VMware VCF pricing, what the per core rate really becomes
VMware Cloud Foundation lists at $350 to $400 per core per year on a one year term. Almost nobody pays that: term length, core scale, ramp structure, and a credible alternative stack into final prices 30 to 55 percent below list, while the 16 core minimum quietly pushes small host estates the other way.
Prepared by Redress Compliance · August 6, 2026 · Broadcom VMware advisory. Based on 35 to 50 renewals and migrations advised 2024 to 2026.
Executive summary
The baseline is the one year subscription at $350 to $400 per core per year, itself already down from the roughly $700 per core opening position Broadcom took after the acquisition in early 2024. Broadcom publishes no official rate card; every enterprise deal is negotiated, which makes the discount structure, not the list price, the real subject of this guide.
The bands are consistent across the market. Three year terms cut 18 to 28 percent. Five year terms cut 28 to 38 percent, the deepest standard band. Estates above roughly 10,000 cores unlock a further 5 to 12 points of scale discount, and ramp structures can price year one below steady state where migration timelines justify it. Stacked with competitive pressure, final prices across our engagements landed 30 to 55 percent below list, with large committed estates reaching effective rates near $100 to $130 per core.
The counterweight is the 16 core minimum per CPU socket. Every socket bills at least 16 cores whatever the silicon carries, and in estates full of small two socket hosts that minimum added 10 to 25 percent of phantom cores, paid for and never used. Host consolidation before the quote is therefore a pricing lever, not just an infrastructure project.
The single most valuable behavior in our engagement file was arriving with an alternative. First Broadcom quotes opened 2 to 4 times above the prior vSphere run rate, and a documented exit evaluation, even a partial one covering a workload slice, moved the final discount 8 to 15 points against estates that negotiated with no alternative on the table.
The price structure, from list to landed
Because Broadcom negotiates everything, VCF pricing behaves as a stack of discounts applied to the per core list. The bands are predictable enough to plan against:
| Lever | Typical effect | What earns it |
|---|---|---|
| One year term | List: $350 to $400 per core | Nothing. The baseline, and the walk away comparison for every other structure |
| Three year term | 18 to 28 percent off | The first meaningful commitment band, where most enterprise deals start |
| Five year term | 28 to 38 percent off | The deepest standard band. Only worth it with renewal caps and exit language attached |
| Scale above 10,000 cores | A further 5 to 12 points | Estate size, counted after the 16 core minimums inflate it |
| Ramp structure | Year one below steady state | A documented migration timeline that justifies deferred consumption |
| Competitive pressure | 8 to 15 further points | A documented exit evaluation the account team believes, even partial |
Stacked, the bands explain the market clearing range: committed multi year enterprise estates land at effective rates around $100 to $130 per core per year, while one year renewals without leverage pay near list. The spread between those two outcomes on a 20,000 core estate is roughly $5M a year, which is the honest budget for the negotiation effort.
The 16 core minimum, and the phantom core bill
VCF licenses per core with a minimum of 16 cores per CPU socket. On modern high core count silicon the minimum is irrelevant; on the long tail of small two socket hosts, 8 and 12 core parts running edge sites, management clusters, and legacy workloads, every socket rounds up, and the rounding is pure cost.
In our engagements the minimum added 10 to 25 percent of phantom cores in exactly the estates least able to absorb it. The response has two parts, and both belong before the quote, because the commit is sized on the inflated count otherwise:
- Consolidate the small hosts. Fewer sockets with more cores per socket converts phantom cores into real ones. A hardware refresh that looked marginal on its own math often clears easily once the licensing delta is counted.
- Scope the estate honestly. Hosts that will retire inside the term, edge sites moving to other platforms, and clusters slated for cloud migration belong in the ramp discussion, not in the committed core count.
The Broadcom VMware renewal survival guide
The full renewal sequence: quote deconstruction, the discount band benchmarks, the exit evaluation method that moves 8 to 15 points, and the clause set for multi year terms.
Get the white paper →Why the first quote opens at 2 to 4 times the old run rate
The opening quote is not a pricing error; it is the product transition doing its work. Perpetual vSphere with support converted into subscription VCF, which bundles the full stack, vSphere, vSAN, NSX, and the management layer, whether or not the estate uses all of it. The bundle repriced the estate at the same time as the metric changed, and the combination is what produces openings at 2 to 4 times the prior run rate. The VCF versus vSphere Foundation comparison covers when the smaller bundle honestly fits, which is the first deconstruction question on every quote.
The negotiation therefore starts with scoping, not percentage haggling: which hosts need the full VCF stack, which fit vSphere Foundation, what the honest core count is after consolidation, and what the term and ramp should look like against the migration reality. Each scoping answer shrinks the base the discount applies to, which is worth more than the discount itself in most estates.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across Broadcom renewals, 2024 to 2026
Across roughly 35 to 50 Broadcom VMware renewals and migrations Fredrik Filipsson advised between 2024 and 2026, final VCF prices landed 30 to 55 percent below list once term, scale, and competitive pressure stacked, and the spread between prepared and unprepared buyers was wider than in any other vendor practice:
A documented alternative evaluation, even partial, moved the final discount that far versus estates with nothing on the table.
First quotes against the prior vSphere run rate, before scoping and negotiation brought the base and the rate down.
The third pattern was the phantom core tax: the 16 core minimum adding 10 to 25 percent in small host estates, almost always discovered after the quote rather than before it. The buyers who did the consolidation math first negotiated from a smaller, honest base. The full licensing mechanics behind the pricing sit in the VCF licensing guide.
Your first five moves
- Deconstruct the quote into base and rate. Core count by host and socket, minimums identified, bundle tier per cluster. The base usually holds more recoverable money than the rate.
- Run the consolidation math on small hosts before committing, so phantom cores become real cores or leave the count.
- Scope VCF versus vSphere Foundation per cluster, and refuse the full stack where the smaller bundle honestly serves.
- Price an exit for a workload slice, documented, even if you intend to stay. It is worth 8 to 15 points and changes the meeting's tone.
- Trade term depth for protections: renewal caps, growth price holds, and ramp years matched to the migration plan. The Broadcom VMware practice and the assessment tools run the sequence with you.
Frequently asked questions
What does VMware Cloud Foundation cost per core in 2026?
List runs $350 to $400 per core per year on a one year subscription. Negotiated enterprise pricing lands well below: three year terms cut 18 to 28 percent, five year terms 28 to 38 percent, scale adds 5 to 12 points above 10,000 cores, and final prices across our engagements landed 30 to 55 percent below list, with large committed estates near $100 to $130 per core.
What is the VCF 16 core minimum and why does it matter?
Every CPU socket licenses at least 16 cores regardless of the physical core count, so small host estates pay for phantom cores: 10 to 25 percent of the licensed count in the estates we reviewed. Consolidating small hosts before the quote converts the phantom cores into real capacity or removes them from the bill.
Why is our Broadcom quote so much higher than what we paid VMware?
First quotes opened 2 to 4 times above the prior vSphere run rate across our engagements, driven by the shift from perpetual licenses with support to a subscription bundle carrying the full VCF stack. The response is deconstruction: bundle tier per cluster, honest core counts, term and ramp structure, then the discount conversation.
How much discount can we actually negotiate on VCF?
The standard bands are 18 to 28 percent at three years and 28 to 38 percent at five, with scale adding 5 to 12 points above 10,000 cores. A documented exit evaluation moved final discounts a further 8 to 15 points in our file. Stacked, prepared estates landed 30 to 55 percent below list.
Is a five year VCF term worth the deeper discount?
Only with protections attached: a renewal cap, price holds on growth cores, and portability language. The deep band locks the estate through Broadcom's next repricing cycle, so an unprotected five year discount trades a visible saving for an invisible future exposure. The clause set is worth more than the extra band depth.
Do we need full VCF or is vSphere Foundation enough?
Per cluster, not per company. VCF bundles vSphere, vSAN, NSX, and the management stack; clusters that use only compute virtualization often fit vSphere Foundation at a materially lower rate. Scoping the bundle tier honestly per cluster shrinks the base before any discount applies, which is usually worth more than the discount itself.
