The new VMware licensing model. Per core, subscription only.

Broadcom replaced VMware perpetual licensing with per core subscription bundles, and the renewal math now turns on core counts, bundle tier, and how credible your exit looks.

What changed in the VMware licensing model under Broadcom?

Broadcom ended perpetual VMware licensing and now sells subscription bundles priced per physical core, with a 16 core minimum per CPU. The portfolio consolidated around VMware Cloud Foundation, with vSphere Foundation as the smaller estate option.

Support is no longer a separate maintenance line. It rides inside the subscription, which is why the renewal compare against old maintenance looks so violent.

Why did Broadcom restructure the portfolio this way?

Bundles raise the revenue floor per customer and make line item negotiation harder. That is the design intent, and the buyer response has to work at the estate level, not the SKU level.

How does per core subscription pricing actually work?

VMware subscriptions meter every physical core on hosts running the software, and each CPU bills at least 16 cores even when it carries fewer. A two socket host with 12 core CPUs therefore licenses 32 cores, not 24.

• Count physical cores: hyperthreading does not change the count; sockets and cores do.
• Apply the minimum: every CPU bills at 16 cores or actual cores, whichever is higher.
• License the cluster: hosts in a cluster running VMware workloads license fully, including failover capacity.

Where do core counts go wrong?

Counts go wrong on decommissioned hosts still in the inventory export, on clusters that no longer run VMware, and on oversized failover capacity. A pre quote scrub of the RVTools export against the live estate is the cheapest savings available.

What did the 2026 price and packaging moves change?

The 2026 moves tightened the bundle structure further: VCF absorbed more of the advanced services, license portability between on premises and supported clouds firmed up, and minimum commit expectations rose for direct enterprise accounts. Broadcom publishes current packaging through the Broadcom support portal and product pages.

Channel economics changed too. Smaller accounts moved to authorized partners for the Broadcom portfolio, which changes who you negotiate with even when the price list does not move.

Does license portability change the buyer math?

Portability helps hybrid estates because the same subscription can follow workloads to supported clouds. It does not lower the price; it removes one excuse for double licensing during migrations.

Which buyer levers still work on a VMware renewal?

Four levers consistently moved VMware renewals in our file: a scrubbed core inventory, bundle tier fit, term length traded against caps, and a funded exit for a defined estate slice. None of them require actually leaving.

1. Scrub the core inventory against the live estate before the quote arrives.
2. Map feature use to the cheapest covering tier, cluster by cluster.
3. Trade a longer term only against a renewal cap and locked unit pricing.
4. Fund a real alternative for the edge or test estate and show the timeline.
5. Co terminate fragmented contracts into one negotiation event.

What does a credible exit look like to Broadcom?

A credible exit names the platform, the budget line, and the migration date for a specific cluster set. A slide that says we are evaluating alternatives moves nothing; a signed pilot moves the discount.

Where the common advice on the new VMware model is wrong

The standard advice says sign the shortest possible term and wait for Broadcom to soften. We disagree. In roughly 20 of the 30 plus VMware estates Fredrik Filipsson advised in 2024 to 2025, short panic terms simply repeated the shock a year later at a higher list, while buyers who segmented the estate, exited the slice with real alternatives, and committed the core under a cap got the durable number. The buyer side move is to decide what stays VMware for three years, cap it, and make the rest visibly contestable. Waiting is not a strategy; segmentation is.

What the engagement data shows

Three cuts of our advisory engagement file frame the size of the opportunity.

What to do next

Five moves turn this analysis into a lower invoice on the next renewal.

A sequence you can run this quarter

1. Export the full host and core inventory and reconcile it against live workloads.
2. Map feature use per cluster to the cheapest covering bundle tier.
3. Price a funded alternative for at least one defined estate slice.
4. Set the renewal calendar 12 months out with a named owner.
5. Draft the cap and portability clauses you will require in the order form.
6. Co terminate fragmented agreements into a single negotiation event.

White Paper · Broadcom / VMware

Broadcom VMware Renewal Survival 2026

The 2026 buyer side reference on Broadcom VMware renewals. Read it free.

Read the white paper

Frequently asked questions

Can you still buy perpetual VMware licenses?

No. Broadcom sells VMware as subscription only, priced per physical core with a 16 core minimum per CPU. Existing perpetual licenses keep running, but support renewals on them have ended in favor of subscription offers.

What is the 16 core minimum in practice?

Every CPU bills at least 16 cores even when it has fewer. A two socket host with 12 core CPUs licenses 32 cores, which is why host hardware shape changes the subscription bill.

How large is the typical renewal increase?

First subscription quotes in our 2024 to 2025 file ran 2 to 3 times prior maintenance spend. Buyers who scrubbed cores, fixed tier fit, and showed a funded exit closed 25 to 40 percent below the opening number.

Is VMware Cloud Foundation mandatory?

No. vSphere Foundation and remaining standalone editions cover estates that do not use the full VCF stack. Paying for VCF where the features go unused is the most common tier mistake we see.

Does a credible exit really change Broadcom pricing?

Yes. In our engagement file the discount moved when a defined estate slice had a named platform, budget, and migration date. Vague evaluation language moved nothing.

Should I sign a one year or three year term?

Sign the longer term only when it buys a renewal cap and locked unit pricing on the clusters you will keep. A short term without leverage just reprices the estate sooner.