Broadcom's VMware Licensing Changes in 2026: What Every Admin Actually Needs to Know

If you've been running VMware in production for any length of time, the last 18 months have been a gut punch. Broadcom completed its $69 billion acquisition of VMware in November 2023, and since then the licensing model has been dismantled and rebuilt from scratch — in Broadcom's favor.

I'm a senior infrastructure engineer. I've deployed vSphere, Horizon, and VCF environments for enterprise clients for over a decade. I've watched this unfold in real time, had the conversations with account reps, run the numbers on renewal quotes, and helped organizations figure out what to do next. This isn't a rehash of press releases. This is what's actually happening on the ground.

What Broadcom Changed (The Short Version)

Broadcom eliminated perpetual licensing for VMware products. Effective February 2024, everything moved to subscription-only, bundled SKUs. The old à la carte model — buy vSphere, add vCenter, tack on vSAN, license NSX separately — is gone.

Here's what replaced it:

• VMware vSphere Foundation (VVF) — Per-core subscription. Replaces vSphere Standard/Enterprise Plus.
• VMware Cloud Foundation (VCF) — Full SDDC stack. Compute, networking (NSX), storage (vSAN), Aria Suite. The "premium" path.
• VMware vSphere Essentials Plus — Still exists, but only for small environments (3 hosts max, 96 cores).

The New Licensing Model, Explained

Per-Core Pricing

The single biggest operational change is the move to per-core licensing. This replaces per-CPU (socket) licensing.

Under the old model, a 2-socket server with 2x 16-core CPUs = 2 CPU licenses. Under VVF and VCF, every physical core must be licensed. Same server = 64 core licenses.

This hits SMBs hardest. A small shop with 3 hosts, each with 2x 8-core processors, used to pay for 6 CPU licenses. Now they need a minimum of 96 core licenses. At VVF pricing (~$120-150/core/year), that's $11,500-14,400/year — before vCenter, before vSAN.

The Bundle Structure

Broadcom wants you in one of two places:

The catch: if you need NSX — and in any serious enterprise environment, you probably do — you're being pushed to VCF pricing, which is meaningfully more expensive.

What Happened to Perpetual Customers

If you had perpetual licenses with active SnS:

1. Your licenses still work until Broadcom's defined EOL dates for your product version.
2. SnS renewals are being offered at escalating prices to "encourage" subscription migration.
3. Some organizations are getting buy-down offers — essentially a credit against VCF if they migrate now.

Broadcom's explicit strategy is attrition. They're not blocking you from running perpetual. They're just making every renewal conversation a pricing shock designed to make the subscription seem inevitable.

VVF vs VCF: Which One Do You Actually Need?

vSphere Foundation (VVF) Is Right For You If:

• You're running a straightforward compute virtualization environment
• You're using third-party storage (NetApp, Pure, HPE) and don't need vSAN
• You don't have NSX in your environment (or can live without it)
• You have a smaller core count (VVF is roughly 30-40% less expensive per core than VCF)
• You're an SMB or mid-market shop that doesn't need the full SDDC stack

VMware Cloud Foundation (VCF) Is Right For You If:

• You're already running NSX and need to maintain that investment
• You want Tanzu and container infrastructure integrated with your hypervisor
• Your environment needs vSAN stretched clusters or advanced erasure coding
• You're a large enterprise that can amortize the cost across 500+ cores
• You care about the Aria Suite automation and operations tools at scale

I'll be direct: VCF is a solid product. The integration between vSphere, NSX, and vSAN at the VCF level is genuinely best-in-class for on-premises software-defined infrastructure. The problem isn't the product — it's the pricing model applied to organizations that used to pay a fraction of this for simpler workloads.

The Math at Different Scales

Rough estimates — actual contract pricing varies based on deal size and Broadcom's discretion.

How Broadcom's Pricing Is Playing Out in the Field

Here's what I'm actually seeing, not what Broadcom's PR team says:

The large enterprises are mostly staying. Not because they're happy, but because migration costs, retraining, and risk at scale make alternatives difficult to justify quickly. They're negotiating hard, taking multi-year deals, and gritting their teeth.

The mid-market is genuinely evaluating alternatives. I'm in conversations with organizations in the 100-500 core range who are running proof-of-concepts on Proxmox, Nutanix, and Microsoft's Azure Stack HCI. Some are well down the migration path.

SMBs are the most at risk of being priced out. A shop with 3-5 hosts that was paying $5,000-8,000/year for vSphere Essentials Plus is now looking at $12,000-15,000/year minimum for VVF.

Service providers are fleeing. The old VSPP was terminated. The replacement — VMware Cloud Service Provider (VCSP) — has pricing that's made many MSPs and hosters either pivot to alternative hypervisors or pass massive cost increases on to customers.

The Products That Got Killed Entirely

Broadcom didn't just reprice. They discontinued a significant portion of the VMware portfolio:

• vSphere Standard — Discontinued (replaced by VVF at higher price)
• vSphere Essentials — Discontinued (Essentials Plus still exists with 3-host cap)
• VMware Workstation / Fusion — Free personal, subscription-only commercial
• vSphere Desktop — Discontinued
• NSX-T as standalone — Absorbed into VCF
• vCloud Director (standalone) — Moved to VCSP program
• VMware Horizon perpetual — Now Omnissa (spun out to KKR in mid-2024)

What Are Organizations Actually Doing?

Option 1: Accept VVF/VCF and Renegotiate Hard

For most enterprise environments, this is the realistic path for the next 1-3 years. Organizations choosing this path are:

• Taking 3-year deals to lock in pricing
• Right-sizing core counts aggressively (consolidating onto fewer, denser hosts)
• Pushing back hard in negotiations — Broadcom is not immune to churn pressure
• Evaluating true workload needs (do you really need VCF, or can VVF cover you?)

Option 2: Migrate to Proxmox VE

Proxmox is the most common VMware alternative I'm seeing evaluated seriously. It's open-source (Debian-based), uses KVM for virtualization and LXC for containers, and has a subscription model for enterprise support that's dramatically less expensive than VMware.

The challenges:

• Feature parity gaps with NSX (Proxmox networking is SDN but not NSX-grade)
• vSAN replacement requires rethinking storage (Ceph integration exists but is operationally different)
• Automation and operations tooling needs rebuilding
• Staff retraining is real — KVM is different from vSphere under the hood

Proxmox is a legitimate hypervisor for workloads that don't need the full VMware SDDC stack. For a 5-10 host environment running standard VMs, I'd take it seriously.

Option 3: Microsoft Azure Stack HCI / Hyper-V

Microsoft's play is Azure Stack HCI, positioning as a hybrid on-premises/Azure option. If you're already a Microsoft shop with Azure commitments, this deserves a look.

The honest assessment: Hyper-V has always lagged VMware in enterprise adoption and ecosystem depth. Azure Stack HCI solves some of that but introduces cloud billing complexity. Right answer for some, not a universal replacement.

Option 4: Nutanix

Nutanix has been the most aggressive in recruiting VMware refugees. Their AHV hypervisor is free with Nutanix software, and they've built migration tooling specifically targeting VMware environments.

Strong option if you're going hyper-converged and want an integrated stack without per-core pricing. Their node-based pricing is more predictable. The caveat: you're trading VMware lock-in for Nutanix lock-in. Not necessarily wrong, but eyes open.

Horizon Is Now Omnissa — What That Means

If you're running VMware Horizon for VDI or virtual apps, your product's ownership changed hands. Broadcom sold the EUC division to KKR in May 2024, and it's now Omnissa.

For existing Horizon environments:

• Existing licenses were converted/transitioned during the spinout
• Omnissa uses a subscription model going forward
• Horizon remains tightly integrated with vSphere (still requires VMware infrastructure underneath)

I'll cover the Omnissa transition in depth in a dedicated article — it deserves more space than a section here.

The 2026 Reality Check

Here's where things stand as of early 2026:

The migration wave is real but slower than headlines suggest. Most organizations are 12-24 months into planning but not 12-24 months into executing. Production migrations of large VMware estates are long cycles.

Broadcom has made some concessions. They extended support windows for some perpetual license customers after significant pushback. They haven't reversed the subscription-only model, but the pressure is visible.

The alternatives are maturing. Proxmox 8.x, Nutanix AHV improvements, and OpenShift Virtualization have all gotten meaningfully better.

Broadcom is betting on the installed base. They paid $69B. They need returns. For large enterprises with complex NSX deployments, that math works. For smaller shops, the calculation is increasingly going the other way.

What Should You Do Right Now?

1. Audit your core count. Know exactly how many physical cores you have licensed. This is your cost baseline.
2. Map your real feature usage. Do you actually use NSX? vSAN? Aria? Many organizations are paying VCF prices for features they barely touch.
3. Get a true VVF vs VCF comparison. If you're not using NSX at scale, VVF may be 30-40% cheaper with minimal capability loss.
4. Talk to Broadcom's renewal team before your SnS expires. The quotes improve when they think you're serious about leaving.
5. Run a Proxmox or Nutanix PoC if you're sub-500 cores. Not necessarily to migrate — just to have leverage in negotiations.
6. Don't panic-migrate. Production migrations are risky. A bad migration is worse than a bad licensing deal. Take the time to do it right.

Bottom Line

Broadcom made a $69 billion bet that VMware's installed base would rather pay up than leave. For most large enterprises, that bet is probably right in the short term. The switching costs are real.

But for SMBs, mid-market shops, and organizations with simpler workloads, the pricing changes have crossed a threshold. The value equation no longer works at the rates Broadcom is charging for VCF on 50-100 cores.

I'll keep covering this as it evolves. If you're working through a licensing decision or migration planning, subscribe below for updates.