AWS Migration Cost Planning Guide: TCO, MAP Credits, and Negotiated Incentives
Most enterprise AWS migrations are budgeted from the wrong number. Finance starts with the AWS-supplied TCO calculator output, builds a business case around the operational savings, and signs a contract that bakes in run-rate spend two to three quarters before the first workload moves. By the time anyone looks at the all-in cost — including assessment, landing zone, refactor, parallel run, cutover, training, and the eighteen unforeseen line items in between — the project is already 30-50% over its original number and the procurement leverage that could have funded the overrun is gone.
This pillar guide is a comprehensive cost planning framework for AWS migration. It draws on patterns from $2.4B+ in AWS spend reviewed, 500+ engagements, and a documented average of 38% reduction against opening AWS pricing positions. The objective is not to oversell what AWS can do for a migration. The objective is to ensure the migration is funded by the right party — meaning that the buyer captures the migration incentives, EDP terms, and ramp protections that AWS reserves for sophisticated counterparties. Most buyers do not.
The Real Cost Structure of an AWS Migration
An AWS migration is not a single line item. It is six distinct cost categories that are almost never modeled together, which is precisely why budgets fail. The categories interact: choices in one phase determine the cost of the next phase, and ignoring those dependencies is the most expensive mistake in cloud cost planning.
1. Assessment and Discovery
Application portfolio discovery, dependency mapping, baseline TCO modeling, and migration wave planning. For a mid-sized enterprise (300-800 applications), this phase typically runs $400,000-$1.4M depending on the depth of the assessment and whether it is delivered by AWS Professional Services, a Premier Tier partner, or an independent advisor. Buyers who skip this phase or accept an AWS-funded discovery without independent validation almost always pay for it later in refactor scope creep.
2. Landing Zone and Foundation
AWS Organizations structure, Control Tower, network architecture, identity federation, security baseline, logging and monitoring foundation, and account vending. A defensible enterprise landing zone runs $300,000-$900,000 to design and stand up, plus ongoing operational cost. Cheap landing zones cost more later. The single largest source of post-migration cost surprises is a landing zone that did not anticipate the security, network, and compliance posture the workloads ultimately required.
3. Application Migration
The actual movement of workloads. Cost depends entirely on the migration pattern chosen for each workload: rehost (lift-and-shift), replatform (lift-and-tinker), refactor (re-architect), repurchase (move to SaaS), retire, or retain. Per-application cost ranges from $5,000 for a clean lift-and-shift of a stateless web tier to $300,000+ for a refactor of a database-heavy monolith. Across a portfolio, the blended cost typically runs $25,000-$80,000 per application, with refactor-heavy portfolios at the top of that range.
4. Parallel Run
The most underestimated cost in any migration. For the period between when a workload is deployed on AWS and when the source environment is decommissioned, you are paying for both. Parallel run periods typically last 3-9 months per workload and consume a meaningful share of the migration budget — often 8-15% of total project cost. This is also the period when negotiated AWS credits matter most, because they directly offset the duplicate spend.
5. Cutover and Stabilization
Final migration events, data sync, DNS cutover, on-call rotation, and the 30-90 day window of elevated incident rate as workloads stabilize on AWS. This phase rarely has its own line item, which is precisely why teams underestimate it. Engineering hours during cutover are 2-3x baseline. Budget for the burn rate explicitly.
6. Training, Tooling, and Operating Model
Cloud certifications, internal training, FinOps tooling, observability stack, and the operating model changes required to run cloud workloads with cloud economics. This is a recurring cost, not a one-time cost, and it is almost always omitted from the migration TCO. Treat it as a permanent 4-8% overhead on cloud run-rate.
How AWS Funds Migrations (and Why You Should Care)
AWS does not want you to budget the full migration cost out of your own balance sheet. AWS wants to fund a meaningful share of it through a portfolio of incentive programs — because every dollar AWS funds in migration is a dollar you commit in run-rate spend with terms that favor AWS. Understanding how the incentive portfolio works is the single highest-leverage piece of migration cost planning. The major mechanisms:
Migration Acceleration Program (MAP)
MAP is the headline program. It provides tiered funding for partner-led assessment and migration execution in exchange for committed AWS consumption. The structure is consistent across tiers: AWS funds a percentage of the migration work, the partner delivers it, and the buyer commits to a specified level of run-rate AWS spend over a defined period. Tiering historically runs from MAP for under $1M annual cloud spend through MAP+ and MAP 2.0 for tens of millions.
The trap: MAP credits are not free. They are paid for with EDP terms, commitment structures, and partner exclusivity provisions that often cost more than the credits themselves over the contract life. We cover the mechanics in detail in the Migration Acceleration Programme guide and the MAP credits negotiation playbook.
POC and Pilot Credits
Before MAP, AWS will typically fund a proof-of-concept or pilot phase to de-risk the decision to migrate. POC credits are essentially marketing spend for AWS and almost always available for workloads of meaningful size. The amount is negotiable and is typically pegged to the expected first-year consumption of the workload under evaluation.
EDP-Embedded Migration Discount
The Enterprise Discount Program itself can include a migration-specific ramp: a steeper discount in years one and two while spend ramps up, leveling off at the standard EDP rate in year three. This mechanic is rarely offered without specific negotiation and is one of the highest-value provisions in any EDP for a customer with active migration work. See our complete EDP negotiation guide for how to structure this.
Partner-Funded Incentives
Premier Tier and Advanced Tier AWS partners have their own incentive pools for funded engagements, lab work, and pre-sales. These are often separate from MAP and stack with it. Buyers who do not know to ask never see them.
The TCO Model That Actually Holds Up
Build the TCO model in three layers. AWS-supplied calculators show the operational savings layer. They do not show the migration cost layer or the negotiated economics layer. The full model is the sum of all three.
Layer 1: Operational Run-Rate
The base AWS spend that the migrated workload will generate. This is what AWS calculators output. Validate it independently — by your own engineering team modeling actual instance families, storage tiers, network egress, and reserved capacity assumptions. AWS calculator outputs are frequently 15-30% low because they assume optimal right-sizing the buyer rarely achieves in year one.
Layer 2: Migration Investment
The six cost categories above, summed across the entire portfolio with explicit wave-by-wave scheduling. Include the parallel-run dual cost explicitly. Most TCO models omit it, which makes the business case look better than it actually is, which is exactly why the project overruns.
Layer 3: Negotiated Economics
EDP discount, MAP credits, POC funding, partner incentives, and any negotiated ramp protection. This layer is what makes or breaks the business case. Buyers who model only Layers 1 and 2 routinely sign EDPs that leave 15-25% of the achievable value on the table — value that competitive procurement could have captured at no cost to the timeline.
The Wave Plan and Its Economic Consequences
Migration waves are usually scheduled around technical dependencies — which is correct from an engineering perspective and disastrous from a financial perspective. The economically optimal wave plan looks different from the technically optimal one in two important ways:
Front-Load the Spend Generators
AWS run-rate spend is what unlocks better EDP terms, MAP credit tiers, and ramp protections. Workloads that generate meaningful spend — large databases, compute-heavy services, data analytics platforms — should migrate early to establish the spend profile that supports the negotiation. Workloads that generate no spend but consume engineering effort should migrate later.
Cluster Refactors to Capture Architectural Savings
Refactor-heavy waves are expensive in one-time cost but generate the most architectural savings. Cluster them so that the operational savings start showing up in the same period that finance is looking at the migration overrun. This is a presentation issue as much as a financial one — and it makes the difference between a board narrative of "migration is paying back" and "migration is exceeding budget".
Stage the Parallel Run
The dual-cost period is the highest-cost period in any migration. Stage waves so that no more than 30-40% of total parallel-run cost is active in any single quarter. This smooths the cash impact and avoids the budget blowout that ends most migration projects in renegotiation with finance.
Negotiating the Migration Before You Start
The single largest determinant of migration cost is the contract that funds it. AWS will negotiate migration terms most aggressively when three conditions are true: the workload has not yet committed to a destination, a competing hyperscaler is in the room, and the buyer has a defensible spend forecast over three years. Lose any one of those conditions and the negotiation collapses to a discounted-credit pricing exercise with predictable AWS-favorable outcomes.
The Pre-Decision Window
The most negotiable moment in any migration is the period between business case approval and destination selection. During that window, MAP credits, EDP discounts, ramp protections, exit terms, and partner incentives are all on the table. Once a destination is selected and announced internally, every one of those levers gets weaker. Run the negotiation in the pre-decision window or accept that you will get a worse deal.
Competitive Tension as Cost Reduction
An Azure migration assessment, a Google Vertex equivalent, or even a credible on-prem refresh quote moves AWS pricing meaningfully. Buyers who run parallel hyperscaler bids during the migration planning phase consistently capture better EDP terms than buyers who do not. The cost of running the parallel bid — perhaps $50,000-$150,000 in advisor time — is dwarfed by the savings. See our multi-cloud leverage service overview for the operating model.
The Three Contract Documents to Negotiate Together
EDP, MAP funding agreement, and partner Statement of Work. These are usually negotiated sequentially and with different counterparts — which is exactly why AWS gets the best of each. Negotiate them in parallel, with the EDP terms anchored on the MAP credit value and the partner SOW anchored on the EDP ramp schedule. This is the single most consistent improvement available to migration buyers and almost nobody does it.
The Five Most Common Migration Cost Mistakes
Patterns from 500+ engagements. Every one of these is avoidable, and most of them are baked in before the first workload migrates.
- Budgeting from the AWS calculator. The AWS-supplied TCO calculator shows operational savings, not migration cost. Building a board-level business case from it guarantees overrun.
- Accepting MAP credits without modeling the cost of the commitment. MAP credits are paid for with EDP terms. Buyers who do not price the commitment over the contract life systematically overpay.
- Sequencing waves for engineering convenience. Migrate spend generators first, parallel-run smoothing second, dependency cleanup third. The engineering-optimal wave plan is the financially worst.
- Negotiating EDP after the destination is announced. Once leadership has committed publicly to AWS, the EDP discount tier drops by 8-15 percentage points relative to what was available in the pre-decision window.
- Omitting parallel run from the budget. The dual-cost period is the highest-cost phase of the migration. Buyers who do not budget it explicitly always go over.
The Workload-Specific Cost Patterns
Different workload types have different migration cost profiles. Knowing the profile in advance is the difference between an accurate forecast and a 30% overrun.
Database Migrations
The most expensive per-workload category, particularly for commercial databases. Oracle and SQL Server license portability rules add a contractual layer to the technical work. Database Migration Service (DMS) reduces the technical risk but not the licensing complexity. See DMS cost analysis and Oracle to AWS migration for the full mechanics.
VMware Estates
VMware Cloud on AWS is often presented as a low-effort migration path. It is — but the run-rate cost is meaningfully higher than EC2-native after migration. The economics work for short-duration migrations that will refactor within 18-24 months, and break down for buyers who treat VMware Cloud as a permanent destination. Detail in VMware Cloud on AWS pricing.
Mainframe Workloads
The highest-cost, highest-risk category. Successful mainframe migrations require specialist partners, multi-year timelines, and explicit AWS investment commitments. Buyers underestimate the operational continuity cost by orders of magnitude. Full treatment in mainframe migration cost analysis.
SAP Estates
SAP on AWS has a well-defined cost pattern with predictable HANA sizing, SAP-specific instance families, and certified architectures. The migration cost is dominated by SAP basis and integration work, not AWS infrastructure. See SAP on AWS migration cost guide.
Planning an AWS migration?
We build defensible three-layer TCO models, negotiate MAP credits and EDP terms in parallel, and capture the migration incentives most buyers leave on the table.
Contact Us →Frequently Asked Questions
How much does an AWS migration actually cost?
A typical enterprise migration runs 12-22% of three-year run-rate spend in one-time and transitional costs across assessment, landing zone, refactor, parallel run, and cutover. MAP credits can offset 15-25% of that when negotiated properly. AWS calculator outputs do not capture this.
What are MAP credits and how do they work?
Migration Acceleration Program credits fund partner-led assessment and refactor work in exchange for committed AWS consumption. The credits are tiered by workload size and can underwrite a meaningful share of the migration if structured before the EDP is signed.
When should migration cost negotiation happen?
Before any signed contract. Migration credits, EDP discount tiers, and ramp schedules are all most negotiable when the workload has not yet committed to AWS as the destination. The pre-decision window is the highest-leverage moment.
Can we run a competing Azure or GCP bid during migration planning?
You should. Documented parallel hyperscaler quotes move AWS pricing by 10-25 percentage points on EDP discount and 30-50% on MAP credit value. The cost of running the bid is dwarfed by the savings.
How long does a typical enterprise migration take?
Mid-sized enterprises (300-800 applications) typically plan 18-36 months for a complete migration. The economic mechanics matter most in the first 6 months, when contract terms are set.
The Bottom Line
An AWS migration is a capital project with operational savings. The capital cost is real, the operational savings are real, and the negotiated economics that connect them is where most of the value is captured or lost. Buyers who budget from the AWS calculator and accept the first MAP offer routinely pay 20-30% more than buyers who model all three layers and negotiate the contracts in parallel.
The discipline is procurement, not engineering. The leverage comes from competitive tension, defensible forecasts, and integrated EDP/MAP/SOW negotiation. None of those require a different engineering team — all of them require treating the migration as a procurement event that funds itself.
If you are budgeting a migration of meaningful scale and have not pressure-tested the TCO model or the incentive structure, the math is overwhelmingly in favor of doing so before the EDP is signed. Contact us for a migration cost review.