MAP-funded migration · 2026 walkthrough

How a MAP-funded AWS migration actually works, end to end (2026).

The AWS Migration Acceleration Program is the most-misunderstood line item in cloud migration. This is the operational walkthrough — the real timeline from discovery to optimize, exactly who does what at each phase (you, the partner, AWS), how the funding flows so a typical migration nets out near $0, and a realistic worked example. Not the brochure; the run-book.

phases
Assess → Mobilize → Migrate
typical duration
4–11 months
funding offset
up to ~100%
your cash cost
~$0
TL;DR
  • A MAP-funded migration runs in three formal phases — Assess (scope, business case, MAP record), Mobilize (landing zone, pilot, runbooks), and Migrate (production waves) — plus an informal Optimize tail. AWS funding is structured to track those phases: an assessment incentive up front, then partner-delivery credits and AWS service credits that scale with what you actually move.
  • The work splits three ways and the split is the whole point. You own decisions and access (priorities, sign-offs, IAM, the people who know the apps). The AWS partner owns delivery (the MAP paperwork, the landing zone, the wave execution, the cutovers). AWS owns the money and the guardrails (the funding pools, the Well-Architected bar, MAP eligibility). Done right, the partner-delivery funding plus the AWS credits offset most or all of the migration cost — so your out-of-pocket is near $0 even though real engineering happened.
  • Honest caveats: MAP has a floor (broadly ~$5K/month of post-migration AWS spend, often framed as a six-figure annual run-rate for the richer tiers), credits are consumption-based not a cheque, and the timeline is driven by your environment's complexity and your team's availability far more than by AWS. A clean 30-server lift-and-shift is months faster than a tangled re-platform — and the funding is sized to the destination spend, not to how hard the journey was.
orientation

IWhat MAP is — and what "MAP-funded" actually changes about a migration

The Migration Acceleration Program is AWS's structured, co-funded path for moving existing workloads onto AWS. Strip away the branding and it is two things bolted together: a methodology (a phased way to run the project) and a funding mechanism (money that offsets the cost of running it).

The methodology half is the three-phase model AWS has standardised across thousands of migrations: Assess, Mobilize, Migrate. It exists because migrations fail in predictable ways — teams skip discovery, under-build the landing zone, then try to move everything at once. The phases force discovery before delivery and a pilot before a production wave. Even unfunded, the structure is worth copying.

The funding half is what makes "MAP-funded" different from "a migration." AWS contributes in two distinct currencies, and conflating them is the single most common misunderstanding. The first currency is AWS service credits applied to your bill — these offset the cost of running the migrated workload during and after the move. The second is partner-delivery funding paid by AWS to the partner doing the work — this offsets the cost of the labour. You feel the first as a smaller invoice; you feel the second as a professional-services engagement that costs you far less than its sticker price, sometimes nothing.

That two-currency structure is why the headline "$0 migration" is real but needs an asterisk. Real engineers do real work for real weeks. The reason your cheque is near zero is that AWS is paying for the destination — they want your workload on AWS for the next decade, so they subsidise the move. The honest framing throughout this piece: the migration is funded, not free of effort. Your team still shows up.

One more orientation point. MAP is not a grant you "win" and then spend however you like. It is consumption- and milestone-based. The assessment incentive is earned by completing the assessment. The migration credits accrue against workloads you actually land on AWS. If the project stalls at 30% migrated, you get roughly 30% of the migration funding — not the full pool. This keeps everyone honest and is the reason partners care as much as you do about actually finishing.

the one-sentence version

MAP gives you a phased method (Assess → Mobilize → Migrate) plus two kinds of AWS money — service credits that shrink your bill and delivery funding that pays the partner — sized to your post-migration spend, so a qualifying migration can net out near $0 in cash even though the work is entirely real.

the end-to-end timeline

IIThe full timeline: discovery to optimize, with realistic durations

Here is the whole arc in order, with the durations a mid-sized migration (say, 40–120 workloads moving off a single source environment) actually sees in 2026. Treat the ranges as honest, not optimistic — the wide bands are where most of the variance lives.

The five stages below map onto MAP's three funded phases plus a pre-phase (Discovery, which precedes the MAP record) and a post-phase (Optimize, which is where the FinOps value compounds after cutover). Funding attaches to the middle three.

0. Discovery & MAP application — 1 to 4 weeks

What happens: A lightweight inventory of the source environment (how many servers, which databases, rough dependencies, current monthly spend), a first-cut business case, and submission of the MAP opportunity into AWS's partner systems. This precedes formal Assess funding — it is the qualification gate.

Who drives: The partner runs the application and writes the business case; you provide the inventory inputs and a sponsor who can say "yes, we intend to migrate." AWS reviews the opportunity for eligibility.

What can stall it: No clear sponsor, or a source environment nobody can describe. If you cannot answer "roughly how much do you spend now and on what," budget extra time here.

1. Assess — 2 to 6 weeks (funded by the assessment incentive)

What happens: Deep discovery. Automated tooling (AWS Application Discovery Service, Migration Evaluator, or partner equivalents) maps dependencies, right-sizes the target, and produces a TCO model and a wave plan. The output is a migration readiness assessment and a directional total-cost-of-ownership case that justifies the spend AWS is about to fund against.

Who drives: The partner runs the assessment and models the target architecture; you grant read access to the environment and validate that the discovered dependencies match reality. AWS funds this phase via the assessment incentive once the assessment is completed and submitted.

What can stall it: Access. Discovery tooling needs to see the environment; procurement or security review of that access is the usual bottleneck. Start the access conversation on day one of Assess.

2. Mobilize — 4 to 10 weeks (delivery funding + early service credits)

What happens: You build the foundation you migrate into. That means the landing zone (multi-account structure, networking, identity, guardrails), the operational runbooks, security baselines, a CI/CD path, and a pilot migration of one or two low-risk workloads to prove the pattern end to end. Nothing production-critical moves yet; you are de-risking.

Who drives: The partner builds the landing zone and runs the pilot; you make the architecture decisions that are actually business decisions (account boundaries, who owns what, compliance constraints) and sign off on the pilot. AWS funding shifts here from the assessment incentive to partner-delivery funding, and the first service credits begin to apply to the pilot workloads now running on AWS.

What can stall it: Treating the landing zone as a checkbox. An under-built landing zone is the number-one cause of painful Migrate phases — every shortcut here becomes a per-wave tax later. This is the phase to spend care, not speed.

3. Migrate — 2 to 6 months (delivery funding + the bulk of service credits)

What happens: Production workloads move in waves — batches of related applications cut over together, each wave a mini-project with its own test, cutover, and validation. Most migrations run 4–12 waves. The strategy per workload is decided in Assess: rehost (lift-and-shift), replatform (lift-and-reshape, e.g. self-managed database → RDS), or refactor (re-architect). Most 2026 migrations are mostly rehost with selective replatforming.

Who drives: The partner executes each wave and runs the cutovers; you own go/no-go on each wave, provide the application owners who validate that things work, and run user acceptance. AWS pays the largest share of both delivery funding and service credits across this phase, because this is where workloads actually land on AWS and start consuming.

What can stall it: Wave validation. The cutover is rarely the hard part; confirming the migrated app behaves correctly under real traffic is. Thin application ownership — nobody who can confidently say "yes this works" — is what turns a two-week wave into a six-week wave.

4. Optimize — ongoing (no MAP funding, but the largest long-run payoff)

What happens: After cutover, you right-size instances, adopt Savings Plans or Reserved Instances against the now-known baseline, retire over-provisioned resources, and apply Well-Architected fixes. This is where a migration stops being a cost and starts being a saving. MAP service credits are typically still burning down here, cushioning the optimisation period.

Who drives: Mostly you, optionally with the partner on a smaller FinOps engagement. AWS's formal funding role has ended — though Well-Architected reviews and other programs can fund follow-on work.

What can stall it: Nothing stalls it; it simply gets skipped. Teams declare victory at cutover and leave 20–40% of the savings on the table. The migration is finished; the value is not.

reading the durations honestly

These bands assume a single source environment and a team that can give the migration real attention. Multiply timelines for: multiple source clouds, heavy refactoring, strict change-freeze windows, or a part-time internal team. The funding does not speed up the calendar — it removes the cost barrier, not the complexity. A funded migration of a hard environment is still a hard migration.

the three-way split

IIIWho does what: you vs the partner vs AWS

The reason MAP migrations work — and the reason the ones that go badly went badly — comes down to the division of labour. Three parties, three clean lanes. When a lane gets crossed (you trying to do delivery, or the partner trying to make business calls), things slow down.

The simplest mental model: you own the decisions and the access; the partner owns the delivery; AWS owns the money and the guardrails. Each party is doing the thing it is uniquely positioned to do. You know your business and control your environment. The partner has done this dozens of times and holds the AWS funding paperwork. AWS holds the chequebook and sets the technical bar.

  • You (the migrating company) — Name a sponsor and a technical point-of-contact. Set priorities (what moves first, what can tolerate downtime). Grant access (read access for discovery, IAM roles for delivery). Provide application owners who can validate each wave. Make the business-level architecture decisions — account boundaries, compliance constraints, downtime tolerance. Sign off on go/no-go at each wave. Your total time is real but bounded: heavy in Discovery and at each cutover, light in between.
  • The AWS partner (the delivery arm) — Files and owns the MAP paperwork (opportunity, assessment, funding claims). Runs discovery tooling and builds the TCO model. Designs and builds the landing zone. Plans and executes the migration waves and cutovers. Manages the AWS relationship and the funding mechanics so you never touch a funding form. Carries the track record AWS underwrites against — which is why partner selection matters more than almost any other decision.
  • AWS (the funder and standard-setter) — Reviews MAP eligibility and approves the opportunity. Provides the assessment incentive, partner-delivery funding, and service credits. Sets the technical bar via the Well-Architected Framework — funding and the richer tiers expect the destination architecture to meet it. Provides solutions architects for design review on larger engagements. AWS does not run your migration; it underwrites and reviews it.
the most common lane-crossing mistake

Companies under-resource the one lane only they can fill: application owners for wave validation. The partner can move the bytes, but only your people can confirm the migrated app is correct. Migrations that staff this lane finish on the optimistic end of every range above; migrations that don't, drift to the pessimistic end.

follow the money

IVHow the funding actually flows to net out near $0

This is the section most guides skip, and it is the one that makes "$0 migration" believable instead of suspicious. There is no single cheque. The near-zero net is the sum of several flows arriving in different currencies at different phases. Trace them and the magic disappears — it is just structured incentives.

Recall the two currencies from Section I: service credits (smaller AWS bill) and partner-delivery funding (cheaper or free professional services). Add the assessment incentive at the front, and you have three flows. Here is the order they arrive and what each one offsets.

Stack the three flows and the picture is clear. The assessment is funded, the delivery is largely or fully funded, and the running cost is credited down. What remains for you in cash is, in a well-sized migration, close to nothing — and what you spend in effort is the part that was never going to be free: your team's time deciding, accessing, and validating. That is the honest shape of a "$0" MAP migration.

Flow 1 — the assessment incentive (front of the project)

AWS funds the Assess phase through an assessment incentive paid once the assessment is completed and submitted. Practically, this means the discovery, TCO modelling, and wave planning — work the partner does — is offset by AWS. For you, the assessment is typically at no cost. This is the first reason your cheque is small: the most analysis-heavy phase is funded before any workload moves.

Flow 2 — partner-delivery funding (Mobilize + Migrate)

During Mobilize and Migrate, AWS pays the partner delivery funding tied to the migration milestones — landing zone built, waves completed, workloads landed. This is the flow that offsets labour. Depending on the size of the migration and the funding tier, this can cover a large share of the partner's engagement cost — and in well-sized migrations, effectively all of it. This is why you can have a real partner doing real delivery without a large professional-services invoice.

The honest nuance: delivery funding is sized to the migration's scale and the post-migration run-rate. A small migration earns less delivery funding, so the partner may charge a residual fee; a large migration can be fully covered. "Near $0" is most literally true in the upper-mid range where the destination spend justifies full funding.

Flow 3 — AWS service credits (Migrate + Optimize tail)

As workloads land on AWS, service credits apply to your monthly bill, offsetting the cost of running the migrated estate during the move and for a window afterward. These credits accrue against actual consumption — they are not a lump sum you draw down freely. The effect: your AWS invoice during and just after migration is heavily discounted, cushioning the period before you have optimised and committed to Savings Plans.

Because credits are consumption-based, they reward actually finishing. A migration that lands 80% of its workloads earns far more service credit than one stalled at 30% — another structural reason every party is aligned on completion.

eligibility, honestly

VDoes your migration qualify? The thresholds and the honest disqualifiers

MAP is not for every migration. There is a floor, and there are profiles where the answer is "use a different path." Knowing this up front saves weeks of pursuing funding you won't receive.

The broad eligibility shape in 2026: an existing workload moving onto AWS from somewhere else (another cloud or on-premise), a credible business case, a delivery partner, and a post-migration AWS run-rate that clears MAP's floor — commonly framed around ~$5K/month of projected AWS spend at the entry end, with the richer funding tiers expecting a six-figure-plus annual run-rate. Below the floor, MAP's overhead isn't worth it and AWS won't size meaningful funding to it.

The thing to internalise: funding is sized to the destination, not the difficulty. Two migrations of identical pain — one landing at $4K/month, one at $40K/month — receive very different funding, because AWS is buying future consumption, not reimbursing your suffering. This is counter-intuitive and worth saying plainly: a hard, small migration may be barely funded; an easy, large one may be fully funded.

  • Strong fit — Moving 30+ workloads off-prem or off another cloud, projected AWS run-rate comfortably above the floor, a real sponsor, and a willingness to commit to AWS as the destination. This is the centre of the MAP target and where "near $0" is most literally achievable.
  • Workable fit — A smaller migration just above the floor (~$5K/month projected). MAP funding is thinner here, so expect the partner to charge a modest residual beyond what AWS funds. Still often far cheaper than an unfunded migration — just not always literally $0.
  • Wrong tool — too small — A handful of servers projecting well under $5K/month. MAP's structure costs more than it returns at this size. A straightforward partner-led or self-run migration (possibly with general Activate credits if you also qualify as a startup) fits better.
  • Wrong tool — no real migration — Net-new workloads built directly on AWS are not a "migration" — they belong to credit programs (Activate, POC funding, Well-Architected), not MAP. MAP requires an existing estate to move.
  • Disqualified — competitor or sanctioned region — AWS will not fund direct competitors, and credit/funding programs follow US export-control rules. These are hard stops independent of migration size.
the floor in plain terms

If your post-migration AWS bill will sit comfortably above ~$5K/month, MAP is worth pursuing and the upper tiers come into reach. If it will sit below that, treat MAP as unlikely and look at a lean partner-led migration or startup credit programs instead. The destination spend, not the migration's complexity, is the gate.

a realistic worked example

VIA realistic example: 60 workloads off a co-located data centre

Abstract walkthroughs hide where the time and money actually go. Here is a representative, anonymised-pattern migration — composite, not a single named customer — traced through every phase with rough durations and how the funding lands.

The setup. A mid-market SaaS company runs ~60 workloads in a co-located data centre: a mix of Linux web/app servers, a couple of self-managed PostgreSQL databases, a message queue, and assorted batch jobs. Current infrastructure spend is roughly $35K/month. They want out of the data centre before a lease renewal nine months away, and they project ~$28K/month on AWS post-migration after right-sizing. That destination run-rate puts them squarely in MAP territory.

Discovery & application (week 0–3). A partner inventories the estate, confirms the ~$28K/month projection is credible, writes the business case around the lease deadline, and submits the MAP opportunity. The sponsor is the VP of Engineering; the lease gives the project urgency, which helps. Funding outcome: opportunity approved, assessment incentive lined up.

Assess (week 3–8). Discovery tooling maps dependencies and surfaces two surprises — an undocumented internal service three other apps depend on, and a database far larger than anyone admitted. The partner builds the TCO model and a wave plan: 8 waves, mostly rehost, with the two PostgreSQL databases replatformed to Amazon RDS. Funding outcome: assessment completed and funded; the company's cash cost for this phase is effectively $0.

Mobilize (week 8–16). The partner builds a multi-account landing zone (separate prod/non-prod, centralised logging and identity, network baseline, guardrails) and runs a pilot: two low-risk batch jobs and one stateless web service. The pilot flushes out an IAM-permissions gap and a DNS-cutover wrinkle — exactly what a pilot is for. Funding outcome: partner-delivery funding begins; first service credits hit the pilot workloads. Company cash cost remains near $0.

Migrate (week 16–36, ~5 months). Eight production waves. Early waves (stateless services) move fast — about a week each. The database replatforming waves are the slow ones: data sync, validation, and a carefully scheduled cutover each take longer, pushing those waves to two to three weeks. The undocumented internal service found in Assess gets its own dedicated wave so its three dependents can be cut over together. Application owners validate each wave; one wave slips a week on validation, which is normal. Funding outcome: the bulk of delivery funding and service credits land here, tracking the workloads as they go live.

Optimize (week 36+). With the full estate on AWS and a known baseline, the team right-sizes over-provisioned instances, commits to Savings Plans against the steady-state, and retires a handful of resources that turned out to be dead. The $28K/month projection settles closer to $23K/month after optimisation. Residual MAP service credits cushion these first months. Funding outcome: no new MAP funding, but the largest recurring saving of the whole project is realised here.

where the money landed

Across the project: the assessment was funded, partner delivery was largely funded by AWS (the company paid a modest residual on the heavier database waves), and service credits offset the running cost through Migrate and into Optimize. Net cash cost to the company was a small fraction of an equivalent unfunded migration — and the recurring AWS bill, post-optimisation, came in below the old data-centre spend. The expensive part was never the cash; it was the engineering attention across ~9 months.

phase-by-phase reference

VIIThe phases side by side: owner, output, funding, and what stalls it

MAP-funded migration · phase-by-phase reference (2026)
PhaseTypical durationWho drivesKey outputAWS fundingWhat stalls it
0. Discovery & application1–4 weeksPartner + your sponsorInventory + business case + MAP opportunityQualification onlyNo sponsor; unknown environment
1. Assess2–6 weeksPartner (you grant access)Readiness assessment + TCO + wave planAssessment incentiveAccess / security review
2. Mobilize4–10 weeksPartner (you decide architecture)Landing zone + runbooks + pilotDelivery funding begins + first creditsUnder-built landing zone
3. Migrate2–6 monthsPartner (you own go/no-go)Production waves cut overBulk of delivery funding + service creditsThin wave validation
4. OptimizeongoingYou (optional partner FinOps)Right-sizing + Savings Plans + WAFR fixesNone (residual credits burn down)Getting skipped entirely
Durations assume a single source environment and a team that can give the migration real attention. The funded phases are 1–3; Discovery precedes funding and Optimize follows it. Funding is sized to your post-migration AWS run-rate, not to migration difficulty.
how to actually start

VIIIHow to start: the first two weeks, concretely

Knowing the arc is useless without a first move. Here is what the opening fortnight looks like if you decide to pursue a MAP-funded migration — deliberately small, because the whole point is to qualify before you commit.

Day 0 — assemble three facts. You need only three things to begin a serious conversation: roughly how much you spend on infrastructure now, roughly what you're running (count of servers, the main databases), and a named sponsor who can authorise the project. You do not need a detailed inventory yet — that is literally what the funded Assess phase produces.

Day 0–3 — get matched to a partner. The single highest-leverage decision is partner selection, because the partner carries the MAP paperwork, the funding relationship, and the delivery track record AWS underwrites against. You want a partner with a real migration history and the AWS tier that unlocks the funding tiers you need — not whoever is nearest. (This is the step CloudRoute handles: it routes your migration to a vetted, MAP-capable AWS partner matched to your stack, region, and size — see the box below.)

Day 3–7 — the qualification conversation. A 30–45 minute call where the partner pressure-tests fit: is the destination run-rate above the floor, is there a real migration (not net-new build), is there a sponsor and a deadline. You leave this call knowing whether MAP is realistic and roughly what funding tier you're looking at. If it's a no, you've spent an hour, not a quarter.

Day 7–14 — the MAP opportunity goes in. If it's a go, the partner drafts the business case and submits the MAP opportunity. Your job is to provide the inventory inputs and confirm the sponsor. From approval, the funded Assess phase begins — and from that point, your cash cost is effectively $0 while real work starts.

That is the entire on-ramp: three facts, one matched partner, one qualification call, one opportunity submission. The depth comes later, funded. The mistake to avoid is over-preparing — teams burn weeks building detailed inventories that the funded Assess phase would have produced for free.

funded vs unfunded

MAP-funded vs an unfunded migration — what actually differs

The phases and the engineering are broadly the same either way; a competent unfunded migration copies MAP's structure. What changes is who pays, how much scrutiny the design gets, and how the economics land. Here is the honest side-by-side.

DimensionUnfunded migrationMAP-funded migration
Assessment costYou pay for discovery + TCOFunded by the assessment incentive
Partner-delivery costFull professional-services invoiceLargely or fully offset by delivery funding
Running cost during moveFull AWS bill from day oneService credits offset consumption
Net cash costSix figures is commonNear $0 when well-sized
Architecture scrutinyWhatever you enforceWell-Architected bar expected
EligibilityAnyonePost-migration run-rate above the floor (~$5K/mo+)
Paperwork overheadNoneMAP records + funding claims (partner-handled)
Best forSmall estates below the MAP floorSubstantial migrations above the floor
The funding does not change the difficulty of the migration — it changes the bill. If your destination spend clears the floor, the funded path is almost always the right one: the same work, with the cash barrier removed and the architecture held to a higher standard.
not sure if you clear the MAP floor?
Get matched with a MAP-capable partner who checks eligibility for free
Start in 3 minutes →
a recent match

A MAP-funded data-centre exit — anonymised

inquiry · mid-market SaaS, data-centre exit, EU-Central
Mid-market SaaS, ~70 workloads in a leased data centre, ~$38K/month current spend, lease renewal forcing a decision

Situation: The team had the AWS skills to run the platform but no bandwidth to plan and execute a 70-workload migration alongside the product roadmap — and no appetite to absorb a six-figure professional-services bill or pay the data-centre bill and the AWS bill in parallel during the cutover. They needed the migration funded and delivered, not just advised on.

What CloudRoute did: Routed within a day to a MAP-capable AWS partner with data-centre-exit and RDS-replatform track record in the EU region. The partner filed the MAP opportunity, ran the funded assessment (8-wave plan, mostly rehost with two databases replatformed to RDS), built the landing zone, and executed the waves against the lease deadline. CloudRoute matched and handed off; the partner owned delivery and every funding claim end to end.

Outcome: Migration completed inside the lease window. Assessment funded; partner delivery largely covered by AWS delivery funding (a modest residual on the heavier database waves); service credits offset the running cost through cutover. Post-optimisation AWS spend landed below the old data-centre cost. The customer's out-of-pocket was a small fraction of an unfunded equivalent — CloudRoute's fee was paid by the partner, so the customer paid $0 to be matched.

engagement window: ~8 months · partner delivery: AWS-funded · customer cash to CloudRoute: $0

faq

Common questions

Is a MAP-funded migration really $0, or is that marketing?
It is genuinely near-$0 in cash for a well-sized migration — but never $0 in effort. AWS funds the assessment, offsets most or all of the partner-delivery cost, and credits down your running bill, so the cheque you write can be tiny or nothing. What is never free is your team's time: deciding priorities, granting access, and validating each migration wave. The migration is funded, not effortless. "Near $0 cash, real engineering attention" is the honest framing.
How long does a MAP-funded migration actually take?
For a single-source-environment migration of 40–120 workloads with a team that can give it real attention: roughly 4 to 11 months end to end. That breaks down as 1–4 weeks discovery, 2–6 weeks Assess, 4–10 weeks Mobilize, and 2–6 months Migrate, plus an open-ended Optimize tail. Multiple source clouds, heavy refactoring, change-freeze windows, or a part-time team push you toward and past the high end. The funding does not speed up the calendar — it removes the cost barrier, not the complexity.
What do I have to do versus the partner versus AWS?
You own decisions and access: name a sponsor, set priorities, grant the access discovery and delivery need, provide application owners to validate each wave, and sign go/no-go. The partner owns delivery: the MAP paperwork, the landing zone, the wave execution, the cutovers, and every funding claim. AWS owns the money and the guardrails: it approves eligibility, provides the funding, and sets the Well-Architected bar. The cleanest migrations keep these three lanes uncrossed.
Does my migration qualify for MAP?
Broadly, you need an existing workload moving onto AWS, a delivery partner, a credible business case, and a post-migration AWS run-rate above MAP's floor — commonly framed around ~$5K/month at the entry end, with richer tiers expecting a six-figure-plus annual run-rate. Crucially, funding is sized to your destination spend, not to how hard the migration is. Net-new builds (no existing estate to move) belong to credit programs like Activate or POC funding, not MAP.
How does the funding actually reach me?
Through three flows, not one cheque. First, an assessment incentive funds the Assess phase. Second, partner-delivery funding paid by AWS to the partner offsets the labour during Mobilize and Migrate. Third, AWS service credits apply to your bill as workloads land, offsetting the running cost. You feel the second as a cheaper-or-free professional-services engagement and the third as a smaller invoice. Stacked together, they net a qualifying migration out to near $0 in cash.
What are migration "waves" and why are they used?
A wave is a batch of related applications cut over together as a mini-project, with its own test, cutover, and validation. Most migrations run 4–12 waves. Waves exist because moving everything at once is how migrations fail — batching limits blast radius, lets each cutover be validated before the next, and keeps the project recoverable. The cutover itself is rarely the hard part; confirming each migrated app behaves correctly under real traffic is, which is why application-owner validation drives wave timelines.
What is the most common reason MAP migrations run late?
Two causes dominate. Early on, it is access — discovery tooling and delivery need to see and touch the environment, and security or procurement review of that access is the usual bottleneck; start it on day one. Later, it is thin wave validation — when there is no application owner who can confidently confirm a migrated app is correct, a two-week wave drifts to six. Notably, neither is an AWS-funding delay; both are inside your organisation, which is also where the fix lives.
What happens after the migration — does the funding continue?
Formal MAP funding ends at cutover, though residual service credits keep burning down into the Optimize phase and cushion it. Optimize is unfunded but is where the largest recurring saving appears: right-sizing, committing to Savings Plans against the now-known baseline, and retiring dead resources routinely recover 20–40% versus the un-optimised post-migration bill. Skipping it is the most common way teams leave money on the table. Separately, programs like Well-Architected reviews can fund follow-on improvement work.

Thinking about a MAP-funded migration?

CloudRoute routes you to a vetted, MAP-capable AWS partner matched to your stack, region, and size — the partner files the MAP paperwork, runs the funded assessment, and delivers the waves. Customer pays $0 to be matched. No procurement, no discovery theatre.

matched within< 24h
funding offsetup to ~100%
cost to be matched$0
How a MAP-funded AWS migration actually works (2026) — full walkthrough · CloudRoute