A discount you pay for whether you use it or not is not a discount
An AWS Enterprise Discount Program trades a multi-year minimum spend commitment for a discount across AWS usage. The discount rises with the size and length of the commit, but so does the shortfall risk if usage falls short, and an over-sized commit is the main failure mode, because you pay the committed amount whether you consume it or not. Size the commit to committed pipeline, not the growth plan, and the shortfall risk disappears.
Prepared by Redress Compliance · August 9, 2026 · AWS advisory. Based on roughly 25 to 35 AWS EDP negotiations advised 2024 to 2025.
Executive summary
An over-sized commit is the main failure mode, because you pay the committed amount whether you consume it or not.
An EDP is a private agreement where you commit to a minimum AWS spend over a multi-year term in exchange for a percentage discount applied broadly across eligible usage, and the three variables are linked: a larger commit and a longer term unlock a deeper discount.
But the commit becomes a floor you pay regardless of consumption.
Most EDPs we reviewed were committed on optimistic growth forecasts, leaving 10 to 25 percent at shortfall risk against realistic usage, and the deeper discount never offset the obligation to pay for usage that never arrived.
Size the commit at or just below your confident baseline, and bank growth as upside that improves the discount rather than spend you have promised to make.
AWS Marketplace spend can count toward the commit, and it is a major and often missed lever.
The scope of eligible spend is negotiated, not listed: direct compute, storage and database usage counts by default, but eligible AWS Marketplace purchases, including third-party software bought through private offers, can often count toward the commit, raising your effective consumption.
Estates that excluded eligible Marketplace spend left 5 to 15 percent of leverage on the table, because routing that spend through the EDP makes the same dollars work toward the floor you have to hit.
The EDP itself is not publicly documented, which is precisely why buyers need a benchmark, since the discount, the commit and what counts toward it are all negotiated.
Savings Plans and Reserved Instances stack on top of the EDP, so the EDP is not the only discount lever. Buyers who treated the EDP as the only lever under-used Savings Plans worth another 10 to 20 percent, layering compute Savings Plans over the EDP for compounding savings.
But Savings Plans and right-sizing reduce the very spend the commit relies on, so model their effect carefully: an optimization clawback that trims consumption after you have committed to a floor pushes you toward a shortfall.
The interaction runs both ways, so the buyer-side move is to stack the discounts while sizing the commit against the spend that survives the optimization, not the pre-optimization run rate.
Size to committed pipeline, because optimistic forecasts cause the shortfalls. The commit should sit at or just below your realistic baseline spend, built from workloads already running plus migrations under contract, with anything speculative in the upside, not the floor.
Three things break an optimistic commit: forecast bias, because growth plans rarely land on schedule; migration slippage, because delayed migrations push spend past the term boundary; and optimization clawback, because right-sizing and Savings Plans reduce the spend the commit relies on.
Pull twelve months of spend from Cost Management, strip out workloads ending within the term, model the floor, and negotiate ramp and true-up provisions that soften a missed quarter.
The EDP variables and their trade
| Variable | Direction | Benefit | Risk |
|---|---|---|---|
| Larger commit | Up | Deeper discount | Shortfall if usage falls |
| Longer term | Up | Better discount, price stability | Locks spend across years |
| Marketplace inclusion | Up | More spend counts toward commit | Requires eligible purchases |
| Conservative commit | Down | Low shortfall risk | Smaller discount |
The three variables are linked: a larger commit and a longer term unlock a deeper discount, and the risk is that the commit becomes a floor you pay regardless of consumption.
The commit is the minimum spend you promise over the term, the term is usually one to five years with longer terms trading for deeper discount, and the discount is a percentage off applied across eligible usage tracked in AWS Cost Management.
Understanding which line items are eligible is the difference between hitting the commit comfortably and facing a shortfall, and because the EDP is not publicly documented, the discount.
The commit and what counts toward it are all negotiated rather than listed, which is exactly why a credible spend benchmark is essential before signing.
The wider AWS negotiation context sits in the AWS contract negotiation guide, and the cloud commit mechanics in the cloud commit and egress benchmark.
Sizing the commit correctly
- Size at or just below your realistic baseline, not the optimistic forecast: a commit you are confident of hitting protects the discount without the shortfall risk, and growth is best treated as upside that improves the discount.
- Build from committed pipeline: base the commit on workloads already running plus migrations under contract, and put anything speculative in the upside, not the floor.
- Forecast bias is the first breaker: growth plans rarely land on schedule, leaving the commit unmet, which is why the floor is built from what is already running, not what is planned.
- Migration slippage is the second: delayed migrations push spend past the term boundary, so a commit sized on an on-time migration schedule shortfalls when the schedule slips.
- Optimization clawback is the third: right-sizing and Savings Plans reduce the very spend the commit relies on, so model their effect before committing, and negotiate ramp and true-up provisions that soften a missed quarter. The wider overspend picture sits in the AWS overspending report.
The AWS EDP negotiation guide
Commit sizing, discount tiers, what counts toward the commit, and the levers that cut an over-sized Enterprise Discount Program.
Get the white paper →What counts toward the commit, and how the discounts stack
The scope of eligible spend is a major lever, and direct service usage, compute, storage and database, counts by default, but eligible AWS Marketplace purchases, including third-party software bought through private offers, can often count toward the commit, raising your effective consumption.
Estates that excluded eligible Marketplace spend left 5 to 15 percent of leverage on the table, because routing that spend through the EDP makes the same dollars work toward the floor you have to hit.
So a benchmark that identifies which Marketplace lines are eligible is worth real money before signing.
Savings Plans are the other half of the picture: they stack on top of the EDP, so the EDP is not your only discount lever, and layering compute Savings Plans over the EDP compounds the savings by another 10 to 20 percent, but they reduce the spend the commit relies on.
So model how they affect the commit carefully.
That interaction is the crux of the sizing problem, because a buyer who commits to a floor on the pre-optimization run rate and then applies Savings Plans and right-sizing has engineered their own shortfall.
The buyer-side move is to stack the discounts while sizing the commit against the spend that survives the optimization, route eligible Marketplace spend through the EDP, and negotiate flexibility.
On an AWS EDP the deepest discount is worthless if the commit outruns your real consumption, so size to the floor and bank growth as upside. The token-economics context for the fast-growing AI line sits in the Bedrock pricing guide.
- 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
What we saw across AWS EDP engagements, 2024 to 2025
Across roughly 25 to 35 AWS Enterprise Discount Program negotiations Fredrik Filipsson advised on between 2024 and 2025, the recurring finding was that commits were sized on optimistic growth rather than committed pipeline, and the common advice drives that mistake.
The standard advice from AWS and many resellers is to commit big, because a larger commit unlocks a deeper discount and the growth will catch up. We disagree:
Of committed spend at shortfall risk against realistic usage, because the commit was sized on a growth plan rather than committed pipeline.
Average improvement in effective discount from right-sizing the commit, routing Marketplace spend through it, and stacking Savings Plans.
In roughly two-thirds of the EDPs we reviewed, the optimistic commit left 10 to 25 percent of spend at shortfall risk, and the deeper discount never offset the obligation to pay for usage that never arrived.
Three patterns recurred: commit over-sizing, with 10 to 25 percent of the committed spend at shortfall risk; the Marketplace blind spot, with estates that excluded eligible Marketplace spend leaving 5 to 15 percent of leverage on the table.
And stacked-discount confusion, with buyers treating the EDP as the only lever and under-using Savings Plans worth another 10 to 20 percent.
The buyer-side move is to size the commit to committed pipeline, route eligible Marketplace spend through the EDP to raise effective consumption, and treat growth as discount-improving upside, because a discount you pay for whether you use it or not is not a discount.
The preparation is modeling and scope: pull twelve months of spend from Cost Management, strip out workloads ending within the term, model the floor, list eligible Marketplace spend, model the effect of Savings Plans and right-sizing.
And negotiate ramp and flexibility provisions for a missed quarter, then take the baseline model and Marketplace plan into the negotiation, because AWS negotiates the commit and discount against credible numbers.
The wider library sits in the AWS practice.
Your first five moves
- Export twelve months of spend by service from Cost Management and separate committed baseline workloads from speculative growth, because the floor is built from what is already running.
- Size the commit at the confident baseline, treating growth as upside, because a discount you pay for whether you use it or not is not a discount.
- List eligible Marketplace spend that can count toward the commit, the major and frequently missed lever worth 5 to 15 percent of leverage.
- Model the effect of Savings Plans and right-sizing on the commit, and stack the discounts while sizing the floor against the spend that survives the optimization.
- Negotiate ramp and flexibility provisions for a missed quarter, and take the baseline model and Marketplace plan into the negotiation. The AWS practice runs the commit table with you.
Frequently asked questions
How does an AWS Enterprise Discount Program work in 2026?
An AWS EDP is a private multi-year agreement where you commit to a minimum spend in exchange for a percentage discount applied broadly across eligible AWS usage. The commit, term and discount are all negotiated rather than listed, so a credible spend benchmark is essential before signing.
The discount rises with the size and length of the commit, but the commit becomes a floor you pay regardless of consumption, which is why sizing it correctly matters more than chasing the deepest discount.
How big should my AWS EDP commit be?
Size the commit at or just below your confident baseline spend, not your optimistic growth forecast. A commit you are sure of hitting protects the discount without shortfall risk, while growth is best treated as upside that improves the discount rather than spend you have promised to make.
Build the floor from workloads already running plus migrations under contract, and put anything speculative in the upside, because forecast bias, migration slippage and optimization clawback all break an optimistic commit.
What happens if I miss my AWS EDP commit?
You generally pay the committed amount regardless of actual consumption, so a missed commit becomes a shortfall you pay for anyway.
This is why over-sizing the commit on optimistic forecasts is the main failure mode: the deeper discount that a larger commit unlocks never offsets the obligation to pay for usage that never arrived.
Building the commit from committed pipeline, and negotiating ramp and true-up provisions that soften a missed quarter, removes most of that risk.
Does AWS Marketplace spend count toward an EDP commit?
Often yes. Eligible AWS Marketplace purchases, including third-party software bought through private offers, can count toward the commit, which raises your effective consumption and makes the same dollars work toward the floor you have to hit.
Routing eligible Marketplace spend through the EDP is a major and frequently missed lever, and estates that excluded it left 5 to 15 percent of leverage on the table, so identifying the eligible Marketplace lines before signing is worth real money.
Should I commit big to get a deeper AWS discount?
Not on optimistic growth. A larger commit unlocks a deeper discount, but if usage falls short you still pay the commit, so the deeper discount never offsets the obligation. In roughly two-thirds of the EDPs we reviewed, the optimistic commit left 10 to 25 percent of spend at shortfall risk.
Sizing the commit to committed pipeline and banking growth as upside protects against that trap, because on an EDP the deepest discount is worthless if the commit outruns your real consumption.
Can I stack Savings Plans on top of an EDP?
Yes. Savings Plans and Reserved Instances apply on top of the EDP, so the EDP is not your only discount lever, and layering compute Savings Plans over the EDP compounds the savings by another 10 to 20 percent.
The catch is that Savings Plans and right-sizing reduce the very spend the commit relies on, so model how they affect the commit carefully: a buyer who sizes the floor on the pre-optimization run rate and then applies Savings Plans engineers their own shortfall.
Negotiating AWS 7: AWS's Playbook, and the Counters
Five moves AWS runs in almost every account: the forecast that is not yours, migration funding that buys your future baseline, the first offer, the scarcity story, and the renewal that reprices your own discipline.
