Broadcom VMware Licensing Changes

Broadcom replaced VMware's 168-plus point products with two main bundles, moved every customer to per-core subscriptions with a 16-core-per-CPU minimum, and ended perpetual license sales, the three changes that reshaped every VMware cost model overnight. If you bought VMware before the acquisition, almost nothing about how you are licensed and billed survives unchanged, and the gaps between the old and new models are exactly where renewal costs balloon.

This guide maps each change and what it means for buyers, then sets out your options under the new model. It extends our VMware by Broadcom licensing guide and the firm's VMware by Broadcom advisory practice.

From 168 products to two bundles

The most visible change is consolidation. The sprawling VMware catalog collapsed into two principal offerings, VMware Cloud Foundation and VMware vSphere Foundation, with a small set of add-ons around them. Customers who licensed a handful of specific products, say vSphere plus a single management tool, now buy a bundle that includes capabilities they may never use. That bundling is the mechanic behind much of the cost increase, because you are no longer paying for what you selected; you are paying for a suite that was assembled for the average customer, not for you.

The consolidation also removed the ability to buy cheap, narrow editions for non-critical workloads. A test-and-development cluster that previously ran on an inexpensive standalone product now sits inside the same per-core bundle as production, which raises its cost disproportionately. The full breakdown of what moved where is in our Broadcom changes analysis.

From perpetual to per-core subscription

The table below contrasts the old and new models on the dimensions that drive cost. The shift from per-socket to per-core with a 16-core floor is the change most likely to surprise teams.

A server with two 8-core CPUs is now billed as 32 cores because of the minimum, even though it has 16 physical cores. Audit your actual core counts and socket population before accepting any quote, because the gap between physical reality and licensable count is pure avoidable cost on lightly populated hosts.

What the changes mean for your cost model

Three practical consequences follow. Recurring cost replaces one-time cost, so multi-year budgeting changes shape and the finance team needs to model an annuity rather than a periodic refresh. The core minimum penalizes lightly populated sockets, so consolidation onto fewer, denser hosts now directly lowers the bill in a way it did not under per-socket pricing. And bundling means the path to savings runs through documenting actual usage to justify a narrower construct or a credit, rather than through picking cheaper individual products, because the cheaper products no longer exist.

Each of these is a lever as well as a cost. Consolidation lowers the core count; usage documentation supports a bundle challenge; and modeling the recurring cost over five years rather than three exposes the renewal cliff in time to plan for it. Our guides to the Broadcom price increase and negotiating with Broadcom show how to pull each one in a live renewal.

Your options under the new model

Buyers have three credible paths. Subscribe and negotiate hard, using right-sizing and a costed exit to discipline the quote. Migrate to an alternative hypervisor where the workload allows and the three-year math favors it. Or adopt a hybrid, keeping critical, feature-dependent workloads on VMware while moving standard workloads to a cheaper platform, which caps your VMware core count at the genuinely VMware-dependent subset.

The decision turns on feature dependence, migration risk, and three-year cost, which we model in our VMware alternatives and VMware migration guides. The renewal-cliff math behind the new subscription model, and how to price it before you commit, is detailed in our VMware Cloud Foundation pricing analysis.

Why the changes happened so fast

The speed of the transition caught many customers unprepared, and understanding the logic helps you plan rather than react. Broadcom's acquisition model favors simplifying a portfolio to its highest-margin core, concentrating on the largest enterprise customers, and converting one-time license revenue into recurring subscription revenue that markets value more highly. The two-bundle structure, the per-core subscription, and the end of perpetual sales are all expressions of that single strategy, which is why they arrived together rather than in stages.

For buyers, the implication is that none of these changes is a temporary policy likely to reverse. They are structural and durable, so planning should assume they persist and that subscription pricing continues to rise at renewal. The customers who absorbed the transition best were those who treated the first renewal not as a one-time shock to survive but as the start of a permanently different cost structure to manage, and who built the consolidation, documentation, and alternative-evaluation capabilities to manage it on an ongoing basis.

The 16-core minimum in practice

The per-core minimum deserves its own attention because it is the change that most often produces an unexpected number. Under the old per-socket model, a host with two lightly populated CPUs was cheap to license. Under the new model, each CPU is licensed for at least 16 cores regardless of how many it physically has, so a server with two 10-core CPUs is billed as 32 cores rather than 20, a 60% overcharge relative to its actual cores. Multiply that across an estate built without the minimum in mind and the impact is substantial.

The defense is consolidation onto fewer, fully populated sockets, which aligns the licensable count with the physical reality and eliminates the minimum penalty. This is a concrete engineering action with a direct financial return under the new model, and it is the first thing to do before any renewal quote is prepared. Audit your socket population and core density now, because every lightly populated socket is paying the minimum penalty every year until you address it.

The hybrid path

Many estates land on a hybrid: keep the feature-dependent, hard-to-migrate workloads on a small VMware footprint and move the standard majority to a cheaper platform. The hybrid caps the VMware core count at the genuinely VMware-dependent subset, which can cut the VMware bill sharply while avoiding the riskiest migrations. It adds operational complexity from running two platforms, so it suits organizations with the engineering depth to manage both, but for a large mixed estate it frequently produces the best blend of cost and risk.

Support and the cost of standing still

One change that buyers underweight is what happens to support. Under the new model, support is bundled into the subscription rather than purchased separately, so letting a subscription lapse means losing updates and support together, not just one of them. For the perpetual licenses you already own, support ends when your existing support contract expires, after which you hold a license to run software that no longer receives security patches.

That makes standing still a real decision with real cost, not a free default. Running unsupported, unpatched virtualization in production is a security exposure that compounds quietly until an incident makes it expensive and visible. The honest options are to subscribe, to migrate, or to accept and explicitly manage the risk of unsupported status for a defined, short transition period while a migration completes. Drifting into unsupported status by inattention is the one path with no upside.

Factor the support reality into every comparison. A subscription that looks expensive includes ongoing patches and support; a perpetual license you keep running does not, and the cost of the resulting risk belongs in the model. When support is priced correctly, the gap between subscribing and standing still narrows, and the case for a deliberate decision, in either direction, strengthens.

An action plan for the new model

Reduce the changes to a concrete plan. Inventory your real core counts and socket population, because the 16-core minimum penalizes lightly populated sockets and consolidation directly lowers the bill. Document which bundle components you actually use, both to support a narrower commercial construct and to protect yourself in any audit. Model the recurring subscription over five years with the renewal cliff included, so the true cost is visible before you commit.

Then evaluate the three paths honestly: negotiate and subscribe, migrate, or run a hybrid that keeps only the feature-dependent workloads on VMware. Price each over the same horizon, including migration and support, and decide on numbers rather than inertia. The estates that handle the transition well are those that treat it as a planning exercise with a clear owner and a costed comparison, rather than a renewal to rubber-stamp under time pressure. Our advisors build that plan and run the negotiation that follows it.

Timing your response to the changes

The window to respond well is the months before your current contract expires, not the weeks after the renewal quote lands. A buyer who begins six months out can complete a consolidation exercise, build a costed alternative, document product usage, and model the five-year cost before the negotiation opens, arriving with a full position rather than reacting under time pressure. A buyer who waits until the quote arrives has none of that and negotiates from weakness.

Use the lead time to decide your walk-away point as well, because the question that determines your real bargaining power is what you will actually do if the vendor holds firm. An organization that has priced and board-approved a migration alternative can credibly decline an unacceptable quote; one that has not will renew on whatever terms are offered. The changes themselves are fixed, but your readiness to respond to them is entirely within your control, and readiness is what separates the buyers who manage the transition from those who simply absorb it.

Finally, treat the first renewal under the new model as the start of an ongoing discipline rather than a single event. Subscription pricing resets at each renewal, so the consolidation, documentation, and alternative-evaluation work you do now becomes a repeatable capability that compounds across future cycles.

The bottom line

Broadcom's changes are comprehensive: two bundles, per-core subscription, a 16-core minimum, and the end of perpetual sales. None of it is reversible, but all of it is plannable. Audit your real core counts, document actual product usage, and price both the negotiated subscription and the exit before you renew. The buyers who treat the new model as a planning problem rather than a billing surprise consistently land the better outcome. Our advisors build that plan across every VMware estate.