Broadcom VMware Practice
vSphere and vSAN subscriptions. The Broadcom era CIO playbook.
Per core bundles replaced a la carte licensing. The edition you pick and the core base you sign decide the next three years.
Per core bundles replaced a la carte licensing. The edition you pick and the core base you sign decide the next three years.
A CIO playbook for vSphere and vSAN in the Broadcom era: what changed, the four realistic subscription paths, core base engineering, and the contract terms that protect the term.
Broadcom replaced perpetual licenses and a la carte SKUs with subscription bundles, sold per core. vSphere and vSAN now arrive primarily inside VMware vSphere Foundation and VMware Cloud Foundation editions rather than as standalone purchases.
The pricing unit is the core, with minimums per CPU. The portfolio simplification is real; so is the repricing that came with it.
Per the vSAN product page, capacity now ships as a per core entitlement inside the bundles with expansion on top. Storage heavy estates need the capacity math done before the renewal call, not after.
Broadcom now sells two bundles, VMware Cloud Foundation and VMware vSphere Foundation, and your job is to match deployed capability to the smaller one that still covers it.
Cloud Foundation is the full private cloud stack: vSphere, vSAN, NSX networking, and the Aria management suite. vSphere Foundation is the compute centric bundle: vSphere plus a smaller vSAN entitlement and lifecycle tooling.
The standalone Standard and Enterprise Plus SKUs that many estates bought a la carte are effectively gone for new vSphere 9 contracts. The vSphere Foundation product page is where most compute estates now land.
The buyer side trap is accepting a VCF quote for a VVF workload. NSX and Aria carry real per core cost. If you do not run software defined networking or the Aria suite, VCF bills capability you never switch on.
The two Broadcom bundles at a glance
| Bundle | vSAN per core | Core stack | Best fit |
|---|---|---|---|
| VMware Cloud Foundation | 1 TiB | vSphere, vSAN, NSX, Aria, automation | Estates running the full private cloud stack |
| vSphere Foundation | 0.25 TiB | vSphere, entry vSAN, lifecycle management | Compute centric estates with modest storage |
| vSAN capacity add on | Per TiB expansion | Storage capacity above the bundle include | Storage heavy clusters |
A compute centric estate with modest shared storage usually fits vSphere Foundation, not VCF. VVF includes 0.25 TiB of vSAN capacity per core; VCF includes 1 TiB per core.
The Broadcom core counting article confirms the vSAN entitlement is tied to purchased cores. Map raw TiB contributed to the cluster before you assume you need the expansion.
Four realistic paths exist: VCF for the full stack estate, VVF for compute centric estates, a minimized bundle footprint with third party tooling around it, and a planned migration off the hypervisor. The right answer is workload math, not loyalty.
Broadcom era paths for a vSphere and vSAN estate
| Path | Best fit | Cost profile | Key risk |
|---|---|---|---|
| VMware Cloud Foundation | Full private cloud stack in active use | Highest per core, broadest capability | Paying for unused components |
| vSphere Foundation | Compute virtualization with some vSAN | Mid range per core | Bundle creep at renewal |
| Minimized footprint | Stable estates with third party ops tooling | Lowest VMware line | Integration effort |
| Hypervisor migration | Estates with portable workloads | Migration cost up front | Execution risk and timeline |
Count licensable cores per host, apply the per CPU minimums, and model consolidation before signing. Hardware refresh onto dense hosts routinely cuts the billable core base 20 to 40 percent, which compounds against every subscription year.
The standard advice is that Broadcom does not negotiate, so take the quote and budget the uplift. We disagree. In roughly 20 of the 25 to 35 Broadcom era renewals Morten Andersen advised in 2024 to 2025, material movement came from edition right sizing, core base engineering, and a credible migration scenario, even where the headline discount stayed thin. The buyer side move is to negotiate the base and the edition, not the percentage. Broadcom holds the rate card; you hold the core count and the workload placement.
Source: Redress Compliance advisory engagement file, 2024 to 2025.
Broadcom changed the price list. You still control the core count, the edition, and the credibility of your exit. Negotiate those.
Broadcom bills a minimum of 16 cores per physical CPU, so low core and sparsely populated hosts pay for cores they do not physically have.
Count every physical core across each host, then apply the 16 core floor per socket. A host with two 10 core CPUs bills 32 cores, not 20. The floor rewards dense hosts and punishes wide, lightly populated ones.
Broadcom briefly went further. In April 2025 it introduced a 72 core minimum per subscription order, which forced tiny clusters to buy cores they could never use. After backlash it reversed the change and returned to the 16 core per CPU floor.
The lesson for buyers is that Broadcom minimum rules move. Confirm the current floor in writing at quote time, because a rule that resets your billable base is worth more than a discount point.
How the 16 core floor bills common hosts
| Host configuration | Physical cores | Billable cores after floor | Waste from floor |
|---|---|---|---|
| 2 sockets x 8 core | 16 | 32 | 16 cores |
| 2 sockets x 10 core | 20 | 32 | 12 cores |
| 2 sockets x 16 core | 32 | 32 | 0 |
| 2 sockets x 32 core | 64 | 64 | 0 |
Host density changes the subscription because the 16 core floor and the per core meter both reward consolidation onto fewer, denser hosts. Sockets at or above 16 cores each waste nothing against the floor.
Consolidating a sprawl of older 8 and 10 core hosts onto modern 32 core hosts routinely cuts the billable core base 20 to 40 percent. That cut compounds against every subscription year in the term.
Rightsizing means cutting billable cores and unused bundle capability before Broadcom quotes, not arguing the number after the quote lands.
Start with a physical core inventory per host, then model a consolidation refresh onto dense silicon. The billable base you present, not the base you inherited, is what the renewal prices against.
Storage is where rightsizing quietly pays. vSAN capacity is metered in raw TiB, and estates over provision it by habit. Measure actual raw contribution with the Broadcom tooling before buying expansion capacity.
Finish the core inventory, the capability map, and a priced consolidation scenario before the vendor call. Walking in with a lower engineered core base changes the starting number, not just the discount.
The levers that move a Broadcom quote are structural: edition choice, the engineered core base, term length, and a credible exit, not the headline discount percentage.
Broadcom guards the per core rate card and concedes little on percentage. Buyers who chase the discount lose. Buyers who change the edition and the core base change the number the rate applies to.
Renewal levers and how to pull them
| Lever | What it moves | How to pull it |
|---|---|---|
| Edition right sizing | Per core rate and included capability | Prove VVF covers the workload instead of VCF |
| Core base engineering | The billable core count | Consolidate onto dense hosts before quoting |
| Term length | Rate and renewal cap | Trade a longer term for a capped uplift |
| Credible exit | Willingness to discount | Price a funded migration scenario |
| Payment timing | Effective cost | Align signature to Broadcom quarter end |
Term length cuts both ways. A longer term can buy a lower rate and a renewal cap, but it also locks your core base. Only commit a long term once the core base is engineered down and true down rights are written in.
Broadcom quarter and year end timing can help buyers who are ready to sign, because sales teams chase bookings against period targets. The leverage only exists if your core base and edition analysis are already finished.
Five terms decide whether year one pricing survives: renewal caps, edition lock language, core count true down rights, support commitments, and audit posture. Subscription estates that skip these renegotiate from zero every cycle.
A funded migration assessment, even one you never execute, is the only leverage Broadcom consistently respects. Estates that priced an exit before renewal negotiated from a different position than estates that could not.
An alternatives pressure play still works in 2026, but only when the migration scenario is priced, scoped, and genuinely executable, not a bluff.
Broadcom has watched customers threaten to leave for two years. A vague threat moves nothing. A funded migration assessment with named target platforms, workload counts, and a timeline is the version the account team takes to its desk.
Credible targets in 2026 include Nutanix, Proxmox, and public cloud rehosting for portable workloads. Each carries migration effort that often exceeds the license saving in year one, which is why the scenario has to be honest about cost.
The play works even for estates that stay. A priced exit reframes the renewal from a forced repricing into a choice, which is the footing from which structural terms get conceded. See our enterprise comparison for the migration risk view.
The wider licensing picture is in the Broadcom VMware licensing pillar and the VCF pillar. The Broadcom VMware practice runs renewal defense end to end.
As per core subscriptions inside bundles, primarily VMware Cloud Foundation and vSphere Foundation, with minimum cores per CPU. Standalone perpetual licensing is gone, and every renewal is a repricing event.
Unprepared estates in our 2024 to 2025 file saw 2x to 4x uplifts, driven by bundle capability they never deployed and unengineered core bases. Prepared estates cut that materially through edition and core base work.
Yes, on structure more than rate. Edition right sizing, core true down rights, renewal caps, and a credible migration scenario all moved outcomes in our engagements, even where headline discounts stayed thin.
Price the scenario either way. Estates with portable workloads gain leverage from a funded migration assessment even if they stay; estates that cannot credibly leave should engineer the core base and lock protective terms instead.
Paying for components you never deploy. Map deployed capability against bundle contents before accepting a VCF quote when VVF or a minimized footprint covers the actual workload.
VCF and VVF differ by scope. VMware Cloud Foundation is the full stack with vSphere, vSAN, NSX, and Aria, and 1 TiB of vSAN per core. vSphere Foundation is compute centric with a 0.25 TiB per core vSAN entitlement. Most compute estates fit VVF.
The minimum is 16 physical cores per CPU. Broadcom bills at least 16 cores for every populated socket even if the processor ships fewer, so low core and sparsely populated hosts pay for cores they do not physically have.
Yes, briefly. Broadcom introduced a 72 core minimum per subscription order in April 2025, then reversed it after backlash and returned to the 16 core per CPU floor. Confirm the current minimum in writing at quote time.
It depends on the bundle. VMware Cloud Foundation includes 1 TiB of vSAN capacity per purchased core. vSphere Foundation includes 0.25 TiB per core, rounded up to the next TiB. Capacity beyond the include is bought as a per TiB expansion.
Start at least nine months before the renewal date. Broadcom timelines punish late starts, and the core inventory, capability map, consolidation model, and a priced exit all take time. Walking in early with an engineered core base is what changes the number.
Core counting worksheets, bundle capability maps, true down clause language, and the migration scenario that creates leverage.
Used across more than five hundred enterprise engagements. Independent. Buyer side. Built for procurement leaders running the next renewal cycle.
