NSX was paid for inside the Cloud Foundation bundle and actively used on only 30 to 50 percent of estates, because a component nobody buys separately never gets a usage review
Bundled reads as free, so nothing gets measured. The estates that never ran NSX still funded it, and still accepted the value story that justified the subscription increase.
Prepared by Redress Compliance · August 18, 2026 · Broadcom and VMware advisory. 20 to 30 renewals reviewed, 2024 to 2025.
Executive summary
NSX was actively used on only 30 to 50 percent of the estates reviewed. Cloud Foundation includes it whether or not it is deployed, so the entitlement arrives without a decision and without a measurement.
VCF subscription raised VMware cost by 40 to 100 percent against the prior perpetual baseline. The value case behind that rise leans on bundled components, and half the estates never turned one of the largest of them on.
Dropping NSX rarely saves license money, because it is already inside the bundle. What the usage number changes is the argument you can make about scope and tier, not a line you can strike.
Carbon Black uncertainty pushed 40 to 60 percent of buyers to evaluate endpoint alternatives. That is a signal about term length and roadmap commitments, not a reason to exit inside a renewal quarter.
What does the Cloud Foundation bundle actually include?
Compute, storage, networking, and management, sold as one subscription priced per core. Broadcom packages NSX as a component of VMware Cloud Foundation rather than as a standalone purchase.
You receive the NSX entitlement whether or not you deploy it. That is the whole mechanism behind this brief.
| Component | What it covers | Where the usage question sits |
|---|---|---|
| vSphere and vSAN | Compute and storage | Always deployed, so never in doubt |
| NSX | Software defined networking and microsegmentation | Deployed on 30 to 50 percent of estates reviewed |
| Aria | Operations and automation | Partially deployed in most estates |
| The subscription itself | Per core, on listed capacity | Listed capacity rather than used capacity sets the bill |
Listed capacity is what inflates the bill
VCF prices per core on subscription. The count that matters is the capacity you declare, not the capacity you use, which is why the core count is the first number to settle. Our Broadcom negotiation brief works the tier and core arithmetic in full.
Why does nobody measure a bundled component?
Because measurement follows purchase. A line item with a price attracts a renewal review, a usage report, and an owner. A component that arrives inside a bundle attracts none of those things.
So the question of whether NSX runs at all is asked in the deployment project, if it is asked, and then never again. Three years later the renewal arrives and nobody in the room knows the answer.
Bundled is not free, it is unpriced. The vendor knows what the component is worth to the package, and prices the package accordingly. The buyer who cannot say whether it is deployed has no way to test that price.
What the usage number is actually for
It is not a case for a refund. It is the evidence that the bundle you are being asked to renew is wider than the estate you run, which is an argument about tier and scope rather than about discount.
The Broadcom and VMware negotiation playbook
The core count, the tier decision, the bundled components worth measuring before renewal, and the clause set that holds the term.
Get the brief →What 20 to 30 Broadcom renewals showed
Across roughly 20 to 30 Broadcom VMware renewals Morten Andersen reviewed between 2024 and 2025, the bundle decision drove cost far more than any feature negotiation. Three patterns held across estate sizes.
VCF subscription raised VMware cost by 40 to 100 percent against the prior perpetual baseline. That is the number every buyer arrives with, and it is the number the vendor expects to defend.
NSX was paid for inside the bundle and actively used on only 30 to 50 percent of estates. The other half carried a networking stack they had not deployed, had not staffed, and in several cases could not confirm the status of without checking.
Carbon Black uncertainty pushed 40 to 60 percent of buyers to evaluate endpoint alternatives. Almost none of them wanted to migrate. They wanted a shorter term and a roadmap commitment, and the evaluation was how they earned the right to ask.
The pattern underneath all three is the same. The buyer arrives arguing about price, and the leverage is sitting in what the estate does and does not run.
- Core counts verified before they are rounded up into the subscription
- Bundle components mapped against what the estate actually deploys
- Renewal uplift exposure modeled over the full term, with the cap to ask for
Should a CIO keep or replace NSX?
Keep it if you run microsegmentation at scale, and treat the entitlement as sunk if you do not. Replacing NSX does not release license money, because the cost is already inside VCF. Capabilities are documented on the VMware NSX product pages.
| Factor | Keep NSX | Replace NSX |
|---|---|---|
| License cost | Already inside the VCF subscription | No subscription saving |
| Operational fit | Strong where segmentation runs at scale | Simpler stack to operate |
| Microsegmentation | Native to the platform | Third party tool and its own contract |
| Best when | The estate uses it heavily | The entitlement is sunk and unused |
Three questions before the decision
- Is NSX deployed and relied on today, and can somebody produce the evidence rather than the assumption?
- Does it reduce operational burden or add to it in the way your team actually runs the network?
- What does the replacement cost to run, fully loaded, rather than what it costs to license?
What should buyers do about the Carbon Black separation?
Shorten the term and demand roadmap commitments before extending, rather than exiting in a hurry. Broadcom moved to separate Carbon Black from VMware, with public context tracked on Broadcom news.
A divested product carries roadmap and integration uncertainty. That is a planning signal about how long to commit, not a reason to fund a migration you had not budgeted.
- Avoid long renewals during an ownership change, because the thing you are committing to may not be the thing that emerges.
- Ask for roadmap commitments in writing before agreeing to any extension beyond the short term.
- Shortlist endpoint alternatives now, which is what makes the shorter term a request the vendor takes seriously.
Time the decision to the ownership, not the renewal
Our brief on Carbon Black under Broadcom works the renewal window in detail, including when the conversation should open.
What the estates measured, 2024 to 2025
Two cuts of the engagement file frame the gap between what is bought and what runs.
Out of every estate carrying the entitlement inside its Cloud Foundation subscription, across the renewals reviewed.
Measured against the prior perpetual baseline, which is the comparison every buyer brings to the table.
The two numbers belong together. A rise of that size is defended with a bundle value story, and the value story assumes components that half the estates never switched on.
Watch the briefing · 5:38The VMware EstateWhat to count and what to prove before the Broadcom renewal conversation opens.
Your first five moves
- Produce evidence of whether NSX is deployed, not an assumption, since only 30 to 50 percent of estates reviewed were actually running it.
- Settle the VCF core count before any feature conversation, because listed capacity rather than used capacity is what sets the subscription.
- Price the bundle against the components you run, and put the unused ones on the table as a scope question rather than a discount request.
- Keep the Carbon Black term short until the ownership settles, and make the roadmap commitment a written condition of any extension.
- Shortlist the endpoint alternative before you need it. The Broadcom and VMware practice runs the deployment evidence before the quote arrives.
Frequently asked questions
Is NSX sold separately from VMware Cloud Foundation?
No. Broadcom packages NSX as a component of Cloud Foundation rather than as a standalone purchase, so the entitlement arrives with the subscription whether or not the estate deploys it.
How many estates actually run NSX?
Between 30 and 50 percent of the estates reviewed, across roughly 20 to 30 Broadcom VMware renewals in 2024 and 2025. The rest carried the entitlement without deploying it.
Does dropping NSX reduce the bill?
Rarely. The cost sits inside the Cloud Foundation subscription, so removing NSX from the estate does not remove a line. What the usage evidence supports is an argument about tier and scope.
How much did VCF raise VMware cost?
By 40 to 100 percent against the prior perpetual baseline in the renewals reviewed. That range is the comparison most buyers bring to the table and the one the vendor expects to defend.
What is the biggest cost lever in a VCF renewal?
The core count. VCF prices per core on listed capacity rather than used capacity, so the declared number moves more money than any feature negotiation.
Should we exit Carbon Black now?
Not on a renewal timetable. Keep the term short, ask for roadmap commitments in writing, and time any move to the ownership change rather than to a renewal quarter.
How many buyers evaluated endpoint alternatives?
Between 40 and 60 percent of the buyers reviewed opened an evaluation after the Carbon Black separation. Most were building the option rather than planning to use it.
Does microsegmentation justify keeping NSX?
Yes, where it runs at scale. NSX is strong value inside a bundle you are already paying for if the estate genuinely uses segmentation, and sunk cost if it does not.
What evidence proves NSX is in use?
Deployment and enforcement data from the platform itself, not a statement from the team that installed it. Several estates reviewed could not confirm the status without checking first.
When should the Broadcom renewal conversation open?
Early enough to produce the deployment evidence first. The core count and the component usage are the two inputs that change the shape of the quote, and neither can be assembled in the final weeks.
