Unprepared renewals landed at 2x to 4x prior spend while prepared estates held the increase to 1.2x to 1.8x, which is the widest preparation spread we track for any vendor
The quote prices the estate the vendor can see: the historical entitlement at the richest plausible bundle. Every correction you bring narrows it.
Prepared by Redress Compliance · August 19, 2026 · Broadcom VMware renewals. 25 to 35 renewals advised, 2024 to 2025.
Executive summary
Unprepared renewals landed at 2x to 4x prior spend, while prepared estates held increases to 1.2x to 1.8x. The spread between those outcomes is the value of preparation.
Roughly 2 in 3 estates were quoted the full stack bundle when their deployment footprint justified the smaller one or less. The quote assumes the richest plausible packaging.
Core count corrections and host consolidation recovered 15 to 30 percent before any discount was discussed. Those corrections are entirely within buyer control.
Small host estates add 15 to 40 percent on the core minimum alone. The per socket floor punishes edge and branch heavy footprints hardest.
What exactly changed in the licensing?
Perpetual licenses and standalone products were replaced with per core subscriptions concentrated into two bundles: the full stack and core virtualization. The standalone line items folded into them, a consolidation laid out on the portfolio page.
Three shifts that set the new bill
- Licensing: subscription only, billed per physical core, with a sixteen core floor per socket.
- Packaging: two main bundles replace the standalone portfolio.
- Support: perpetual support renewals end at expiry, which forces the conversion moment.
What did not change
The technology. Existing deployments run undisturbed until the support clock forces the commercial conversion, which means the deadline is contractual rather than technical. The bundle scopes are published for the full stack and the core virtualization tier.
How much more are estates actually paying?
It depends almost entirely on preparation. The spread between a prepared and a reactive renewal is the widest we track for any vendor.
| Profile | Typical outcome | Main driver | What closed the gap |
|---|---|---|---|
| Prepared, counts corrected | 1.2x to 1.8x | Defensible cores, right bundle | Inventory evidence brought to the table |
| Partial preparation | 1.8x to 2.5x | Bundle fit challenged late | Component justification demanded after the quote |
| Reactive renewal | 2x to 4x | Quoted count and full bundle accepted | Nothing; the vendor's numbers stood |
| Small host estates | Add 15 to 40 percent | The sixteen core minimum penalty | Host consolidation before the count was fixed |
Why the spread is so wide
Because the quote prices the estate the vendor can see: the historical entitlement at the richest plausible bundle. Every correction you bring narrows the number, and the corrections are entirely within buyer control.
The Broadcom VMware renewal survival guide
The core count rebuild, the bundle challenge, the migration signal and the buyer side moves before the conversion date.
Get the brief →What 25 to 35 Broadcom VMware renewals showed
Across roughly 25 to 35 Broadcom VMware renewals Fredrik Filipsson advised between 2024 and 2025, the increase was real but the worst of it was avoidable. Three patterns recur.
- Unprepared renewals landed at 2x to 4x prior spend, while prepared estates held increases to 1.2x to 1.8x.
- Roughly 2 in 3 estates were quoted the full stack bundle when their deployment footprint justified the smaller one or less.
- Core count corrections and host consolidation recovered 15 to 30 percent before any discount was discussed.
The increase is not the negotiation. The count and the bundle are, and both are settled before a discount is ever mentioned.
- Inventory in, core sizing out: three pricing scenarios across discount bands
- Your quote benchmarked against real closed transactions
- Exit and alternative scenarios priced so you negotiate with a credible walkaway
Which estates absorb the worst of it?
Estates with many small hosts, heavy standalone product history, and no migration credibility. Each of the three inflates a different line.
Four profiles that pay most
- Edge and branch heavy: dozens of low core hosts each paying the per socket floor.
- Core virtualization only estates quoted the full stack: paying for components they never deploy.
- Regulated and change averse: no credible exit means no pricing counterweight.
- Mid market without direct coverage: less negotiation room in the narrowed channel.
Who came out relatively fine
Dense, consolidated data centre estates already using much of the full stack. Their per core economics were tolerable and the bundle matched the deployment, which is precisely the profile the model was built around.
Watch the briefing · 4:49Negotiating Broadcom: The VMware EstateRebuilding the count, contesting the bundle, and the evidence that has to be on the table before the rate is discussed.
How should buyers respond before the renewal?
In four moves, run in order. Rate is the last conversation, not the first.
The sequence that produced the 1.2x to 1.8x outcomes
- Inventory hosts, sockets and cores, removing dead and migration scheduled hosts with evidence.
- Map deployed components against both bundles and demand component level justification.
- Consolidate low core hosts where the refresh math works, before the count is fixed.
- Stand up a migration assessment on the portable workload tier, visibly.
Does the migration threat have to be real?
The assessment has to be real; the full migration does not. Vendors verify production signals, and a funded assessment with a named platform and a workload list is the minimum credible signal. The diversification case sits in the virtualization diversification playbook.
Pull the entitlement records early
They live in the support portal and reconstructing them late is what forces a reactive renewal. The core mechanics sit in the core licensing reference.
Where the common advice on the change is wrong
The standard advice is that the increase is inevitable, so buyers should sign the longest term available to delay the next one. We disagree.
Term without rate protection buys the wrong thing
A long term locks the count and the bundle you accepted, which on a reactive renewal is a 2x to 4x outcome held for years rather than one.
The buyer side move is to correct the count, contest the bundle, and only then trade term for written rate protection. Negotiate term length against rate protection, never term for its own sake. The estate level view sits in the licensing pillar and the negotiation sequence in the negotiation series.
What the renewals measured, 2024 to 2025
Two cuts of the engagement file, and neither depends on the vendor conceding a rate.
Where the deployment footprint justified the core virtualization tier or less, so the packaging was set by the quote rather than by the estate.
From core count corrections and host consolidation alone, using evidence the buyer already held.
Both are counting exercises. The discount conversation happens after them or it happens against the wrong number.
Your first five moves
- Rebuild the core count from your own inventory, removing dead and migration scheduled hosts with evidence, because count corrections recovered 15 to 30 percent on their own.
- Map deployed components against both bundles and demand justification, since 2 in 3 estates were quoted the full stack when the smaller tier fit.
- Consolidate low core hosts before the count is fixed, as small host estates add 15 to 40 percent on the sixteen core minimum alone.
- Stand up a funded migration assessment on the portable tier, visibly, because a named platform and a workload list is the minimum credible signal.
- Trade term only for written rate protection. The Broadcom practice and the spend health check rebuild the count before the quote arrives.
Frequently asked questions
What changed in the licensing?
Perpetual licenses and standalone products became per core subscriptions concentrated into two bundles, with a sixteen core floor per socket and support renewals ending at expiry.
How much more do estates pay?
Prepared estates held increases to 1.2x to 1.8x prior spend. Unprepared estates renewing reactively landed at 2x to 4x.
Why is the spread so wide?
Because the quote prices the estate the vendor can see: historical entitlement at the richest plausible bundle. Every correction the buyer brings narrows it.
How often is the bundle wrong?
Roughly 2 in 3 estates were quoted the full stack when their deployment footprint justified the core virtualization tier or less.
What do count corrections recover?
Between 15 and 30 percent, from core count corrections and host consolidation alone, before any discount is discussed.
Who absorbs the worst of it?
Edge and branch heavy estates paying the socket floor on many small hosts, estates quoted components they never deploy, and buyers with no credible exit.
Who came out relatively fine?
Dense consolidated data centre estates already using much of the full stack. Their per core economics were tolerable and the bundle matched the deployment.
Does the migration threat have to be real?
The assessment does; the migration does not. A funded assessment with a named platform and a workload list is the minimum signal a vendor will verify.
Is a long term the right defense?
Not on its own. A long term locks the count and bundle you accepted, so it holds a bad outcome for years rather than one. Trade term only for written rate protection.
What order should the work run in?
Count, bundle, consolidation, migration signal, then rate. The discount conversation happens after those or it happens against the wrong number.
