Pricing both bundles before the quote arrives was worth 15 to 25 percent on the first offer
Both bundles license per core with the same per processor minimum, so the sticker comparison looks simple and is not. The feature gap decides whether the premium buys anything, and in most estates it did not: the capabilities justifying the higher bundle were still undeployed a year after signature. The cheapest defence is arithmetic done before the seller frames the choice.
Prepared by Redress Compliance · August 11, 2026 · Broadcom advisory. Based on 25 to 35 Broadcom VMware renewals advised, 2024 and 2025.
Executive summary
In about 7 out of 10 estates, the features that justified the higher bundle premium were not deployed a year later.
The networking, storage, and management layers that distinguish the full private cloud stack from the compute focused option were bought and then not implemented, which makes the premium a purchase of intent rather than of capability.
Separately, 30 to 50 percent of estates on the full stack used only the compute layer in practice, which is the same finding measured from the other direction.
Buyers who priced both bundles against each other before the quote arrived landed 15 to 25 percent lower on first offer.
That is the single most cost effective action in this decision and it requires no negotiation at all: the arithmetic is public, the core count is yours, and doing it in advance means the seller frames a comparison you have already made rather than one you are meeting for the first time.
Both bundles carry the same sixteen core per processor minimum, which adds 10 to 25 percent on hosts with small core processors. That means the bundle choice and the core minimum are separate problems and both apply regardless of which you pick.
An estate that solves the bundle question and ignores the minimum has fixed one of two independent inflations, and the minimum is the one that follows the hardware rather than the software decision.
Estates under 5,000 cores had materially more leverage than the public messaging suggests, provided they credibly scoped a partial migration. The prevailing narrative is that only very large estates can negotiate, and our engagements did not support it.
What mattered was whether a partial migration had been scoped and costed, not the absolute size of the footprint, which puts the leverage inside the buyer's control rather than in the size of their invoice.
What each bundle actually contains
| Layer | Full private cloud stack | Compute focused option |
|---|---|---|
| Hypervisor and compute | Included | Included |
| Software defined networking | Included | Not included |
| Software defined storage | Included | Entitlement differs |
| Cloud management and automation | Included | Not included |
| Licensing unit | Per core with a per processor minimum | Per core with the same minimum |
The bundle question and the core minimum question are independent, and conflating them is how estates fix half the problem.
The bundle decides which software layers you are entitled to and therefore what the premium buys; the minimum decides how many cores you are billed for regardless of which bundle you chose.
A host with two small core processors bills at the floor on either option, so an estate that correctly downgrades to the compute focused bundle while leaving a horizontally built fleet in place has removed the software overreach and kept the hardware inflation.
Solve both, in that order, because the bundle decision is available at the next renewal while the hardware layout moves only on a refresh cycle. The wider metric analysis sits in the VMware licensing comparison.
The feature usage audit that settles it
- Prove which layers are actually deployed, not which are entitled, because in 7 of 10 estates the distinguishing capabilities were still not in production a year after purchase.
- Price both bundles at your own core count before the quote arrives, which was worth 15 to 25 percent on the first offer and costs nothing but the arithmetic.
- Count the minimum separately, since it applies on either bundle and adds 10 to 25 percent on small core hosts, making it a hardware question rather than a licensing one.
- Scope a partial migration if you are under 5,000 cores, because leverage in our file tracked whether an alternative had been costed rather than how large the estate was.
- Treat the entitlement as the ceiling, not the plan, and revisit the bundle at each renewal, since a deployment that never happened in year one rarely happens in year three.
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 premium is bought on intent
A seven in ten non deployment rate a full year after signature is not a delivery failure, it is a purchasing pattern, and it is worth understanding why it recurs.
The higher bundle is presented during a renewal shaped by a large increase, at a moment when the buyer is looking for a reason the increase is justified.
A full private cloud stack supplies that reason: it reframes a price rise as an upgrade, and it does so with capabilities that are genuinely valuable to organisations that deploy them. What is missing from that conversation is a deployment plan, because nobody in the room owns one.
The networking and management layers require project work, skills, and a roadmap slot that the renewal negotiation neither creates nor schedules, so the entitlement arrives and the implementation does not. A year later the estate is running compute on a private cloud licence.
The correction is not to distrust the higher bundle, which is the right answer for organisations actually building the stack, but to require the same evidence you would require of any other project before funding it: who owns the deployment, which quarter it starts.
And what the first workload on the new layers will be.
If those three answers exist, the premium buys capability. If they do not, it buys intent, and intent renews annually at the same price as capability.
Doing the bundle arithmetic before the quote arrives protects against this too, because a buyer who has already priced both options is comparing a known alternative rather than evaluating a single proposal, which is exactly the difference that produced 15 to 25 percent on first offer.
The exit economics sit in the migration analysis.
- 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 bundle decisions, 2024 and 2025
Across roughly 25 to 35 Broadcom VMware renewals advised in 2024 and 2025, the bundle choice was rarely made on features. It was made on the core count arithmetic and the discount ladder:
Estates where the networking, storage, and management capabilities justifying the higher bundle were still not in production a year later.
Lower first offer achieved by buyers who priced both bundles against each other before the seller framed the comparison.
Three patterns recurred: in about 7 out of 10 estates the features justifying the higher bundle were not deployed a year later; buyers who priced both bundles before the quote arrived landed 15 to 25 percent lower on first offer.
And estates under 5,000 cores had materially more leverage than the public messaging suggests where they credibly scoped a partial migration.
Alongside those, 30 to 50 percent of full stack estates used only the compute layer, and the per processor minimum added 10 to 25 percent on small core hosts regardless of bundle. The buyer side move is a feature usage audit that proves which bundle the estate actually needs.
The wider library sits in the Broadcom practice.
Your first five moves
- Run a feature usage audit that proves deployment, not entitlement, because in 7 of 10 estates the capabilities justifying the premium were still undeployed a year after purchase.
- Price both bundles at your own core count before the quote arrives, which was worth 15 to 25 percent on first offer and requires only arithmetic you already have the inputs for.
- Require a deployment plan before funding the premium: who owns it, which quarter it starts, and what the first workload on the new layers will be.
- Count the per processor minimum separately from the bundle, since it adds 10 to 25 percent on small core hosts on either option and follows the hardware refresh cycle rather than the renewal.
- Scope a costed partial migration even under 5,000 cores, because leverage tracked whether an alternative existed rather than the size of the estate. The Broadcom practice runs the audit with you.
Frequently asked questions
What is the difference between the two bundles?
The full private cloud stack bundles compute with software defined networking, storage, and cloud management as one integrated platform. The compute focused option covers the hypervisor layer.
Both are sold as per core subscriptions with the same per processor minimum, so the difference is entitlement scope rather than licensing mechanism.
How often is the premium actually used?
Not often. In about 7 out of 10 estates the features justifying the higher bundle were still not deployed a year after signature, and 30 to 50 percent of estates on the full stack used only the compute layer in practice. Those are the same finding measured from two directions.
Why does the premium get bought if it is not deployed?
Because it is presented during a renewal shaped by a large increase, at the moment a buyer is looking for a reason the increase is justified. It reframes a price rise as an upgrade.
What the conversation lacks is a deployment plan, because the new layers need project work and a roadmap slot that the negotiation neither creates nor schedules.
What is the cheapest way to improve the outcome?
Price both bundles at your own core count before the quote arrives. Buyers who did landed 15 to 25 percent lower on first offer, and it requires no negotiation: the arithmetic is public and the core count is yours.
It changes the conversation from evaluating a proposal to comparing a known alternative.
Does the core minimum depend on the bundle?
No. Both carry the same per processor minimum, so a host with small core processors bills at the floor either way, adding 10 to 25 percent.
That makes it an independent problem: an estate that correctly downgrades the bundle while leaving a horizontally built fleet in place has fixed only half of the inflation.
Do smaller estates have any leverage?
More than the public messaging suggests. Estates under 5,000 cores negotiated materially better outcomes in our file where they credibly scoped a partial migration, which means leverage tracked whether an alternative had been costed rather than the absolute size of the footprint.
That places it inside the buyer's control.
What should be required before funding the higher bundle?
The same three answers any project needs: who owns the deployment, which quarter it starts, and what the first workload on the new layers will be. With those, the premium buys capability. Without them it buys intent, and intent renews annually at the same price as capability.
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.
