Broadcom VMware, what actually changed and what it costs
Broadcom turned VMware into a subscription only business anchored on VMware Cloud Foundation, priced per core with a sixteen core per CPU minimum, and opening quotes ran several times the old perpetual cost. With a core audit and a credible exit, most buyers settle far below the first number, which is the whole defense in one sentence.
Prepared by Redress Compliance · August 6, 2026 · Broadcom VMware advisory. Based on 35 to 50 renewals and migrations advised 2024 to 2026.
Executive summary
The model change is total. Perpetual licenses are gone, the product catalog collapsed into subscription bundles anchored on VMware Cloud Foundation, and the metric moved to per core with a sixteen core minimum per CPU socket.
An estate that bought vSphere licenses once and paid support now rents the full stack annually, whether or not it uses the full stack.
The repricing rode the repackaging. The bundle carries vSphere, vSAN, NSX, and the management layer as one subscription, and converting a support renewal into a full stack rental is how opening quotes arrived at several times the prior run rate.
The support reduction story, existing support commitments wound down to push the subscription conversion, supplies the pressure; the bundle supplies the price.
The defense is arithmetic plus an alternative.
A core audit corrects the base, the sixteen core minimums and small host phantom cores identified before the quote, and a credible exit evaluation corrects the rate: with both in hand, most buyers settle far below the first number.
In the 30 to 55 percent below list range the negotiated market actually clears at.
The renewal cycle is now contracted. Subscription terms mean the negotiation recurs on Broadcom's calendar, with the discount bands, ramp structures, and renewal caps set at each signature deciding the next cycle's starting point.
The estates that hold flat renewals treat each cycle as the next one's preparation, alternatives maintained, cores audited, and clauses banked.
What changed, item by item
| Before Broadcom | After | The consequence |
|---|---|---|
| Perpetual licenses plus annual support | Subscription only, term based | The estate rents what it owned, and the renewal recurs on contract |
| A la carte products | Bundles anchored on VCF, with vSphere Foundation below it | Paying for the stack, not the components in use |
| Per processor constructs | Per core, sixteen core minimum per socket | Small host estates carry 10 to 25 percent phantom cores |
| Channel breadth | A concentrated partner program | Fewer routes, and the direct relationship carrying more of the negotiation |
The support reduction story is the pressure mechanism. Winding down support on perpetual estates converts a maintenance decision into a platform decision on Broadcom's timeline: keep running unsupported, convert to subscription, or leave.
Naming the mechanism matters, because the answer to manufactured urgency is a prepared calendar, not a faster signature.
What it costs now, and where the number bends
The price structure underneath the quotes is consistent: VCF lists at $350 to $400 per core per year, term bands cut 18 to 38 percent, scale adds points above 10,000 cores, and competitive pressure moves 8 to 15 more, the full stack worked in the VCF pricing brief.
The changes piece of the story is that every one of those bands is applied to a base the sixteen core minimum and the bundle tier decisions inflate by default, which is why the deconstruction precedes the discount.
The per cluster question, VCF or the smaller vSphere Foundation, and the per host question, consolidate or carry phantom cores, together move more money than the headline discount in most estates. The per core calculator and the VCF versus VVF calculator run both in minutes.
The VMware negotiation playbook
The buyer side framework covering VCF pricing, the per core model, the support reduction defense, and the moves that hold a flat Broadcom renewal.
Get the white paper →Defending a flat renewal, the two instruments
The settlements far below opening quotes shared two instruments, and neither was rhetorical.
The core audit corrected the base: RVTools level inventory, socket and core counts verified before Broadcom rounds them, small hosts consolidated or scoped out, retiring clusters moved to the ramp discussion.
The exit evaluation corrected the rate: an alternative costed to executability, even for a workload slice, moving the final discount 8 to 15 points against estates that negotiated with nothing on the table.
The exit landscape is real enough to price: Hyper V for the Windows heavy estate, where Datacenter licensed estates saved 50 percent plus, and the wider field in the alternatives comparison guide.
The finding that travels: mixed outcomes are the norm, 20 to 40 percent of workloads stayed on VMware after evaluation, and the evaluation paid either way, because the quote moved whether or not the workloads did.
- 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 renewals and migrations, 2024 to 2026
Across roughly 35 to 50 Broadcom VMware renewals and migrations Fredrik Filipsson advised between 2024 and 2026, the change mechanics produced the same arc:
First quotes against the prior vSphere run rate, the bundle and metric conversion doing the inflating.
The discount shift from a documented alternative, even partial, against estates with nothing on the table.
The cycle finding matters most going forward: the subscription model makes this negotiation recurring, and the second renewal arrives with whatever clauses the first one banked.
Renewal caps, price holds on growth cores, and portability language written at this signature are the discount of the next one, which is the discipline the perpetual era never demanded and the subscription era punishes estates for lacking.
Your first five moves
- Audit the cores before the quote: sockets, minimums, phantom cores, and consolidation candidates, so the base is yours.
- Scope the bundle per cluster, VCF where the stack is used, vSphere Foundation where it is not, and refuse the estate wide default.
- Price an exit to executability, even for a slice; it moves 8 to 15 points whether or not anything migrates.
- Answer the support reduction pressure with a calendar, not a signature: the urgency is manufactured, the preparation window is yours.
- Bank the clauses for the next cycle: renewal caps, growth price holds, and portability, because the subscription negotiation now recurs. The Broadcom VMware practice runs the sequence with you.
Frequently asked questions
What did Broadcom change about VMware licensing?
Everything structural: perpetual licenses ended, the catalog collapsed into subscription bundles anchored on VMware Cloud Foundation, and the metric moved to per core with a sixteen core minimum per CPU socket.
An estate that owned vSphere and paid support now rents the full stack on term, with the negotiation recurring at each renewal.
Why is our Broadcom quote so much higher than what we paid VMware?
The bundle and the metric conversion: a support renewal became a full stack subscription, priced per core with minimums, which produced opening quotes at 2 to 4 times prior run rates across our engagements.
The openings are positions: with a core audit and a credible exit, settlements landed 30 to 55 percent below list.
What is the support reduction story?
The pressure mechanism: support on perpetual estates wound down to force the platform decision, keep running unsupported, convert to subscription, or leave, on Broadcom's timeline.
The counter is a prepared calendar and a priced alternative, because manufactured urgency only works on estates without either.
Do we need full VCF or is vSphere Foundation enough?
Per cluster, not per estate: VCF bundles vSphere, vSAN, NSX, and management, and clusters using only compute virtualization often fit the smaller bundle at a materially lower rate.
Scoping the tier honestly per cluster shrinks the base before any discount applies, which is usually worth more than the discount.
Is leaving VMware realistic?
Partially, and that is enough: mixed outcomes are the norm, with 20 to 40 percent of workloads typically staying after evaluation, Datacenter licensed estates saving 50 percent plus on what moved to Hyper V.
And the documented evaluation moving Broadcom quotes 8 to 15 points whether or not anything migrated.
The evaluation pays either way.
How do we keep the second Broadcom renewal from repeating the first?
Bank the clauses now: renewal caps, price holds on growth cores, ramp structures matched to real plans, and portability language.
The subscription model makes the negotiation recurring, and each cycle starts from whatever the previous signature protected, which is the discipline the perpetual era never required.
