VMware

What VCF to AWS native virtualization migration economics actually look like in 2026.

The pitch is a clean move from VCF to a native AWS posture. The economics are not clean. They run across a 36 to 54 month window, with a cumulative cost that often exceeds the VCF renewal it was meant to displace, even before the operating model shift is priced.

By The Negotiation Desk · VMware · Archetype: The Exit

Published 13 Mar 2026

The VCF to AWS native virtualization migration is the most asked about exit path the Desk fielded in the first five months of 2026. The question almost always arrives the same way. A buyer is reading a VCF renewal that has moved 28 to 41 percent above the prior contract on the per core basis, and the buyer's leadership is asking what it would cost to land the same workloads on AWS without a VMware presence. The pitch from the hyperscaler side is that the move is cheaper in the long run, simplifies the operating model, and removes a vendor concentration that has become uncomfortable. The pitch from the seller's side is that the move is more expensive than the buyer's quote, takes longer than the buyer's quote, and never finishes for the workloads that matter. Both sides are partly right. The piece below describes the actual cost shape across the cohort the Desk has helped scope.

The piece is not a recommendation. The Desk's view is that the migration economics are favourable for a meaningful subset of estates, neutral for another subset, and unfavourable for a third. Which bucket an estate sits in is a function of three variables that the buyer can measure before the migration starts. The piece walks through the variables, the cost components, the duration band, the realised savings, and the open ended cost lines that the pitch decks do not price.

The three variables that decide the bucket

The first variable is the workload composition. Estates dominated by stateless application tiers and modern data services migrate at the lower end of the cost band. Estates dominated by stateful workloads with custom storage profiles, large monolithic databases, or strict licensing on the guest operating system migrate at the higher end. The composition variable can swing the cumulative migration cost by 40 to 70 percent on the same core count.

The second variable is the network and identity dependency profile. Estates with deep dependencies on the VMware network virtualization stack, with custom NSX policies that have accreted across multiple years, and with identity flows that cross the VMware management plane carry a higher rework cost on the AWS side. The dependency variable can extend the migration duration by 30 to 50 percent and can produce a category of cost that is rarely priced in the opening business case.

The third variable is the buyer's operational readiness. Estates run by teams with active AWS engineering depth migrate faster and at a lower realised cost than estates where the AWS posture has to be built alongside the migration itself. The operational variable is the hardest to price honestly because the buyer's leadership often overstates the readiness on the way into the business case. The Desk's correction factor on the readiness variable, against the cohort the Desk has scoped, is a 25 to 40 percent increase on the realised migration duration over the buyer's stated duration.

"The migration is rarely a single move. It is a portfolio of moves across a multi year window, with parallel VCF licensing running across the window because the migration never finishes on its scheduled date for the workloads that matter."Migration Lead, The Desk

The cost components the business case has to price

The business case has to price six components. The first is the migration labour, which the Desk's cohort runs between $4.8M and $11.2M for an estate in the 10,000 to 15,000 core range. The variance is driven by the workload composition and the operational readiness variables. The second is the rework cost on the network and identity layer, which runs between $1.6M and $4.4M for the same estate range. The third is the new AWS run rate at the destination, which on the cohort runs at roughly 0.94 to 1.18 times the equivalent VCF run rate on the same workload footprint, with the variance driven by the workload composition and the use of reserved capacity on the AWS side. The fourth is the parallel VCF licensing cost across the migration window, which runs at roughly 35 to 65 percent of the steady state VCF run rate across the migration period because the seller will not reduce the licensed footprint linearly with the migration progress.

The fifth is the exit cost on the prior VCF contract, which is determined by the term structure at the point the migration decision is made. Buyers who decide to migrate in year one of a five year prepaid contract face a meaningfully larger exit cost than buyers who decide in year four. The sixth is the operational change cost on the buyer side, which prices the team restructuring, the tool change, and the runbook rebuild work that has to happen across the migration window. The sixth component is the one most often omitted from the opening business case, and the one most often responsible for the variance between the planned migration cost and the realised migration cost.

The duration band

The Desk's cohort across 11 in flight migrations runs at a duration of 36 to 54 months from the decision to migrate to the final decommission of the VCF estate. Migrations that complete inside 36 months are rare and almost always involve estates dominated by stateless workloads and teams with prior AWS depth. Migrations that extend beyond 54 months are common and typically involve a long tail of stateful workloads that the business case originally assumed would be migrated but that operationally remain on VCF because the migration economics for those individual workloads do not work. The long tail produces a permanent residual VCF footprint of 8 to 22 percent of the original estate, which the buyer must continue to licence at the renewal rate even after the migration has otherwise completed.

The realised savings on the cohort

The realised savings across the cohort, measured as the cumulative cost difference between the migrated state and a counterfactual VCF renewal across the same period, run between negative 4 percent and positive 27 percent. The median realised saving across the cohort is positive 11 percent of the counterfactual cumulative cost. The variance is driven primarily by the workload composition variable and secondarily by the operational readiness variable. The negative outliers are typically estates dominated by stateful workloads where the AWS run rate exceeded the equivalent VCF run rate and the migration labour cost was not recovered across the migration window. The positive outliers are typically estates that completed the migration early enough in the original VCF contract to avoid the residual licensing cost across the migration window.

In flight VCF to AWS migrations in the Desk's cohort11

Duration band from decision to decommission36 to 54 months

Migration labour on a 10,000 to 15,000 core estate$4.8M to $11.2M

AWS run rate against equivalent VCF run rate0.94 to 1.18 times

Median realised cumulative saving vs counterfactual VCF11%

Residual VCF footprint after migration8 to 22%

What the pitch decks rarely price

Two cost lines are rarely in the pitch decks. The first is the parallel VCF licensing across the migration window, which is invariably understated because the buyer's plan assumes a linear reduction in the VCF footprint that the seller will not contractually accept. The seller's renewal terms typically require a minimum committed core count across the term, which constrains the buyer's ability to reduce the licensed footprint until the contract ends. The second is the residual VCF footprint after the migration otherwise completes, which the pitch deck assumes is zero. The Desk's cohort runs the residual at 8 to 22 percent of the original estate, which means the buyer carries a permanent VCF renewal alongside the AWS posture rather than retiring the VMware contract entirely.

When the migration is the right move

The migration is the right move for estates where the workload composition is dominated by stateless tiers, the operational readiness on the AWS side is already in place, and the VCF contract term aligns to allow the migration to land before the residual licensing window opens. For estates that meet all three criteria the realised savings on the cohort have run between 18 and 27 percent of the counterfactual cumulative cost. For estates that meet two of the three the savings have run between 5 and 14 percent. For estates that meet one or none the migration economics have not produced savings against the counterfactual, even before the operational change cost is priced.

What we have seen on live deals this quarter

A global services firm scoped a VCF to AWS migration in February against an 18,000 core estate dominated by stateless application tiers, with an existing AWS engineering team carrying a significant production footprint already. The Desk's scoping concluded the migration would land at a realised saving of roughly 22 percent of the counterfactual VCF cumulative cost across the 48 month window, with a residual VCF footprint of roughly 9 percent. The firm proceeded with the migration. A regional insurer scoped the same move against an 11,200 core estate dominated by stateful database workloads with deep network virtualization dependencies. The Desk's scoping concluded the migration would land at a realised cost of roughly negative 6 percent against the counterfactual VCF cumulative cost. The insurer did not proceed with the migration and instead negotiated the VCF renewal down by 31 percent of the opening quote, which produced a better cumulative outcome than the migration would have.

The takeaway

The VCF to AWS native virtualization migration runs 36 to 54 months from decision to decommission and produces a median realised saving of 11 percent of the counterfactual VCF cumulative cost across the Desk's cohort. The variance across the cohort is wider than the median, and the bucket the estate sits in is decided before the migration starts.
The three variables that decide the bucket are workload composition, network and identity dependency profile, and operational readiness on the AWS side. The business case has to measure all three honestly before the migration is approved.
The pitch decks rarely price the parallel VCF licensing across the migration window or the 8 to 22 percent residual VCF footprint after the migration otherwise completes. Both lines are material, and both have to be in the business case before the migration economics can be read cleanly.