduced core count persists through every year of the subscription and carries into the next renewal baseline, while a discount is renegotiated from scratch each cycle.
One is a permanent change to the quantity, the other is a temporary change to the rate.
Consolidation onto denser hosts recovered 15 to 25 percent of the increase before any discount conversation.
Read alongside our exit economics work, where a costed alternative quote cut the renewal by an average of 22 percent whether or not the estate migrated, the levers stack in a specific order: consolidate first so the base is right, then bring a costed exit quote against the corrected base.
Then negotiate rate.
Each step operates on the output of the one before it, which is why running them in the reverse order loses most of the value.
The old model against the new one
| Dimension | Pre acquisition | VCF today | Buyer impact |
|---|---|---|---|
| Licence type | Perpetual plus support | Subscription only | Recurring cost, no ownership |
| Unit | Per processor or per core | Per core with a minimum | Higher billable core count |
| Scope | Pick the products you need | Bundled platform | Pay for unused capability |
| Renewal | A support renewal | A full subscription renewal | Larger annual exposure |
The retroactive part is what makes this unusual, and it is worth naming explicitly.
Under a per processor perpetual model, building horizontally on many modest hosts was the sensible answer: it spread failure domains, kept per host blast radius small, and cost nothing extra in licensing because the unit was the socket. The core minimum inverts that.
The same layout now bills for cores the processors do not have, so an architecture chosen for resilience is charged as though it were chosen for capacity.
Nobody made a bad decision; the metric changed underneath a decision that was already made, and the remedy is a hardware refresh cycle rather than a negotiation. The bundle comparison sits in the VVF against VCF guide.
The levers, and why the order changes the total
- Consolidate first. Fewer, denser hosts cut the billable core count and recovered 15 to 25 percent of the increase before anybody discussed price, and the reduction persists across every year of the subscription.
- Decommission idle hosts and reclaim stranded capacity, which is the same lever with no engineering risk attached, because a host that runs nothing still bills at the floor.
- Challenge the bundle scope, since buyers who needed only compute are paying for storage, networking, and management, and the scope question is separable from the core count question.
- Bring a costed exit quote against the corrected base, which cut renewals by an average of 22 percent in our exit work whether or not the estate actually migrated.
- Negotiate term, ramp, and price protection last, because a rate applied to a corrected base compounds correctly while a rate applied to an inflated one simply discounts the phantom cores. The exit arithmetic sits in the migration economics analysis.
The Broadcom VMware negotiation brief
The bundle comparison, the per core arithmetic, the minimum core rule, and the renewal moves that hold against an opening position.
Get the white paper →Why the base beats the rate, in numbers
The reseller advice is to accept the bundle and negotiate a discount on the new list, and in 22 of the 35 renewals we reviewed that advice pointed the buyer at the smaller of the two available levers. The reason is structural rather than tactical.
A discount is a percentage applied to a quantity for a term, renegotiated from zero at the next renewal, and it decays as list prices move underneath it.
A core reduction is a permanent change to the quantity itself: it applies in year one, applies again in years two and three of the same subscription, and carries into the renewal baseline as the number the next negotiation starts from.
Twenty percent off an inflated base and twenty percent off the base itself look identical on the first invoice and diverge on every one after it.
That is before the compounding effect of the two working together, since a discount negotiated against a corrected base is worth more in absolute terms than the same percentage against a padded one. Sequencing matters for the same reason at the exit stage.
Our migration work found a costed alternative quote cut the renewal by an average of 22 percent regardless of whether the estate moved, but that percentage is applied to whatever base exists when the quote is tabled.
So consolidating first and quoting second is worth materially more than doing it the other way round.
Three steps, each operating on the output of the last: correct the base, price the alternative against it, then argue the rate. The full alternatives landscape sits in the alternatives comparison.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across Broadcom VMware renewals, 2024 to 2025
Across roughly 30 to 40 Broadcom VMware renewals benchmarked between 2024 and 2025, the move to subscription raised annual cost for most estates, with the steepest jumps where customers had licensed only the hypervisor:
Billable cores added by the per processor minimum on hosts whose processors do not physically carry that many cores.
Renewals where the discount conversation displaced the billable core count as the thing that actually decided the invoice.
Three patterns recurred: annual cost rising 2 to 5 times for hypervisor only estates moving to the full bundle, core minimums adding 10 to 30 percent of phantom cores on hosts with low core processors.
And consolidation onto denser hosts recovering 15 to 25 percent of the increase before any discount.
The buyer side move is to redesign the estate first so the minimums stop charging for cores that do not exist, then price a credible alternative against the corrected base, and only then negotiate rate, term, and price protection.
Under this metric your hardware layout is a licensing decision rather than only an engineering one.
Your first five moves
- Count true physical cores per host and identify where the minimum is charging for cores that do not exist, which added 10 to 30 percent to the billable base in our file.
- Consolidate onto fewer, denser hosts and decommission idle ones, since that recovered 15 to 25 percent before any discount and the reduction persists across every subscription year.
- Challenge the bundle scope separately from the core count, because paying for storage, networking, and management you do not deploy is a different argument from paying for phantom cores.
- Bring a costed exit quote against the corrected base, not the original one, as the average 22 percent renewal cut applies to whatever number is on the table when it is produced.
- Negotiate rate, term, and price protection last, because a percentage against a corrected base compounds and the same percentage against a padded base just discounts the padding. The Broadcom practice runs the core audit with you.
Frequently asked questions
What changed with VMware Cloud Foundation?
The portfolio collapsed into bundles with VCF as the flagship, packaging compute, storage, networking, and management into one subscription sold per core.
Perpetual licensing ended for new purchases, so customers who needed only the hypervisor now pay for capability they may never deploy, which is the core of the increase.
How do core minimums catch buyers out?
A host with two low core processors bills at the per processor floor rather than at its real core count, so an estate built on many small hosts pays for capacity that does not physically exist.
In our file that added 10 to 30 percent of phantom cores, and it falls hardest on estates that built horizontally for resilience under the old model.
Why did costs rise so much for hypervisor only estates?
Three changes stacked. The bundle widened scope so compute buyers now pay for storage, networking, and management. Subscription replaced a one time purchase plus support, converting capital into recurring cost. And the core minimum inflated the billable count.
Annual cost rose 2 to 5 times where the previous footprint was narrowest.
Is negotiating a discount the right first move?
No, and in 22 of the 35 renewals we reviewed it displaced the lever that actually mattered.
Cutting billable cores by 20 percent beats a 20 percent discount, because the reduced count persists through every year of the subscription and carries into the next renewal baseline, while a discount is renegotiated from zero each cycle.
How much does consolidation recover?
Between 15 and 25 percent of the increase before any discount conversation, by packing workloads onto fewer, denser hosts so the per processor minimum stops charging for cores that are not there.
Decommissioning idle hosts is the same lever with no engineering risk, since a host running nothing still bills at the floor.
When is an exit worth modelling?
Once the base is corrected.
A costed alternative quote cut renewals by an average of 22 percent in our migration work whether or not the estate actually moved, but that percentage applies to whatever base is on the table when the quote is produced.
So consolidating first and quoting second is worth materially more than the reverse order.
Is hardware layout really a licensing decision now?
Yes, and retroactively. Under a per processor perpetual model, building horizontally on modest hosts was sound engineering that cost nothing extra in licensing. The core minimum inverts that, charging an architecture chosen for resilience as though it were chosen for capacity.
The remedy runs on a hardware refresh cycle rather than through negotiation.
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.
