First quotes for comparable estates spread across a 2 to 4 times range, so there was no list price to argue against

Buyers arrive at a conversion looking for the benchmark rate, and there is not one. Comparable estates were quoted across a 2 to 4 times spread before any negotiation, which means the opening number describes the account team rather than the market, and only your own alternative sets a floor.

Prepared by Redress Compliance · August 16, 2026 · Broadcom VMware advisory. 20 to 30 conversions benchmarked, 2024 to 2025.

The licences you own did not expire. The support model that made them operable did, and the comparison base moved with it.

• Run unsupported. Viable for stable, isolated, internal workloads with compensating security controls. Unsupported vSphere keeps running, but every unpatched vulnerability accumulates risk the security team has to own explicitly, so treat it as a priced and time boxed bridge rather than a strategy.
• Third party support. Cuts support cost materially and provides no new versions or vendor security patches, which makes it a fit for a frozen tier rather than a developing one.
• Negotiated conversion. Convert where VMware remains strategic, against a credible alternative, on the shortest term you can hold. This is the right answer for the core, and only for the core.
• Migration. Nutanix, Hyper V, or cloud native paths for the segments where the three year total cost genuinely clears, which is rarely the whole estate and rarely nothing.
• Segment before you quote. Estates that split workloads first converted a smaller core and cut total spend further than any discount recovered across the whole footprint.
• Quote Broadcom only for the core that must stay, with a costed alternative visible for everything else. The levers are the alternative, the term, and the timing, in that order.

The instinctive response to a conversion quote is to ask what other people are paying, and to treat the answer as the target. That instinct works when a vendor prices from a list with a discount band around it. It fails here, because across the conversions we benchmarked the first quotes for comparable estates spread across a factor of two to four before any negotiation took place. A spread that wide is not a distribution around a market price. It is the absence of one.

What produces the spread is worth understanding, because it tells you which levers are real. Broadcom rebuilt the commercial model around bundles and a per core metric, but the conversion of an existing perpetual estate has no natural anchor: there is no prior subscription to uplift from, and the old support figure is explicitly not the base. What the account team has instead is your footprint, your renewal date, and an assessment of how likely you are to leave. The opening number is therefore a read on the account, not a read on the product, which is why two similar estates receive numbers that differ by more than any discount could bridge.

That explains the sharpest split in the file. Buyers who arrived with a costed migration alternative landed 25 to 40 percent below their first quote; buyers without one signed close to it. The alternative does not work as a threat, and account teams have heard every version of the threat. It works because it supplies the anchor the situation is missing. A funded, dated plan for a defined segment converts an open ended question about willingness to pay into a bounded comparison, and a bounded comparison is the only thing that behaves like a list price in this negotiation.

The second lever is segmentation, and it beats discounting because it changes the quantity rather than the rate. An estate that decides which workloads must stay on VMware, which can sit on a frozen tier under third party support, and which clear the migration math, arrives quoting Broadcom for a smaller core. Estates that did this cut total spend further than any discount recovered across the whole footprint. The mechanics of the perpetual right itself, and what survives non renewal, are covered in perpetual licence options; the exit costing sits in the market share read, which is also the honest account of how much of a migration plan actually gets executed. The wider library is in the VMware practice.

Across roughly 20 to 30 Broadcom VMware renewals and conversions benchmarked, quote variance rather than list price defined the negotiation:

The third pattern was structural. Estates that segmented workloads before negotiating converted a smaller core and cut total spend further than any discount alone achieved, because the discount applies to a quantity and segmentation changes the quantity.

Perpetual SKUs left the price list at the end of 2023 and have not returned. Existing licences run, but support contracts lapse at term end with renewal generally unavailable, which is what makes this a forced decision rather than an optional one.

Watch the briefing · 4:135 Tactics That Move the Number in a Broadcom NegotiationWhat shifts a conversion quote once the alternative is costed and the estate is segmented.

Did our perpetual VMware licences expire?

No. The licences you own remain yours and the software keeps running. What ended is the support and update path that made them operable long term. Broadcom stopped selling perpetual SKUs at the end of 2023 and support contracts on existing licences lapse at term end, with renewal generally unavailable.

Why is the subscription quote so much higher than our support bill?

Because the comparison base changed. You were renewing support at roughly 20 percent of licence value. You are now repurchasing the licence annually, inside a bundle whose scope you pay for whether you consume it or not. First quotes commonly land 2 to 4 times the prior run rate.

Is there a market rate we can benchmark to?

Not in the usual sense. First quotes for comparable estates spread across a factor of two to four before any negotiation, which is the absence of a market price rather than a distribution around one. The anchor has to come from your own costed alternative.

What actually moves the number?

A costed migration alternative, the term, and the timing, in that order. Buyers holding a funded and dated plan for a defined segment landed 25 to 40 percent below their first quote. Buyers without one signed close to it, regardless of how hard they pushed on discount.

What are the four options for a perpetual holder?

Run unsupported, buy third party support, convert to subscription at a negotiated price, or migrate off. Most enterprises blend the last three across workload tiers rather than picking one for the whole estate, because the right answer differs by tier.

Is running unsupported a real option?

As a bridge, for stable and isolated internal workloads with compensating security controls. Unsupported vSphere keeps running, but every unpatched vulnerability accumulates risk the security team must own explicitly. Price it and time box it rather than treating it as a strategy.

Does segmenting really beat negotiating harder?

Yes, because a discount applies to a quantity and segmentation changes the quantity. Estates that split workloads before negotiating converted a smaller core and cut total spend further than any discount recovered across the whole footprint.

What survives if we do not renew?

Under perpetual, the licences were owned and support was optional, so non renewal left you with running software. Under subscription, nothing survives non renewal. That difference in exit position is the part of the conversion that deserves the most attention before signing a long term.