You are registered. Your playbook is ready. Read the full 2026 edition of the VMware After Broadcom per-core renewal defense below.
Prepared by Atonement Licensing · buyer-side advisory · last reviewed June 2026. Figures are list-level or clearly labelled indicative ranges. The 3,000-core estate used below is a representative benchmark scenario for illustration, not a quote.
Executive summary
A VMware renewal quote under Broadcom is an opening position, not a fixed cost, and it is built around Broadcom's per-core subscription targets rather than the capacity you actually run. The model changed in late 2023, but the negotiation did not disappear. The gap between the quote and a defensible number lives in three places the buyer controls: the bundle edition, the core count Broadcom attributes to your hosts, and the term. Buyers who negotiate only the headline discount leave the structural money behind.
On a representative 3,000-core benchmark estate, an opening proposal that defaults to VMware Cloud Foundation across an unaudited core count models near $1.05M per year at indicative list-level pricing. The same workload, after an accurate core-count audit and a move to vSphere Foundation where the full stack is unused, models near $0.62M per year — a Year 1 reduction of roughly 40 percent on the same compute. The decision in front of procurement is which clusters genuinely need the full private-cloud stack, how many cores you actually license once phantom and retired hosts are stripped out, what term protects the price, and whether a costed alternative is on the table.
This playbook explains how Broadcom assembles a quote and where that gap hides: the per-core subscription mechanics and the 16-core-per-CPU minimum, the VCF-versus-VVF bundle decision, the core-count audit that funds everything else, the levers that move a Broadcom deal in sequence, the 150-day renewal timeline, the exit options that reset the anchor, and the contract terms that protect the deal in years two and three. Read it before your renewal window opens, not after Broadcom sends the quote.
How Broadcom builds a VMware quote
Broadcom acquired VMware in late 2023 and rebuilt the commercial model within months. Perpetual licenses are gone for new purchases. Most standalone products, including standalone vSphere editions and many add-ons, were folded into two main bundles. Pricing moved from per-CPU to per-core subscription, billed annually across a one, three, or five year term. The result is a quote that looks like a single number but is really three separate decisions stacked on top of one another.
Every quote is assembled from the bundle edition, the total core count Broadcom attributes to your hosts, and the term. Each of those is negotiable or correctable, and each is usually set in Broadcom's favor on the first pass. The account team works to its own quarter and fiscal-year targets and has more room to move than the opening figure suggests, but only against a buyer who has taken the quote apart.
The single blended quote exists so you react to the total instead of the components. Ask for the quote decomposed by host, by bundle SKU, and by year, then price each line on its own. The bundle line and the core count are where the largest corrections live, and neither survives contact with your own inventory data.
Action. Treat the bundle, the core count, and the term as three separate negotiations. Decompose the quote by host and SKU before any pricing conversation begins.
2Per-core subscription mechanics and the 16-core minimum
VMware subscriptions are now priced per physical core, with a minimum of 16 cores charged per physical CPU. A processor with fewer than 16 active cores is still counted as 16. A processor with more than 16 cores is counted at its actual core count. This single rule reshapes which hardware is economic to license, and it is the most common source of phantom cost in a Broadcom quote.
The practical effect is that older, low-core-count servers carry a hidden penalty, while consolidating workloads onto fewer, denser hosts can lower the licensed core total. The table below shows how the minimum changes the effective count for common configurations.
| Host configuration | Physical cores | Cores charged | Effect |
|---|---|---|---|
| 2 CPUs, 8 cores each | 16 | 32 | Minimum doubles the charge |
| 2 CPUs, 12 cores each | 24 | 32 | Minimum adds 8 phantom cores |
| 2 CPUs, 16 cores each | 32 | 32 | No penalty at the threshold |
| 2 CPUs, 24 cores each | 48 | 48 | Charged at actual cores |
Before you accept any core total, confirm it against your own configuration management data. Decommissioned hosts, lab clusters, and disaster-recovery nodes that no longer run production are commonly left in the count. Each removed two-CPU host saves at least 32 charged cores. The minimum also changes hardware-refresh economics: replacing two low-core sockets with newer high-density CPUs can hold the same compute capacity on fewer licensed cores once the per-CPU floor is taken into account. Model the licensed core total alongside the hardware cost rather than treating them as separate decisions.
The share of charged cores we routinely find that trace to the 16-core minimum on low-core hosts, retired nodes, and standby capacity left in the count (indicative).
The reduction in licensed cores achievable on a refresh that moves the same workload onto fewer, denser hosts above the per-CPU floor (indicative).
Action. Count your own cores from configuration data, strip retired and standby hosts, and model consolidation onto denser hardware before accepting Broadcom's core total.
3VCF and VVF: choosing the right bundle
Broadcom sells VMware mainly through two bundles. VMware Cloud Foundation, or VCF, is the full private-cloud stack: compute, storage through vSAN, networking through NSX, and management through Aria. VMware vSphere Foundation, or VVF, is the smaller bundle aimed at compute and basic management, with a capped vSAN entitlement rather than the full software-defined storage layer.
The account default leans toward VCF because it carries the higher per-core price. Many estates run vSphere for compute, use external or array-based storage, and have never deployed NSX or the full Aria suite. Those estates are paying for capability they do not use when they accept VCF. The bundle should follow the features actually enabled in your environment, not the account team's preference.
| Capability | vSphere Foundation | Cloud Foundation |
|---|---|---|
| Compute virtualization | Included | Included |
| vSAN storage | Capped per-core entitlement | Full entitlement |
| NSX networking | Not included | Included |
| Aria management suite | Limited | Full |
| Best fit | Compute estates on external storage | Full private-cloud stack in use |
Action. Pull a feature-enablement report per cluster, then quote VVF for every cluster that does not genuinely consume the full stack.
4The core count audit
The fastest saving in most VMware renewals is an accurate core count, not a discount. Broadcom prices on the cores you license, so every host you remove and every workload you consolidate reduces the bill directly. Run this audit before you respond to a quote, and start from your own inventory rather than the vendor's figure.
Pull the physical CPU and core layout of every host in every cluster, then separate production from non-production, retired, and standby capacity. Identify low-core-count hosts that trigger the 16-core minimum and model whether consolidation onto denser servers lowers the licensed total. Confirm which clusters genuinely need the bundle and which could move to a smaller edition. The bar chart below shows where over-licensing concentrates on the benchmark estate, expressed as the indicative share of charged cores each correction recovers.
The outcome is a defensible core number you own. When Broadcom quotes a higher figure, you can show the difference line by line rather than arguing in the abstract.
Action. Build an independent host, core, and cluster inventory, reconcile it against the quote, and present the corrected core number as your starting position.
5The levers that move a Broadcom deal
Discount is one lever among several, and it is the last one to pull. Buyers who negotiate only on the headline percentage leave the structural value on the table. Use the levers in sequence, starting with the ones that cost Broadcom the least to give and protect you the most.
| Lever | What it does | When it works best |
|---|---|---|
| 1. Core count accuracy | Strip inactive hosts and phantom cores from the quote | Always, before any pricing talk |
| 2. Bundle right-sizing | Move from VCF to VVF where the full stack is unused | When vSAN, NSX, and Aria are not deployed |
| 3. Term length | Trade a longer commit for a deeper discount and price hold | When your roadmap is stable for three years |
| 4. Price cap | Cap renewal uplift for the full term in writing | Always; uncapped uplift is the quiet cost |
| 5. Ramp and co-termination | Align contracts and phase growth into the commit | When multiple agreements renew apart |
| 6. Migration alternative | Bring a costed exit to reset the anchor | When a second platform is viable |
| 7. Support continuity | Protect support for licenses during transition | When you plan a partial migration |
| 8. Discount | The headline percentage, last | After every structural term is set |
The order matters. If you spend your negotiating room on discount first, you have nothing left to trade for the price cap or the bundle change, which are worth more over a three-year term than a few extra points off list.
Facing a VMware renewal in the next two quarters? Our advisors run this playbook with you.
Software Licensing AdvisoryAction. Sequence the negotiation: fix the cores and the bundle, lock the term and the cap, hold a costed alternative, and only then negotiate the discount.
Broadcom prices differently against a customer who can leave than against one who cannot. The costed alternative is the whole negotiation.6
The 150-day renewal timeline
A negotiating position is built, not found. By the time Broadcom sends a renewal quote, the buyers who do well have already done the work. This is the timeline we run, and the phases below sit on top of it.
Inventory and configure
Build an independent host, core, and feature inventory, then decide the bundle and model the right-sized core count before any pricing starts.
Cost the alternative
Scope and cost at least one migration alternative end to end on a representative cluster, so a credible option sits in your file.
Anchor and close
Open the commercial talk with your structure first, press for the cap, term, and support continuity, and close at a quarter end that helps you.
| Days before renewal | What to do | Why |
|---|---|---|
| 150 to 120 | Build an independent host, core, and feature inventory | You cannot negotiate what you cannot measure |
| 120 to 90 | Decide the bundle and model the right-sized core count | Set the configuration before pricing starts |
| 90 to 60 | Cost at least one migration alternative end to end | An alternative is the source of real bargaining power |
| 60 to 30 | Open the commercial talk with your structure first | Anchor on your terms, not the quote |
| 30 to 10 | Press for the price cap, term, and support continuity | Protect the terms that compound over the deal |
| 10 to 0 | Close at quarter end where the timing helps you | Timing pressure works in the buyer's favor |
Action. Start at day 150. A renewal opened 30 days out is negotiated on Broadcom's information, with no time to audit cores or cost an alternative.
7Exit options and migration alternatives
An exit option you have actually costed is the strongest lever a VMware buyer holds. You do not need to migrate to benefit from it. A credible, board-ready alternative changes the anchor of the conversation. The realistic alternatives vary in maturity, operational fit, and migration cost, and hypervisor choice is only part of the picture; management, storage, networking, and backup integration all carry switching cost.
| Alternative | Profile | Main consideration |
|---|---|---|
| Microsoft Hyper-V | Mature, often already licensed via Windows Server | Feature parity for advanced storage and networking |
| Nutanix AHV | Hyperconverged platform with built-in management | Hardware and licensing model change |
| Proxmox VE | Open source, low licensing cost | Enterprise support and operational maturity |
| Red Hat OpenShift Virtualization | Container and VM platform on Kubernetes | Skills and architectural shift |
Sequence the alternative properly. A migration that targets the easiest workloads first, such as test, development, and non-critical production, builds operational confidence and a real cost record without betting the core estate. That phased path is also the most credible thing to put in front of Broadcom, because it shows a route you have already started rather than a threat you cannot execute. Most enterprises that run this analysis decide to stay on VMware for a defined period while building optionality, which is a reasonable outcome; the value is in arriving at the renewal with a number that proves you could move.
Scope a representative cluster, not the whole estate, to produce a defensible migration number. A costed plan for 10 to 15 percent of the workload is enough to discipline the quote, and it costs a fraction of a full-estate study. Broadcom reads the difference between a real plan and a bluff quickly.
Want an independent read on your VMware renewal and a costed alternative? Talk to our advisors.
Book a 30 minute callAction. Cost one phased alternative on a representative cluster and keep it in the file, whether or not you intend to move this cycle.
8Contract terms that protect the deal
The pricing number is not the whole negotiation. The terms around it decide what the deal costs in years two and three. Subscription pricing makes year-end uplift the quiet cost of the deal, so cap the increase in writing for the full term and at the first renewal. A longer term can earn a deeper discount, but only commit to growth you can foresee, and phase increases rather than buying ahead of need. Protect support continuity for any perpetual licenses you still run, since those lose support when the contract lapses, which is the pressure point Broadcom uses at renewal.
Two further clauses repay the effort to negotiate them. A reduction right, even a limited one, lets you lower the committed core count at renewal if your estate shrinks through consolidation or migration, so growth is not the only direction the contract allows. And clear definitions matter: confirm exactly how cores are counted, how bundle entitlements are measured, and what triggers a true-up, so a later disagreement turns on agreed language rather than the vendor's reading. Precise terms today prevent an expensive interpretation fight in year two.
Audit the cores to the floor, match the bundle to the features actually in use, lock the term with a written uplift cap and a reduction right, protect support continuity for any perpetual estate, and keep a costed alternative on the table — then, and only then, negotiate the discount. The durable value in a Broadcom renewal lives in the configuration and the clause set, not in the headline percentage, and that is where the multi-year gap is recovered.
Key takeaways
- The first Broadcom quote is built to be moved. Prepare 150 days out, not 30.
- Audit your own core count before discussing price, and remove retired, standby, and lab hosts.
- Watch the 16-core-per-CPU minimum; low-core hosts and consolidation decide the licensed total.
- Choose VVF over VCF wherever NSX and full Aria are not in use.
- Use the levers in sequence: cores, bundle, term, cap, alternative, then discount.
- Cost a migration alternative on a representative cluster even if you intend to stay.
- Win the price cap, reduction right, and support continuity before the discount; they compound over the term.
Frequently asked questions
Why did our VMware renewal increase so much under Broadcom?
Broadcom retired perpetual licenses and most standalone products, moving customers to per-core subscriptions sold mainly as VMware Cloud Foundation and vSphere Foundation bundles. The bundle plus a per-CPU core minimum often raises the effective price even when your usage is unchanged, which is why an accurate core count and the right bundle matter more than the headline discount.
What is the per-core minimum on VMware subscriptions?
Broadcom applies a minimum of 16 cores per physical CPU for subscription licensing. A CPU with fewer than 16 active cores is still counted as 16, so low-core-count hosts inflate the subscription unless you consolidate onto denser servers or refresh hardware above the per-CPU floor.
Should we buy VMware Cloud Foundation or vSphere Foundation?
Choose VMware Cloud Foundation only when you use the full stack, including vSAN, NSX, and Aria. Many estates run compute and storage that vSphere Foundation covers at a lower per-core price. Map enabled features per cluster before accepting the larger bundle, because the account default leans to VCF.
Can we still buy perpetual VMware licenses?
No. Broadcom moved VMware to subscription only and stopped selling new perpetual licenses. Existing perpetual licenses still run, but they no longer receive support once your support contract lapses, which is the pressure point at renewal and the reason to negotiate support continuity for any perpetual estate you keep.
What are the realistic alternatives to VMware?
Common alternatives include Microsoft Hyper-V, Nutanix AHV, Proxmox VE, and Red Hat OpenShift Virtualization. Migration cost and operational risk vary, but a costed alternative on a representative cluster is what resets a Broadcom quote even when you choose to stay.
Get this playbook applied to your contract. Confidential assessment within one business day.
Book a 30 minute callThis playbook accompanies the VMware Broadcom Renewal Defense overview page. Related research: the VMware Broadcom Licensing Guide 2026, our software licensing advisory service, and the VMware and Broadcom vendor profile.