An acquired entity's AWS accounts do not inherit your private pricing rate, and without a pre-agreed step-down of 25 to 40 percent a divestiture leaves you paying commit on spend you no longer own
The AWS Private Pricing Addendum defines eligible accounts by who opened them, which means joining a target's accounts to your payer under AWS Organizations extends the payer relationship but not the discount. On the other side, a carve-out removes the spend but not the commitment, and AWS will only agree a 25 to 40 percent step-down if you negotiate it at signature rather than at the moment of the event. Both problems are solved with two paragraphs of redline that cost nothing at signing and are unbuyable afterward.
Prepared by Redress Compliance · August 31, 2026 · AWS advisory practice. EDP and PPA negotiations benchmarked 2024 to 2026, plus published addendum text from SEC and public sector filings.
Executive summary
The anti-pull-in clause is the single most expensive sentence in the addendum: accounts joined to an eligible payer via AWS Organizations that were not opened by you or your affiliates get zero of your 5 to 20 percent discount.
A buyer acquiring a $12M-a-year AWS estate and consolidating it under the existing payer will see the support bill rise immediately while the discounted rate stays behind, because eligibility is tested at account-opening, not at billing.
A divestiture removes the revenue but leaves the commitment, and AWS concedes a 25 to 40 percent step-down on defined triggers only when it is drafted at signature.
On a $30M annual commit, that is a $7.5M to $12M swing per contract year, and the identical request made after the carve-out is announced is priced as a renewal concession, not a contract right.
Joint and several liability survives the deal unless you sever it, which means a divested entity's accounts remain your exposure and your accounts remain theirs.
The published addendum text makes the customer and all affiliates owning eligible accounts jointly and severally liable for the entire commitment, so a carve-out without an account-transfer and release mechanic leaves two balance sheets on one shortfall.
The mis-scoped account clawback is a repayment event, not a billing correction, and it is enforced retroactively across the discount term.
AWS reserves the right to invoice the full value of discounting received on any account not owned by you or an affiliate, so a target's accounts run at your rate for nine months post-close becomes a lump-sum invoice rather than a rate adjustment.
How AWS decides which accounts get your rate, and why an acquisition fails the test
Four mechanics decide who gets your rate, and an acquired company fails on at least three of them the day the deal closes. The addendum defines an Affiliate by control, not by ownership percentage, so a majority-acquired target is technically inside the definition on close.
That is where the good news ends. Eligibility is separately tested at the account level: the discount applies only to accounts opened by you or your Affiliates, associated with a location matching the contracting AWS Party, and registered with email addresses issued by you or your Affiliates.
A target's accounts were opened years ago, by a company that was not your Affiliate at the time, under its own domain. They fail the origination test permanently, because you cannot retroactively change who opened an account.
Then comes the anti-pull-in clause, which explicitly says the addendum will not apply to an Eligible Payer Account joined via AWS Organizations to an account that is not an Eligible Account. That closes the obvious workaround.
And the clawback provision lets AWS invoice you the full value of discounting received on accounts not owned by you or your Affiliates, which converts a scoping error from a billing dispute into a repayment demand.
Add the representation that you have full power and authority to bind your Affiliates and that all affiliate accounts were opened by that Affiliate for its own use: a pre-existing acquired entity fails that warranty on its face.
And joint and several liability means you have signed up the target's balance sheet without the target having signed anything.
| Mechanic | Test applied | Where an acquisition breaks |
|---|---|---|
| Affiliate definition | Direct or indirect control | Passes on close for majority deals |
| Account eligibility | Who opened it, which domain, which region | Fails: opened pre-deal under target domain |
| Organizations join | Payer link does not confer the rate | Fails: consolidation moves billing, not discount |
| Clawback | Discount on non-owned accounts is repayable | Fails: mis-scoping becomes a repayment event |
| Affiliate binding rep | You warrant authority to bind | Fails: target never signed and may be jointly liable |
Read the table as a sequence, not a checklist. Control gets the target inside the Affiliate definition, which lulls buyers into thinking the work is done, and then the account-opening test quietly excludes every account the target already runs.
That gap between "is an Affiliate" and "has Eligible Accounts" is the whole problem, and it is only closable by an AWS-signed amendment that names the acquired accounts.
What an acquisition actually costs you before the amendment lands
The interim between close and amendment is where the money leaks, and it leaks in three directions at once.
First, Enterprise Support enrolls every linked account under the payer, priced at 10 percent of the first $150K of monthly spend, 7 percent to $500K, 5 percent to $1M, and 3 percent beyond, with a $15K floor.
Consolidate a target running $400K a month and your support bill rises immediately at your marginal band, while the target's spend earns no discount at all because its accounts are not Eligible Accounts. You have imported the cost and left the benefit behind.
Second, the anti-stacking prohibition kills whatever the target was already getting: its own credits, its MAP funding, a standalone PPA it negotiated last year.
The moment those accounts sit under your payer, the addendum's bar on combining discounting with any other discount gives AWS a clean argument to switch them off, and in practice AWS will not run two commercial constructs against one payer.
Third, the confidentiality clause covers the existence and terms of the addendum, so you cannot drop your PPA into a diligence data room without AWS consent. That is not academic.
If the target's advisors need to model post-close run rate, you are either seeking consent under time pressure or negotiating blind, which is precisely when AWS discovers it has leverage it did not have last quarter.
The asymmetry is deliberate: cost consolidates instantly through the payer, benefit consolidates only through a signed amendment on AWS timelines.
Assume 60 to 120 days of that gap based on what we see in practice, model it as a hard number in the deal case, and treat pre-agreed amendment mechanics as one of the clauses that decide your next three years rather than an operational detail for the FinOps team.
Protect your AWS EDP commit with 6 flexibility clauses
Six flexibility clauses protect an AWS EDP commit: rollover, carryforward, over commit caps, under commit relief, and clean exit ramps.
Get the white paper →The divestiture problem: your commit does not shrink when your estate does
The commitment is a payment obligation. It is not a usage forecast, and AWS will say so plainly the first time you ask for relief. Where buyers lose control is in a detail almost nobody diligences at signature: which shortfall trigger sits in the addendum.
Live AWS private pricing addenda contain two structures. The first tests Commitment-Eligible Fees in a Contract Year against the Contract Year Commitment. The second tests fees across the entire Discount Term against the Total Commitment.
Read as licensing mechanics, this looks like drafting preference. Read as leverage, it is the difference between a cash call inside 90 days of your carve-out closing and a deferred exposure you have two more years to consume against.
Model it against a real structure. Public AWS addenda run three Contract Years (the UK DVLA agreement, for example, runs a Discount Term from April 2024 to March 2027 split into three Contract Years), so assume a $30M annual commitment across three years, $90M total.
A Year 2 divestiture strips 30 percent of consumption. On a contract-year trigger, Year 2 lands at roughly $21M against a $30M Contract Year Commitment and AWS invoices a $9M shortfall payment at the year boundary, then does it again in Year 3.
On a total-commitment trigger, the same event produces a $9M gap in Year 2 that you can partially reabsorb through organic growth, migration acceleration, or a Year 3 overspend, with the reckoning arriving once at term end against $90M rather than twice against $30M.
That is a two-year timing swing on $18M of exposure, decided by which sentence your predecessor signed. Before you negotiate anything about corporate change, pull the addendum and find the trigger.
If it is contract-year, your first ask is not the step-down, it is a conversion to term-level measurement, which AWS concedes more readily because it does not reduce the number, only the cadence.
The step-down (25 to 40 percent commitment reduction on a qualifying divestiture, sized to divested run-rate) is the second ask, and as covered in our work on AWS private pricing agreement redlines, it is priced at zero at signature and effectively unbuyable once the event is public.
Analysis: AWS prices corporate change as an event, and your only defense is to price it as a term
AWS does not have a corporate-change policy. It has a corporate-change negotiation, and it runs that negotiation at the moment your leverage is at its structural minimum. This is not an accident of process.
It is the predictable output of an incentive system where the account team is measured on committed spend growth, and where every M&A event arrives on the customer's clock, not the vendor's.
Start with the timing asymmetry, because everything else follows from it. A signed acquisition or divestiture has a close date, a disclosed rationale, an integration plan with named owners, and usually a synergy number that a CFO has already put in front of a board. AWS has none of that.
It has a request.
When you approach the account team in the week after announcement asking to extend your rate to the target's accounts, or to reduce commitment because a business unit is leaving, you have told them three things at once: you have a deadline, you have a public commitment to a number.
And you cannot walk.
No competent seller trades hard-won contract value against a counterparty in that position, and AWS sellers are competent.
Now layer the incentive. An acquisition is not a problem for the account team, it is the best quarter they will have all year. New accounts, new workloads, an integration program that consumes services, and a customer who needs an amendment.
The response is entirely predictable: yes to extending the rate, priced as an uplift to Total Commitment, usually with term extension attached. A divestiture is the mirror image and it is a defend-the-number problem. The team's compensation and forecast both assume your commitment holds.
Their opening position will be that the commitment is a payment obligation independent of what you own, which is contractually correct, followed by an offer to keep the number whole in exchange for a longer term or a broader service scope.
That trade is worth taking only if you priced it before the event.
Which brings us to the observed pattern that matters most. Step-down language is conceded at signature and refused mid-term, and the reason is mechanical rather than cultural.
At signature AWS is competing for the commitment, and a conditional future reduction costs nothing against a deal it does not yet have. Mid-term it holds the commitment, and the same language is a direct write-down of contracted revenue in the current fiscal period. The clause has not changed.
The vendor's position on it has.
Buyers used to bridge a post-divestiture gap by routing third-party software through AWS Marketplace, and that route has narrowed.
The 25 percent marketplace contribution cap toward commit is a separate mechanism from the discount, and since the May 2025 tightening it applies only to products fully deployed on AWS, so anything with components running outside AWS infrastructure no longer retires commitment.
If your gap-closing plan assumed marketplace absorption at scale, requalify it against the current rule before you rely on it in a negotiation.
Here is the argument buyers underuse. A pre-agreed pull-in right, one that lets you bring an acquired entity's accounts under your rate on notice rather than by amendment, is an AWS revenue growth mechanism. It removes friction from exactly the scenario the account team wants.
That makes it tradeable, and it should be traded for the thing AWS resists: a symmetric step-down. Present them as one article, not two asks. You are not asking for relief, you are offering a faster path to incremental commitment in exchange for proportionality when the estate shrinks.
The conclusion is narrow and it is the only defensible position. Do not negotiate M&A twice.
Write a single corporate-change article that covers both directions, defines the affiliate control test as of the amendment date rather than the signature date, sets the step-down at 25 to 40 percent of divested run-rate, caps the pull-in uplift formula in advance.
And severs joint and several liability for divested affiliates at close.
Two paragraphs. Zero cost at signature. Unbuyable afterward.
The redline that pre-agrees both directions
Both problems, inbound and outbound, are solved by four paragraphs drafted before you sign.
Start by importing the assignment formulation AWS already publishes in its own Customer Agreement into the addendum itself: either party may unilaterally assign, in whole or in part, to any present or future Affiliate or to any entity which acquires the operating assets to fulfil its obligations.
That language is AWS's own drafting, which removes the argument that you are asking for something unusual. Without it, the addendum sits behind a consent gate at exactly the moment you have no time and no leverage, because your acquirer's closing checklist has a date on it and AWS knows it.
Second, add an automatic pull-in right: any entity in which you acquire a controlling interest during the Discount Term becomes an Eligible Account holder at the then-current discount rate.
Effective the later of the closing date or thirty days after written notice, with the acquired entity's spend counting toward the Contract Year Commitment from that date.
Draft around the account-origin problem directly, because the standard clause limits eligibility to accounts opened by you or your Affiliates, and an acquired estate fails that test on its face.
Say so in the amendment: the origin requirement is waived for accounts acquired through a change of control.
Third, the step-down.
Define the trigger (sale, spin-off, or transfer of a business unit representing more than an agreed percentage of your estate), define the baseline (trailing twelve months of Commitment-Eligible Fees attributable to the divested accounts, measured from Cost and Usage Reports).
And set the reduction at 25 to 40 percent of the remaining Contract Year Commitments, prorated.
Fourth, sever joint and several liability on transfer with an express release of the divested affiliate, and carve AWS pricing terms out of confidentiality for bona fide diligence under NDA.
The related work on AWS private pricing agreement clause redlines covers the surrounding protections that travel with these four.
The asymmetry to exploit: AWS reviews pull-in language as a growth clause and step-down language as a risk clause, and the two are cheapest when they arrive in the same redline.
Present them as a matched pair (you accept the commitment increase on acquisition, AWS accepts the reduction on divestiture) and the step-down reads as symmetry rather than as a hedge against your own forecast.
What the evidence shows about how AWS responds
Across benchmarked negotiations, reductions in this range are agreed pre-signature and effectively unavailable once the divestiture is announced.
AWS EDP discounts run 5 to 20 percent stepped by commitment threshold, so a stranded commit costs far more than the rate protects.
The evidence base here is published addendum text from SEC exhibits (the VTEX Form 20-F filing) and the UK public sector agreement executed under the One Government Value Agreement 2.0, read against 20 to 25 EDP and PPA negotiations benchmarked between 2024 and 2026 across that 5 to 20 percent band.
Four patterns recur with enough consistency to plan against. First, AWS concedes the pull-in right far more readily than the step-down, for the obvious reason that pull-in grows the commitment and step-down shrinks it.
Expect the pull-in paragraph to survive largely intact, sometimes with a notice period stretched from thirty to sixty days. Second, the step-down is a signature-window item only.
Raised mid-term, after a divestiture is public or even rumored, it is refused, and the standard response is an offer to roll the shortfall into an extended term rather than reduce it.
Third, AWS will trade the step-down for something: a term extension from three years to four or five, or a commitment floor below which the reduction cannot take you.
Both are usually worth accepting, since a floor at 60 to 75 percent of the original commitment still removes the majority of the stranded exposure. Fourth, the ineligible-services list stays unilateral.
The 30-day right to add a newly generally available service survives every redline we have seen, so build your forecast on services already in scope rather than on what AWS may launch.
On assignment, uncapped transfer rights are rarely accepted, but the operating-assets carve-out (AWS's own language) is accepted routinely because refusing it means arguing against a clause AWS publishes itself.
That is the highest-probability win in this section, and it should be the first item you table.
- 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
Your first five moves
- Pull the executed addendum and identify which shortfall trigger you signed, because a Contract Year test bills a divestiture gap inside twelve months while a Total Commitment test defers it to term end, and the two demand completely different negotiating postures (owner: contract manager, 48 hours).
- Model the commit gap at a 30 percent estate reduction before you talk to AWS, using your actual Contract Year Commitment: on a $20M year, losing 30 percent of spend leaves roughly $6M of commitment with no workload behind it, and that number, not a principle, is what buys you a step-down clause.
- Table the symmetric corporate-change article now, at renewal or mid-term amendment, not when the deal is signed, and demand both directions in one paragraph: acquired-entity accounts admitted to your rate within 60 days of close, and commitment reduced 25 to 40 percent proportionate to divested spend. Anchor it against the clause redlines that decide the next three years, since AWS concedes structural language far more readily inside a discount conversation than inside an M&A one.
- Price the step-down against a term extension rather than against discount, because AWS protects the 5 to 20 percent band harder than it protects duration. In our negotiation experience, a fourth year, or a higher committed floor in years one and two, buys a symmetric reduction right that AWS will not sell for basis points.
- Audit which accounts currently receiving your rate fail the account-opening test, and do it before AWS does, because out-of-scope accounts are not a billing correction, they are a repayment event for the full discount value received. Any account opened by an entity you later acquired, or registered with a non-corporate email domain, goes on a remediation list this quarter, with an amendment request attached rather than a hope that nobody reconciles it.
Frequently asked questions
Does an acquired company automatically get my AWS EDP discount?
No. The Private Pricing Addendum limits discounting to accounts opened by you or your affiliates and registered with your email domains, and it expressly excludes accounts joined to an eligible payer via AWS Organizations that were not themselves eligible.
Consolidating billing gets you one invoice, not one rate. You need a signed amendment adding the acquired entity's accounts as eligible accounts before any discount applies.
Can I assign my AWS PPA to an acquirer without AWS consent?
Under the standard baseline, neither party may assign without prior written consent, but the AWS Customer Agreement contains a carve-out permitting unilateral assignment to an affiliate or to an entity acquiring the operating assets used to fulfil the obligations.
That carve-out is not always mirrored in the private pricing addendum. Import it explicitly into the addendum so the discount travels with the business rather than dying at closing.
What is an AWS step-down clause and will AWS agree to one?
A step-down permits you to reduce the annual commitment on a defined trigger such as a divestiture, business unit closure or documented AWS service failure. Observed concessions run 25 to 40 percent of the annual commit.
AWS will not volunteer it, and it is materially easier to secure at signature or renewal than mid-term once a transaction has been announced.
What happens to my commit if I sell a division that generates a third of my AWS spend?
Nothing, unless the addendum says otherwise. The commitment is a payment obligation, not a usage forecast, so removing the spend leaves the shortfall payment intact.
On a $30M annual commit, a 30 percent estate reduction creates roughly a $9M gap per contract year, payable either in that contract year or at term end depending on which shortfall trigger your addendum uses.
Can I use AWS Marketplace to cover a post-divestiture shortfall?
Only partially. Marketplace purchases retire up to 25 percent of the commitment, and since a May 2025 tightening only SaaS products fully deployed on AWS qualify, so products with components running outside AWS no longer count.
That caps the bridge at a quarter of the commit and narrows the eligible catalog. Treat it as a partial mitigation, not a solution.
Are my affiliates liable for the AWS commitment after I divest them?
Yes, unless you sever it. The addendum makes the customer and every affiliate owning eligible accounts jointly and severally liable for all obligations.
A carve-out without an account transfer plus an express release leaves the divested entity exposed to your shortfall and leaves you exposed to theirs. Build the release and transfer mechanic into the transaction documents and get AWS countersignature.
Can I show my AWS PPA pricing to a potential acquirer during diligence?
Not without care. The addendum treats the existence and terms of the agreement as confidential, which is a live problem when a buyer's counsel expects to see committed spend obligations in the data room.
Negotiate an explicit carve-out permitting disclosure to bona fide transaction counterparties under a written NDA, or you will be seeking AWS consent under deal timelines.
