AWS MAP for VMware Migration: 3 Post-Broadcom Paths (2026)

Path 1: VMware Cloud on AWS (VMC) Deep Dive

VMware Cloud on AWS runs VMware’s full Software-Defined Data Center stack — vSphere, vSAN, NSX, and HCX — on dedicated AWS i3en or i4i bare-metal hosts. The control plane is managed by VMware. The hardware lives in AWS. From your administrator’s perspective, it’s the same vCenter they used on-premises, just running in us-west-2.

Migration uses VMware HCX, which supports vMotion across the WAN. VMs move live with sub-second cutover and no IP changes. Production workloads can migrate during business hours. Existing automation (vRealize, Ansible, Terraform vSphere provider) continues working unchanged. Backup tooling (Veeam, Commvault, Rubrik) integrates the same way it did on-premises.

MAP credits apply to the AWS infrastructure layer — the bare-metal host hours, EBS storage, data transfer, and adjacent AWS services consumed by VMC workloads. VMware’s software subscription, billed through Broadcom or AWS Marketplace, is not MAP-eligible. Practical implication: VMC offsets your AWS spend but does not reduce your VMware spend. Customers using VMC for permanent residency rather than as a transit path generally see lower net savings than rehost or refactor options.

Where VMC genuinely wins: timeboxed datacenter exits. If your colo contract expires in six months and you have 800 VMs to move, VMC compresses the move to weeks rather than quarters. Many customers then use VMC as a staging environment, gradually peeling workloads off into native EC2 or managed services over 12-24 months — the "Relocate then Refactor" arc.

Path 2: Rehost to EC2 Deep Dive

Rehosting moves VMware VMs to native EC2 instances using AWS Application Migration Service (AWS MGN). The result is identical operating systems and applications running on AWS compute, with VMware removed from the stack. This is the path most VMware exits ultimately take because it delivers the largest licensing savings with manageable risk.

AWS MGN deploys a lightweight agent inside each source VM (Windows or Linux). The agent performs continuous block-level replication to a staging area in your AWS account, typically maintaining sub-5-second RPO once steady state is reached. When you trigger cutover, MGN snapshots the latest state, launches an EC2 instance from the snapshot, runs post-launch scripts to install drivers and reconfigure network settings, and validates connectivity. Cutover windows are typically 15-30 minutes per VM. Agentless replication via vCenter integration is available for environments with strict no-agent policies.

Forrester’s 2024 Total Economic Impact study on AWS MGN documented $3.5M in three-year savings for a composite organization migrating 500 VMs, with migrations completing 70% faster than manual lift methods. The same study found that 60% of total savings came from post-migration rightsizing rather than licensing elimination alone — VMware environments routinely run at 20-30% average utilization due to host pinning, while EC2 instances can be sized to actual P95 demand.

For a deeper technical walkthrough of MGN agents, replication topology, and wave orchestration, see our complete AWS migration guide and the companion tooling reference at essential AWS migration tools.

Path 3: Refactor to Cloud-Native Deep Dive

Refactoring replaces VMware-hosted applications with managed AWS equivalents. Done selectively, it delivers the highest steady-state savings of any path. Done indiscriminately, it explodes timelines and budgets. The discipline is knowing which workloads earn the refactor investment and which should stay on rehost.

Strong refactor candidates have one or more of these traits: a database backend portable to Amazon RDS or Aurora; a containerized or containerizable application that fits ECS or EKS; statefulness that maps cleanly to S3, DynamoDB, or ElastiCache; or a batch processing pattern suited to AWS Batch or Lambda. Self-hosted MySQL, PostgreSQL, and SQL Server instances are near-universal refactor wins — MAP credits cover RDS consumption from day one, and managed backups, patching, and Multi-AZ failover remove three operational burdens.

Containerized workloads already running on VMware Tanzu, OpenShift, or vanilla Kubernetes move to EKS with moderate effort: registry migration via ECR, ingress remapping, IAM Roles for Service Accounts, and observability stack replacement. Workloads orchestrated by vSphere with Tanzu translate particularly cleanly because the underlying CRDs are upstream Kubernetes.

Avoid refactoring legacy applications with hard-coded dependencies on Windows Server features, kernel modules, proprietary middleware, or specific network topologies. These are rehost candidates — run them unchanged on EC2 and revisit modernization once they are no longer firefighting on Broadcom pricing. Equally, do not refactor workloads scheduled for retirement within 18 months; the engineering investment will not pay back.

How AWS MAP Credits Cover VMware Migrations

MAP credits for VMware engagements come in two tranches: assessment funding (typically $50K-$250K depending on estate size) and migration credits (a percentage of forecasted AWS consumption, usually 15-25% over the first two years). Credits draw down quarterly against actual AWS spend on migrated workloads, so accurate forecasting during the assessment phase materially affects total funding.

Eligibility specifics for VMware: a documented VMware estate of any size qualifies for the assessment tier; a forecast minimum AWS commitment (currently $250K annual run rate for the standard MAP tier, $1M for MAP for Windows Workloads expansions) is required for migration credits. The forecast must include EC2, EBS, data transfer, and any target managed services. VMware Cloud on AWS bare-metal host consumption counts toward the AWS commitment; the VMware software fees do not.

For a complete walkthrough of credit calculation, drawdown mechanics, and application-to-approval timelines, see our CFO guide to maximizing AWS MAP credits. For sister estates with significant Windows Server licensing, also review AWS MAP for Windows workloads — Microsoft and VMware credits can stack within a single engagement when both estates exist. The KB explainer at what are AWS MAP credits covers the foundational concepts in one page.

Migration Timeline: What a Typical VMware MAP Engagement Looks Like

VMware MAP engagements for estates of 200-500 servers run 6-12 months end to end. Smaller estates (under 100 VMs) often complete in 16 weeks. Larger estates (over 1,000 VMs, complex dependencies, regulated industries) extend to 18-24 months. The waved approach below is the canonical pattern and applies regardless of which path mix you choose.

1. Weeks 1-4 — Assessment. Deploy AWS Application Discovery Service or Migration Evaluator. Inventory VMs, applications, dependencies, and runtime characteristics. Categorize each workload into one of the three paths. Output: a wave plan, a refined AWS spend forecast, and a signed MAP application. Funded by MAP assessment credits.
2. Weeks 5-8 — Mobilization. Build the AWS landing zone (multi-account via Control Tower), Direct Connect or VPN to on-premises, replication infrastructure (MGN staging area, HCX appliance for VMC), backup, monitoring, and IAM. Pilot 5-10 non-production VMs end to end. Resolve any tooling gaps before production waves start.
3. Weeks 9-24 — Migration. Run waves of 20-50 VMs every two weeks. Each wave: replicate, test in isolated VPC, schedule cutover, execute cutover, validate, decommission source. Decommission cadence is financially important — every shut-down VMware host reduces your licensing obligation.
4. Weeks 25-40 — Optimization. Rightsize EC2 instances based on 30 days of CloudWatch data. Apply Savings Plans or Reserved Instances. Implement FinOps controls so the projected savings actually materialize.

Wave sequencing should prioritize early host decommissions. Migrating ten low-traffic VMs from one VMware host first lets you turn off that host in week six and start reducing licensing exposure immediately — a meaningfully better financial outcome than migrating ten high-value VMs spread across ten different hosts.

Frequently Asked Questions

How much do MAP credits actually cover for a VMware migration?

For standard-tier MAP engagements with at least $250K projected annual AWS run rate, credits typically equal 15-25% of forecast AWS spend over the first 24 months. A 500-VM estate with a $1.5M annual forecast might receive $225K-$375K in migration credits plus $50K-$150K in assessment funding. The exact percentage depends on workload mix, partner involvement, and AWS account team negotiation.

Can I use MAP credits to fund VMware Cloud on AWS subscription fees?

No. MAP credits apply only to AWS consumption. VMware subscription fees on VMware Cloud on AWS — billed through Broadcom or AWS Marketplace — are not credit-funded. This is the structural reason VMC alone does not solve the Broadcom pricing problem and most customers use VMC as a transit path rather than a permanent destination.

What is the realistic timeline to exit VMware completely?

For estates of 200-500 servers using a mostly-rehost strategy, 6-12 months from MAP signing to final VMware host decommission is typical. Larger estates (1,000+ VMs) or refactor-heavy strategies extend to 18-24 months. The fastest published case studies show 90-day exits for sub-100-VM estates using VMC as an intermediate landing zone.

Do I need a MAP-approved partner or can I run the migration in-house?

You can self-deliver, but most enterprises engage a MAP Migration Competency partner for two reasons: partners typically secure higher credit allocations through joint funding mechanisms, and partner expertise compresses the assessment-to-cutover timeline by 30-50%. Self-delivery is more common for organizations with prior large-scale AWS migration experience.

What happens to my Broadcom contract if I exit early?

It depends on contract language. Many post-acquisition Broadcom subscriptions are multi-year with no early termination clause — meaning you pay through the term regardless of usage. Some legacy VMware perpetual licenses with separate maintenance contracts allow maintenance cancellation. Audit your contracts before signing a migration plan; accelerating migration to align with renewal date often provides a cleaner exit than mid-term termination.

Next Steps: Free VMware Estate Assessment

Opsio has delivered 25+ AWS MAP engagements covering 3,000+ VMware VMs across Nordic, DACH, and UK enterprises. We provide a complimentary VMware estate assessment that quantifies your Broadcom exposure, models the cost of each migration path, and feeds directly into a MAP-maximized funding application — typically completed in three weeks.

Our wave planning methodology specifically targets early VMware host decommissions, helping clients eliminate 40-60% of VMware licensing spend within the first 90 days of migration execution. As an AWS Migration Competency partner with VMware Certified Professionals on staff alongside AWS Solutions Architects, we run both the technical execution and the financial modeling in a single engagement.

Start with our AWS migration services team for the free VMware assessment, or explore our broader AWS managed services if you need 24/7 operations after cutover. Cited sources: Gartner "Cost Optimization Strategies for VMware Customers After Broadcom Acquisition" (August 2024); Forrester Total Economic Impact of AWS Application Migration Service (2024); Forrester Q1 2025 VMware Migration Sentiment Survey; AWS re:Invent 2024 sessions MAM218 and ENT306.