AWS EDP Negotiation: The Ten Lever Playbook
An EDP is a multi year take or pay obligation, not a discount. The headline 3 to 40 percent is the easy negotiation. The Private Pricing Addendum is where the deal is won or lost.
Prepared by Redress Compliance · June 2026 · Representative AWS estate scenario (benchmark scenario, not a quote)
Executive Summary
An AWS Enterprise Discount Program, now papered as a Private Pricing Agreement, is the largest single contractual commitment most cloud mature enterprises sign. AWS presents it as a discount. Structurally it is a take or pay obligation against your cloud spend for one to five years, three being the common term.
The discount on the front page runs roughly 5 to 15 percent at $1M to $5M of commitment, 10 to 25 percent at $5M to $25M, and 20 to 40 percent above $25M. That number is the easy negotiation.
The hard negotiation is everything in the Private Pricing Addendum, the PPA: Marketplace inclusion, the net spend mechanic, true down rights, the growth allowance, and which services retire the commitment.
The expensive mistake is the overcommit. The commitment retires on your net bill after Savings Plans and Reserved Instances apply, so deep reservation coverage shrinks the very number that burns down your commit. A confident forecast plus aggressive optimization is the fastest route to an end of term true up.
This playbook gives you the ten buyer levers, the commitment math, and the contractual provisions that decide whether the EDP is leverage or liability. Read it once, hand the recommendations to FinOps and procurement, and use the timeline as the project plan.
Why is an EDP a commitment to spend, not a discount on services?
Read the contract title, not the slide. An EDP discounts your rates only if you reach a minimum committed spend over the term. Miss the number and you owe the gap anyway. That is the definition of take or pay, and it is the single fact the account team will soften in every conversation.
The discount is real, but it is the consequence of the commitment, not the substance of the deal. AWS does not publish its discount tiers, and account teams are instructed not to share benchmark data with customers. So buyers negotiate the one number AWS will discuss, the percentage, and ignore the number that actually carries the risk, the commit.
Three contract mechanics decide the outcome, and none of them sit on the front page:
| Mechanic | What it means | Why it carries the risk |
|---|---|---|
| Net spend basis | The commitment retires on your final bill after Savings Plans and RIs apply, not on list usage. | Optimizing harder shrinks the number that burns down the commit, pushing you toward shortfall. |
| End of term true up | Unconsumed commitment is billed as a lump at the end of the period. | A confident forecast can convert into a seven figure budget spike with no service in return. |
| Service eligibility | Not every line on your bill retires the commitment in the same way; Marketplace and some third party charges are capped or excluded. | You can hit your spend target and still miss your commitment if the spend lands in the wrong bucket. |
Why does the Private Pricing Addendum matter more than the EDP front page?
The front page shows a term, a commit, and a percentage. The Private Pricing Addendum shows whether that percentage survives contact with your actual consumption. The PPA is where AWS defines what counts, what is capped, what rolls over, and what happens when you are wrong about demand.
We read these addenda for a living, and the same provisions decide the same outcomes every time. The percentage varies by a few points between a well run and a poorly run negotiation. The PPA terms vary by millions.
| PPA provision | Default AWS posture | What a strong buyer secures |
|---|---|---|
| Marketplace eligibility | Capped at 25 percent of commitment, eligible SaaS only. | Explicit confirmation that planned SaaS qualifies, and the cap treated as a floor to use, not a ceiling to fear. |
| True down / reset | Not offered. Commitment is fixed for the term. | A mid term reset right tied to a defined business event or consumption variance. |
| Growth allowance | Overage billed at the discounted rate, no better. | Incremental discount on spend above the commit, so growth is rewarded not just tolerated. |
| Ramp profile | Flat or front loaded annual commit. | A ramp that tracks your real adoption curve, lowest in year one. |
| Service scope | Generic "AWS services" language. | Named inclusion of data transfer, egress, and AI services such as Bedrock in the discounted scope. |
Notice what is missing from the AWS default column: every provision that protects the buyer is one you have to ask for. Silence in the PPA is an AWS friendly term.
The absence of a true down right means the commitment is fixed. The absence of a growth clause means your success funds AWS at no extra benefit to you.
How do the EDP discount tiers scale, and where should you position your number?
The discount scales with total committed spend across the term, not annual run rate. Crossing a tier boundary is worth more than squeezing a few points inside a tier, so the highest leverage move is to position your commit just above the next break, never just below it.
| Total commitment band | Typical discount range | Negotiation note |
|---|---|---|
| $1M to $5M | 5 to 15 percent | Entry band. The percentage is modest; flexibility terms matter more than the rate here. |
| $5M to $25M | 10 to 25 percent | The crowded middle. Most enterprise EDPs land here; tier positioning pays. |
| $25M and above | 20 to 40 percent | Strategic band. AWS will invest in winning the workload; demand more than rate. |
Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025. AWS does not publish these tiers; ranges reflect deals we have advised and are confirmed against your estate during delivery.
Why is Marketplace EDP inclusion now table stakes, and what changed in 2025?
AWS Marketplace lets you route third party software purchases through your AWS bill, and that spend can retire your EDP commitment. For a SaaS heavy enterprise, this reshapes the math: software you were buying anyway helps clear a commit you were already making.
Two mechanics define the value, and both moved against the buyer in 2025. First, the May 1, 2025 SaaS policy change narrowed eligibility: only SaaS products fully deployed on AWS infrastructure now retire commitment. Before that, partial AWS deployment qualified. Invoices dated before May 1, 2025 were grandfathered.
Second, Marketplace retirement is generally capped at 25 percent of total commitment. That cap is not a problem to fear; it is a budget to use. A buyer who maps eligible SaaS to the cap before signing turns the commit from a spend obligation into a procurement consolidation.
| Marketplace rule | Before May 2025 | From May 1, 2025 |
|---|---|---|
| Deployment test | Partial AWS deployment qualified. | SaaS must be fully deployed on AWS to count. |
| Retirement share | 100 percent of eligible invoice, capped near 25 percent of commitment. | Unchanged 25 percent cap; eligibility pool is smaller. |
| Grandfathering | Not applicable. | Invoices before May 1, 2025 still count under the old test. |
How do Reserved Instances and Savings Plans interact with the EDP, and what is the dual counting question?
This is the provision that creates the most expensive surprises, and the one buyers understand least. The short answer: Savings Plans and Reserved Instances stack underneath the EDP, and the EDP commitment retires on the net result. The two discounts do not double count in your favor.
Mechanically, RI and Savings Plan pricing applies first, lowering the unit cost, and the EDP discount applies to the spend that remains. The upfront and recurring charges from RIs and Savings Plans do count toward your commitment, so they are not lost.
But because the commitment tracks net spend, every dollar you save with a reservation is a dollar that no longer burns down your commit.
Why this matters for your commit size
If you size the EDP commit against your current on demand bill, then deploy Savings Plans to optimize, you can cut your net bill by 30 to 40 percent and walk straight into a shortfall. The optimization that should save money instead triggers a true up.
The right sequencing
Model your reservation strategy first, project the net bill after RIs and Savings Plans, then size the EDP commit to that lower number with a defensible floor. Optimization and commitment are one decision, not two. Treating them separately is the most common seven figure error we repair.
Net bill reduction available from reservation coverage.
On the compute heavy estates we benchmarked in 2024 to 2025, full Savings Plan and RI coverage cut the net bill by 30 to 40 percent against on demand. Sized into the commit, that protects you; ignored, it creates a shortfall.
EDPs we reviewed were oversized at signature.
Roughly one in four EDPs we assessed before signature in 2024 to 2025 carried a commit set against the on demand bill, ignoring planned reservation savings, with overcommitment of 12 to 20 percent.
Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025. Confirmed against your estate during delivery.
Which flexibility provisions turn an EDP commit from risk into opportunity?
The difference between a good EDP and a dangerous one is not the discount. It is the set of provisions that let the commitment flex when your business does. AWS will not volunteer these. Each one is a lever you raise during negotiation, before the commit number is locked.
Below is a representative estate, Meridian Retail Group, a three year EDP with a $40M total commitment, ramped to match adoption. The numbers are a benchmark scenario, not a quote, and every row is internally consistent.
| Commitment year | Committed spend | Forecast net consumption | Position |
|---|---|---|---|
| Year 1 | $12.0M | $12.0M | On track, ramp matches adoption |
| Year 2 | $13.5M | $13.5M | On track |
| Year 3 | $14.5M | $14.5M | On track |
| Total | $40.0M | $40.0M | Right sized to forecast floor |
Now contrast the right sized deal with the commit AWS proposed first: a flat $16M per year, $48M total, sized to an optimistic growth story. If consumption lands at the $40M forecast, the buyer owes a $8M end of term true up for services never used.
The flexibility provisions that close that $8M gap are specific and nameable. Secure them in the PPA before the commit locks:
| Provision | What it protects | How AWS resists it |
|---|---|---|
| True down / reset right | Lets you reset the commit on a defined event or variance, killing the true up. | Claims it is not standard; concede only late, only with executive escalation. |
| Ramped commitment | Lowest commit in year one, rising with real adoption. | Prefers flat or front loaded so revenue books sooner. |
| Growth allowance | Extra discount on spend above the commit. | Offers the same discounted rate, no incremental benefit. |
| Marketplace flexibility | Maximizes the 25 percent SaaS retirement against the commit. | Leaves eligibility vague so less spend qualifies. |
What counter moves will AWS make, and how do you handle them?
The account team runs a consistent playbook. Recognizing each move removes its power, because most of them work only on a buyer who has not seen them before. Here are the counters we see in nearly every EDP cycle.
| AWS move | What they say | Your response |
|---|---|---|
| Anchor high on the commit | "To unlock the next tier you need to commit at $48M." | Separate tier from commit. Ask for the tier at your defensible number, then trade for it elsewhere. |
| Discount as the whole story | "We have moved you to 28 percent, this is a strong deal." | Accept the rate, then negotiate true down, growth, and Marketplace before signing anything. |
| Quarter end urgency | "This pricing is only available if we close this month." | Use their clock, do not serve it. Real concessions appear near AWS period ends, so let the pressure run both ways. |
| Forecast inflation | "Your AI and data growth means you will easily exceed this commit." | Commit to demonstrated consumption only. Keep speculative AI and analytics demand in flexible tranches. |
| True down is impossible | "We do not offer commitment resets, it is not how the program works." | Escalate. Resets exist; they are granted to buyers with a credible alternative and withheld from those without. |
Where the common advice on EDP commitments is wrong
The standard reseller and account team pitch is to commit big, because the top discount tier only opens at a high number and a larger commit nets a lower unit rate. We disagree.
On the EDPs we advised in 2024 to 2025, the net spend mechanic meant an aggressive commit paired with normal Savings Plan optimization produced a shortfall in roughly one in four cases.
The buyer side move is to size the commit to the floor of your defensible forecast, take a slightly lower headline tier if you must, and put the saved leverage into true down and growth provisions. A 28 percent discount on spend you never make beats nothing, but 22 percent on spend you actually consume beats it.
How do you use the mid term renegotiation moment?
An EDP is not a sign and forget contract. The midpoint of the term, usually around month 18 of a three year deal, is a live negotiation window that most buyers leave unused. AWS wants your renewal and your growth, and that want is leverage you hold before the term even ends.
Size and protect
Model reservations first, size the commit to the net forecast floor, and write true down, ramp, growth, and Marketplace flexibility into the PPA. Confirm eligible SaaS in writing. Keep a multicloud alternative credible.
Track and reopen
Monitor net burn down against commit monthly. If you are ahead of plan, reopen for a larger discount tied to growth. If behind, trigger the true down right early, before the shortfall compounds.
Time and reposition
Start the next negotiation six months before term end, timed to an AWS period close. Bring the alternative, the consumption record, and the unused Marketplace headroom to the table as leverage.
The buyer who treats the EDP as a living commitment, monitored and reopened, lands a materially better renewal than the buyer who signs and waits for the true up letter. Leverage is highest while AWS still wants something from you, which is always, as long as your workload can move.
What are the ten buyer levers, in order of impact?
Every section above reduces to ten moves. Run them in this order. The first four protect against the overcommit; the rest convert the commitment into leverage.
| # | Lever | Why it matters |
|---|---|---|
| 1 | Size to the net forecast floor | Defends against the true up; the single most valuable move. |
| 2 | Model reservations before the commit | Savings Plans and RIs shrink the net bill that retires the commit. |
| 3 | Win a true down / reset right | Turns a fixed obligation into one that flexes with the business. |
| 4 | Ramp the commitment | Matches the obligation to your adoption curve, lowest in year one. |
| 5 | Maximize Marketplace eligibility | Uses the 25 percent cap to retire commit with software you buy anyway. |
| 6 | Negotiate a growth allowance | Rewards spend above the commit with incremental discount. |
| 7 | Name the service scope | Pulls data transfer, egress, and Bedrock into the discounted scope. |
| 8 | Position just above the tier break | A tier crossing beats rate haggling inside a band. |
| 9 | Time signature to AWS period end | Real concessions surface near AWS quarter and year close. |
| 10 | Keep a multicloud BATNA credible | A movable workload is the only reason AWS grants flexibility. |
Recommendation
Negotiate the commitment, not the discount, and protect it with the PPA. The percentage on the front page is the part AWS will discuss freely, which is exactly why it matters least.
The deal is won in the addendum, where the net spend mechanic, the Marketplace cap, and the true down right decide whether your EDP is leverage or a liability with a deadline.
- Right size before you optimize. Model Savings Plans and RIs first, project the net bill, and size the commit to that floor. The commit retires on net spend, so an oversized commit plus normal optimization is the fastest path to a true up.
- Make the flexibility real. A true down right, a ramp, a growth allowance, and confirmed Marketplace eligibility move more value than any rate concession, and AWS grants them only to a buyer whose workload can credibly move.
Redress Compliance runs this playbook as a standing engagement: size the commit, draft the PPA asks, sit on your side of the table through signature and the mid term reopening. We are glad to tie a meaningful part of the fee to delivered value.
