The decision rule is architectural, not commercial: choose VCF if vSAN will be your primary storage, NSX your network layer and the automated lifecycle genuinely in use — choose VVF if you need enterprise-class compute virtualization and will run storage, networking and automation on other tooling. Both are per-core subscriptions on the same metric, so paying for VCF’s stack and deploying only the hypervisor is the single most common — and most expensive — sizing error in the post-Broadcom catalogue.
Published 18 December 2025 · Last reviewed 5 February 2026
Since Broadcom completed its acquisition of VMware and withdrew perpetual licensing in early 2024, the portfolio has collapsed into two principal subscriptions, both sold per physical core. VMware Cloud Foundation (VCF) is the full private-cloud platform: vSphere, vSAN with a per-core storage entitlement of 1 TiB, NSX networking, the advanced operations and automation layer descended from Aria and SDDC Manager, and integrated Kubernetes runtime services. vSphere Foundation (VVF) is the slimmer enterprise tier: vSphere with vCenter, the foundational operations toolset, Kubernetes runtime, and — since April 2025 — a 0.25 TiB-per-core vSAN entitlement of its own.
Editions below VVF exist (vSphere Standard and the Essentials Plus kit) but the practical enterprise choice in 2026 is between these two bundles, and the choice has consequences across a three-to-five-year term. Everything in the old catalogue — standalone vSAN, NSX, Aria SKUs, perpetual keys with support and subscription (SnS) — has either been folded into these bundles or closed to new purchase. What happens to an existing perpetual estate is its own decision, covered in VMware perpetual + SnS vs Broadcom subscription.
Both bundles share the same counting rules, which is why the comparison is about included components rather than metric mechanics: every physical core in a licensed host is counted, and each CPU is counted as at least 16 cores regardless of its actual core count. (A 72-core per-order minimum announced for April 2025 was withdrawn within days; the per-CPU floor is the one that stands.) The metric is identical; what you get per core is not.
This guide compares licensing programs as published by the vendor; it is general information, not legal or licensing advice for your situation, and it names no firms. Program terms change — verify current terms against your own quote. The firm directory lists VMware-capable advisors with balanced pros and cons, listed, not ranked.
| DIMENSION | VCF — CLOUD FOUNDATION | VVF — VSPHERE FOUNDATION |
|---|---|---|
| Positioning | Full software-defined data centre: the private-cloud platform | Enterprise compute virtualization with management and base operations |
| Compute & management | vSphere + vCenter, fleet-level lifecycle automation (SDDC Manager lineage) | vSphere + vCenter; lifecycle managed per cluster with standard tooling |
| Storage (vSAN) | Included at 1 TiB per licensed core; add-on capacity per TiB beyond that | Included at 0.25 TiB per licensed core (since April 2025); add-on per TiB |
| Networking | NSX included; advanced services (load balancing, firewalling) as add-ons | Not included — bring your own network and security stack |
| Operations & automation | Advanced operations, automation and log analytics included | Foundational operations toolset only |
| Kubernetes | Integrated runtime and fleet services | Runtime services included |
| Metric | Identical for both: per physical core; 16-core minimum per CPU (a 72-core per-order minimum announced for April 2025 was withdrawn within days); subscription terms typically one, three or five years | |
| Commitment shape | Larger per-core spend; economics favour estates that consume the whole stack; multi-year terms quoted with deeper discount pressure | Smaller per-core spend; less included, so less shelfware risk on conservative architectures |
| Exit & renewal dynamics | vSAN as primary storage and NSX as the network layer raise replatforming cost — renewal leverage must be built early | Hypervisor-level dependency only; credible alternatives are easier to stand up, which itself is renewal leverage |
Read the last two rows together: they are the commercial heart of the comparison. The more of VCF’s stack you deploy, the better its per-core economics look on paper — and the harder your next renewal becomes, because the cost of credibly walking away rises with every layer you adopt. VVF buys less and binds less. Neither is the wrong answer; signing one while behaving like the other is.
VCF fits estates that are genuinely platform-shaped: vSAN as primary storage rather than an experiment, NSX doing real segmentation or routing work, the automation layer deploying and patching clusters rather than sitting installed-but-idle, and a team sized to operate all of it. It also fits organisations consolidating onto fewer, larger clusters, where the 1 TiB-per-core storage entitlement absorbs what would otherwise be a separate storage purchase.
VVF fits the much larger population of estates that run vSphere as excellent compute virtualization on top of third-party storage arrays and existing network and security tooling. Since the April 2025 addition of a 0.25 TiB-per-core vSAN entitlement, VVF also covers moderate hyperconverged use — management clusters, branch sites, modest production vSAN — without forcing the full-stack purchase.
The test that decides it: list the VCF-only components and write next to each the date your team last used it in production. If NSX and the automation layer have no dates, you are describing a VVF estate. Conversely, if you are buying VVF plus a storage array refresh plus standalone network virtualization, total the three quotes before concluding VVF is the smaller commitment. Segmented estates — VCF on the strategic clusters, VVF on general-purpose compute — are common and legitimate, with the practical caveats covered in the FAQ.
Core counts are now the audit surface. Per-core subscription licensing makes the compliance question simple to state and easy to get wrong: every physical core in every licensed host, with the 16-core-per-CPU floor applied per socket. Hardware refreshes that move to higher-core-count CPUs silently expand the licence requirement; an accurate, dated core inventory is the cheapest compliance asset you can hold, and the first thing a compliance assessment establishes.
Renewal leverage must be designed in at signature. A subscription estate has no terminal value: when the term ends, so does the right to run. That moves the negotiation weight from the initial purchase to the renewal, where the vendor knows your replatforming cost precisely. Buyers who keep leverage do three things: they keep at least part of the estate on architectures with credible alternatives, they co-term agreements so renewal is one event with full estate weight behind it, and they start the renewal conversation early — the dynamics are the same as any major VMware renewal negotiation.
Term length is a two-sided instrument. Multi-year terms attract deeper discounting and protect against list-price movement, but they also lock the architecture decision for the duration. A three-year VCF commitment signed for a stack you adopt in year two is rational; the same commitment signed on aspiration is shelfware with a renewal cliff attached.
Add-ons deserve the same scrutiny as the bundle. vSAN capacity beyond the included entitlement, advanced network security services, and disaster-recovery components are all separately priced lines. A quote that looks lean at the bundle level can carry its real weight in add-ons — model the add-on growth path, not just the day-one configuration. Structuring all of this is exactly what a VMware licensing advisory engagement exists to do.
The minimums bite small footprints. The 16-core-per-CPU floor penalises low-core-count processors and can make a small site or edge location cost more than its hardware footprint suggests. (A 72-core per-order minimum announced for April 2025 was withdrawn within days but still circulates as if in force.) Price the real floors into any distributed-estate design before committing to the architecture.
Included storage is a function of cores, not need. Because vSAN entitlement scales with licensed cores, storage-heavy clusters with modest CPU counts are structurally short of entitlement and accumulate per-TiB add-ons. Model TiB-per-core for each cluster; the ratio decides whether the included capacity is a benefit or a rounding error.
Buying VCF as “future-proofing”. Paying the full-stack rate for components you might adopt is the most common sizing error in the current catalogue. The components do not expire, but your term does — and unconsumed entitlement strengthens the vendor’s renewal position, not yours, because your own usage data will show you paid for headroom you never used.
Assuming downgrade is symmetrical. Moving VVF→VCF at renewal is an upsell the vendor will gladly process; moving VCF→VVF means unwinding any dependency on NSX, the automation layer and the larger storage entitlement first. If you want the downgrade option live, keep the deployment within VVF-shaped boundaries during the term — or accept that the option is theoretical.
Forgetting that advice has incentives. Much VCF/VVF guidance reaches buyers through channels that earn margin on the outcome. The independence test applies here with full force: ask who profits from the configuration being proposed.
Operationally yes, and segmented estates are common: VCF on the clusters that consume the full stack, VVF on general-purpose compute. The practical constraints are contractual rather than technical — order minimums apply per purchase, the management tooling differs between tiers, and running two agreements means two renewal events unless they are co-termed. Model the split cluster by cluster before quoting.
No. Broadcom withdrew perpetual VMware licenses and support renewals on them; both VCF and VVF are term subscriptions. Existing perpetual keys keep functioning but stop receiving updates and support once the support contract expires — the trade-offs of that position are the subject of the perpetual-versus-subscription guide.
Each bundle carries an included vSAN entitlement that scales with licensed cores — a larger per-core allowance on VCF, a smaller one on VVF. Capacity beyond the entitlement is licensed separately per TiB. Because the entitlement is a function of core count, storage-heavy clusters with modest CPU footprints are the configurations most likely to need add-on capacity.
Contractually it is a renegotiation, not a toggle: the term ends and you order what you need next. The constraint is architectural — if NSX, the automation layer or the larger vSAN entitlement have become load-bearing during the term, a downgrade carries migration work. The realistic downgrade window exists only if you have kept the deployment within VVF-shaped boundaries.
Per CPU only. Each physical CPU is counted as at least 16 cores even if it has fewer. A 72-core per-order minimum announced for April 2025 was withdrawn within days; verify the floors on your own quote. Small sites and edge deployments feel the 16-core floor hardest.
No. This is a directory, not a ranking. Firms are listed alphabetically with balanced pros and cons. Independence is shown as a pro and reseller, Big-Four or vendor-side-audit ties as a con, both stated as factual trade-offs for you to weigh.
Sizing VCF against VVF, pricing the minimums into your architecture and designing renewal leverage before signature is precisely the work of a VMware licensing advisory engagement. The Broadcom VMware hub maps the vendor’s wider licensing world, and the directory lists every firm covering VMware — with balanced pros and cons, listed, not ranked. New to choosing an advisor at all? Start with how to choose a software licensing consultant.
What Broadcom changed, and your three end states →
Every firm for this work →
The renewal-event playbook and its firms →
Who your advisor really works for →
Filtered, listed, not ranked →
Every field guide on the site →
Tell us the vendor, the service you need and where things stand, and we will route your brief to firms that genuinely cover that combination. The directory and matching are free for buyers, no vendor ever sees your brief, and we add no markup.
Our weekly dispatch on vendor audit programs, regional developments and one buyer move. Subscribe to The Licensing Radar.