What Replaced vSphere Enterprise Plus
VMware vSphere Enterprise Plus was the cornerstone enterprise VMware licence — a per-socket perpetual licence that included vSphere's full feature set (DRS, HA, vMotion, FT, Storage vMotion) with annual support as a separate renewal. For most enterprises, vSphere Enterprise Plus represented the primary hypervisor cost and was the reference point for all VMware licence negotiations.
Broadcom replaced vSphere Enterprise Plus with two subscription tiers: VMware Cloud Foundation (VCF) and VMware vSphere Foundation (VVF). The choice between them is determined by whether your clusters require the full HCI stack (vSphere + vSAN + NSX) or compute-only virtualisation (vSphere without vSAN or NSX). Neither tier offers the per-socket perpetual model that vSphere Enterprise Plus used — both are per-physical-core annual subscriptions.
VMware vSphere Foundation (VVF)
- vSphere (full feature set including DRS, HA, vMotion, FT)
- vCenter Server (manages VVF clusters)
- Tanzu Kubernetes Grid (basic)
- VMware Aria Suite Starter
- Production support included
- No vSAN (storage is external)
- No NSX (networking is standard)
VMware Cloud Foundation (VCF)
- Everything in VVF
- vSAN (hyperconverged storage)
- NSX (software-defined networking)
- VMware Aria Suite (operations, automation)
- vSphere with Tanzu (full Kubernetes)
- HCX (workload mobility)
- Production support included
VCF vs VVF: Understanding the Price Difference
The approximately $30 per core per year difference between VVF and VCF represents the vSAN and NSX component costs, plus the expanded Aria Suite and advanced Kubernetes capabilities. For organisations that previously purchased vSphere, vSAN, and NSX separately, the VCF subscription price is broadly comparable to the sum of those components — though still significantly higher than the perpetual amortisation cost of the previous position.
The critical error many enterprises make is accepting a uniform VCF quote when only a subset of their clusters actually require vSAN and NSX. Clusters running vSphere only — without hyperconverged storage and without NSX networking — are appropriately licenced under VVF, not VCF. The $30 per core per year difference multiplied across hundreds or thousands of compute-only cores is material. We cover the estate segmentation mechanics in detail below.
"Broadcom's initial proposal almost always maps 100% of an estate to VCF. The correct answer is usually 40–70% VCF, 30–60% VVF — and the difference in annual cost is significant."
Cost Impact Modelling: Real Numbers
The following model represents a typical mid-enterprise VMware environment: 800 physical cores across production and non-production clusters, previously licenced on vSphere Enterprise Plus perpetual with a mix of vSAN and NSX in production and plain vSphere in non-production and DR environments.
Note: The above model uses indicative figures for illustrative purposes. Actual per-core list prices vary by tier selection and Broadcom's published pricing at time of contract. The model illustrates three important principles: even with a significant enterprise discount, VCF pricing is materially higher than previous perpetual positions; correct VCF/VVF segmentation reduces cost by 10–15% even at list price; and both segmentation and discount negotiation together represent the most effective cost management combination.
Estate Segmentation: Correctly Splitting VCF and VVF
The most systematic cost reduction available before negotiation begins is correct segmentation of your estate between VCF (full stack) and VVF (compute-only) tiers. The segmentation logic is straightforward: clusters that run vSAN (using local storage in the hosts) require VCF. Clusters that run NSX for overlay networking or microsegmentation require VCF. All other clusters — those using external storage and standard networking — are appropriately licenced under VVF.
In practice, many enterprise environments have a majority of their compute cores in non-HCI clusters — particularly in development, test, UAT, and disaster recovery tiers where HCI functionality was not deployed. These clusters should be VVF. Production environments with mature HCI deployments typically require VCF. The ratio varies, but in our experience with enterprise VMware estates, 30–50% of total cores qualify for VVF rather than VCF — generating meaningful cost savings from the segmentation alone.
DR and Standby Cluster Licensing
Disaster recovery standby clusters deserve special attention. Many Broadcom initial proposals include DR standby hosts at full VCF or VVF rates, treating them identically to production clusters. If your DR clusters are vSphere-only (no vSAN, no NSX), they should qualify for VVF pricing, not VCF. Additionally, for DR clusters where hosts are powered off except during testing and actual recovery events, there is a case for negotiating a reduced-rate DR licence — Broadcom has offered these in some engagements, though they are not a standard published offering and require specific commercial negotiation.
Core Count Discipline
Per-core pricing makes core count discipline as important as tier selection. Every core reduction in your committed count reduces your annual subscription cost by the per-core rate — and that saving compounds over the full subscription term. See our detailed discussion of core count validation in the subscription migration guide.
Three core count disciplines are critical for vSphere licensing specifically. First, processor generation selection for future hardware purchases — modern processors continue to increase core counts with each generation (Intel Sapphire Rapids, AMD Genoa). A server purchased in 2025 with 96 cores per CPU represents significantly more VCF licence exposure than a 2020-era 32-core server with the same workload capacity. Before the next hardware refresh, model the VCF core count implications of server selection — in some cases, selecting fewer higher-frequency cores for workload consolidation is commercially preferable to maximising core count per node.
Second, cluster rightsizing — removing underutilised hosts from production clusters before migration reduces your committed core count without reducing workload capacity. In many enterprises, average cluster utilisation is 30–50% at peak, meaning hosts can be removed from clusters and either decommissioned or placed in a separate lower-tier pool without operational impact.
Third, test and development rationalisation — legacy test environments often have VMware footprints that were established when perpetual licences had no ongoing cost. Under per-core subscription, every test host is an annual cost. Audit test environments for hosts that can be consolidated, decommissioned, or migrated to a cheaper alternative (Proxmox, Hyper-V) before counting them in VCF scope.
Negotiating vSphere Pricing with Broadcom
With correct estate segmentation and validated core counts established, the negotiation parameters for vSphere pricing are: total annual subscription value (higher value unlocks better per-core rates), commitment term (3-year minimum for meaningful discounts; 5-year for maximum), bundle complexity (simpler estates with clean VCF/VVF separation are easier to discount than complex mixed environments), and competitive pressure from alternatives evaluations.
Broadcom's discount structure for vSphere-based VCF is volume-tiered. For enterprise customers committing $500K–$1M annually on a 3-year term, 20–28% below list is achievable with standard negotiation. For customers at $1M–$3M annually, 28–35% below list. For customers above $3M annually with a credible alternatives track, 35–45% below list has been achieved in our engagements. The key is that these discounts are not automatically offered — they require structured negotiation with a prepared commercial position, supported by core count analysis and alternatives modelling.
For the complete negotiation playbook, see our VMware Broadcom Survival Guide. For a current assessment of your specific vSphere pricing position, contact our VMware advisory team.