In this article

We break down how AWS’s Enterprise Discount Program, Azure’s Enterprise Agreements, and GCP’s Workload Agreements actually work, why the discount comes attached to a real financial commitment rather than just usage-based savings, and what that commitment structure costs an enterprise that outgrows or under-uses it.


Your AWS bill crossed a threshold where your account team started talking about an Enterprise Discount Program. The pitch sounds simple: commit to a spend level, get a meaningful discount. What doesn’t get explained as clearly is what you’re actually signing up for once that commitment is in place.

This isn’t a minor procurement detail. For any enterprise spending seven+ figures a year on cloud infrastructure, understanding what an EDP, Enterprise Agreement, or Workload Agreement actually locks in matters as much as the discount percentage itself.

What These Programs Actually Are

AWS’s Enterprise Discount Program, Microsoft’s Enterprise Agreements (sometimes referenced by the related Microsoft Azure Consumption Commitment, or MACC), and Google’s Workload Agreements all work on the same basic structure: an enterprise commits to a minimum spend level over a defined term, typically one to five years, in exchange for a discount percentage that scales with the size and length of the commitment.

AWS EDPs generally become available starting around $1 million in annual spend, with discount tiers commonly cited across multiple industry sources in the range of roughly 5% at entry level scaling to 25-30% or more for the largest, longest commitments, though AWS doesn’t publish these rates and actual terms are individually negotiated. Google’s Workload Agreements follow a similar tiered pattern, with steeper discounts available above $100 million in annual spend.

These agreements aren’t hypothetical or purely a sales pitch. Because they’re sometimes filed as material contract exhibits by public companies, real examples are occasionally visible on the public record. One such filing, an AWS Enterprise Discount Program Addendum between AWS and a publicly traded customer, filed with the SEC, shows a real contracted discount rate of 9% for the agreement’s first two contract years, a concrete illustration of how these terms are actually structured in a signed, legally binding document rather than a marketing estimate.

The terms of these agreements are negotiated privately between the enterprise and the cloud provider and generally aren’t published, which is itself part of what makes evaluating whether a given deal is actually good difficult without external benchmarking data or, as above, the rare public filing.

The Discount Comes With a Growth Floor, Not Just a Ceiling

The part of these agreements that gets less attention than the discount percentage is the commitment structure itself. Multiple independent sources covering AWS EDP terms confirm the same core mechanism: each year’s committed spend is generally required to be at or above the prior year’s, a built-in floor that only moves upward, regardless of whether your actual infrastructure needs grow at the same pace. If your usage plateaus or shrinks, whether from a business slowdown, a successful cost optimization project, or workloads moving elsewhere, the commitment doesn’t shrink with it, and by default, most agreements don’t include a step-down clause that would let it.

This creates a specific, real risk: shortfall liability. If actual usage comes in under the committed level, the enterprise typically still owes the difference. Analysis from contract-review firm Redress, based on a review of roughly 25 to 35 AWS EDP commitments, found that committed spend floors were frequently set 15 to 30% above conservative usage projections specifically to reach a deeper discount tier, and that shortfall liability ran 10 to 20% of total commitment value when growth didn’t keep pace with the ramp schedule negotiated upfront.

Other industry guidance describes the same failure mode from a different angle: AWS sales teams sometimes propose aggressive year-over-year ramp schedules that a customer’s actual usage growth doesn’t support, and by the time that gap becomes clear, typically in the later years of a three- or five-year term, AWS is generally unwilling to reduce the committed amount mid-term without a significant concession elsewhere. In other words, the discount is calibrated against an assumption of continued growth, and the cost of guessing wrong falls on the customer.

What This Means in Practice

None of this makes an EDP, Enterprise Agreement, or Workload Agreement a bad deal in every case. For an organization with genuinely predictable, growing cloud usage and the internal expertise to negotiate favorable terms, these programs can produce real, substantial savings, and walking away from a well-structured commitment purely on principle would leave money on the table. The issue isn’t that the discount is fake. It’s that the discount is priced against a specific bet about your future usage, and if that bet turns out wrong, the cost isn’t hypothetical.

This matters more, not less, as infrastructure strategy shifts. An enterprise locked into a multi-year, growth-floor commitment has a real financial disincentive to move workloads elsewhere, repatriate to private infrastructure, or right-size usage down, even when doing so would otherwise make sense, because doing so risks triggering the exact shortfall liability the commitment was structured to avoid. That’s a form of lock-in that doesn’t show up in a technical architecture diagram, but it’s just as real as the proprietary API dependencies we’ve written about separately in what your cloud API choice is actually costing you.

The Alternative Isn’t No Discount, It’s No Bet

Fixed, transparent pricing doesn’t require guessing your usage two to five years out to access a reasonable rate. OpenMetal’s dedicated hardware pricing is the same rate whether you’re evaluating month to month or committing to a multi-year term for additional savings, and multi-year terms lower the cost further without a growth floor attached, a shrinking workload doesn’t create a shortfall liability, because the price was never contingent on hitting an ever-increasing spend target in the first place. Our breakdown of the hidden costs in cloud computing covers the broader category of costs that don’t show up on the advertised price, including this kind of commitment structure.

Who Should Actually Look Closely at This

  • Worth a close look if you’re approaching an EDP, Enterprise Agreement, or Workload Agreement renewal and your usage growth has slowed, shifted, or become less predictable than when the original commitment was signed
  • Worth a close look if you’re being pitched a deeper discount tier that requires a significantly larger commitment than your current spend, since that gap is exactly where shortfall risk concentrates
  • Less urgent if your usage is genuinely predictable and growing at or above the committed pace, since the discount is doing what it’s designed to do in that scenario
  • Worth modeling either way what a shortfall would actually cost against what a fixed-price alternative would cost over the same term, rather than comparing only the headline discount percentages

Getting Started

Current bare metal and hosted private cloud pricing is fixed and published, not privately negotiated against a spend commitment. For a full total cost of ownership comparison against hyperscaler pricing, including multi-year term options, see our guide to calculating TCO for hosted private clouds.

FAQ

What is an AWS Enterprise Discount Program?

An AWS Enterprise Discount Program (EDP) is a private pricing agreement for enterprises with significant AWS spend, generally starting around $1 million annually, offering a discount on total billing in exchange for a committed minimum spend over a term typically ranging from one to five years.

What happens if my company doesn’t hit its EDP commitment?

Enterprises generally still owe the committed amount even if actual usage comes in lower, a structure known as shortfall liability. Industry analysis of AWS EDP commitments has found shortfall exposure commonly running 10 to 20% of total commitment value when usage growth doesn’t keep pace with the negotiated ramp schedule, and multiple independent sources describe AWS as generally unwilling to reduce a committed amount mid-term without a significant concession elsewhere.

Does the commitment level in an Enterprise Discount Program decrease if usage drops?

No. Standard terms typically require each year’s committed spend to be at or above the prior year’s level, meaning the commitment floor only moves upward regardless of whether actual usage grows at the same rate.

Is fixed, non-committed pricing available as an alternative to hyperscaler enterprise agreements?

Yes. Dedicated infrastructure providers like OpenMetal offer published, fixed pricing that doesn’t require a multi-year spend commitment to access competitive rates, with additional discounts available for longer terms but without a shortfall liability structure attached.


Chat With Our Team

We’re available to answer questions and provide information.

Reach Out

Schedule a Consultation

Get a deeper assessment and discuss your unique requirements.

Schedule Consultation

Try It Out

Take a peek under the hood of our cloud platform or launch a trial.

Trial Options

 

 

 Read More on the OpenMetal Blog

The Real Cost of AWS, Azure, and GCP Enterprise Discount Commitments

Aug 31, 2026

We break down how AWS’s Enterprise Discount Program, Azure’s Enterprise Agreements, and GCP’s Workload Agreements actually work, why the discount comes attached to a real financial commitment rather than just usage-based savings, and what that commitment structure costs an enterprise that outgrows or under-uses it.

When to Actually Store Everything in S3 and When Not To

Aug 24, 2026

We look at the real case for using S3-compatible object storage as a default backend, where that approach genuinely breaks down on latency, how serious platforms solve it with a caching layer rather than abandoning object storage entirely, and how to make the same call on a Ceph-based cluster.

Infrastructure for Internet-Wide Security Scanning and Attack Surface Management

Aug 17, 2026

We look at why attack surface management and cyber-risk scanning companies, businesses that scan large swaths of the public internet as their core product, run into trouble on hyperscaler infrastructure, what actually needs to be true about a provider’s acceptable use policy and IP allocation to support this workload, and how that differs from running an internal penetration testing lab.

The Real Cost Math Behind Self-Hosted GitHub Actions Runners

Aug 11, 2026

We work through the actual cost crossover between GitHub-hosted Actions runners and self-hosted runners on dedicated bare metal, using GitHub’s current 2026 rates, correct a common misconception about a self-hosted runner fee that never took effect, and cover what a self-hosted build pipeline needs beyond just cheaper compute.

Comparing OpenMetal, Hetzner, and OVHcloud for Proxmox VE Hosting

Aug 10, 2026

We compare three dedicated server providers commonly considered for Proxmox VE hosting, OpenMetal, Hetzner, and OVHcloud, across real hardware specs, current pricing, storage architecture, and support model, so you can match the provider to your actual workload rather than just the sticker price.

Migrating Off Azure: Entra ID and Cosmos DB Are the Hard Part

Jul 17, 2026

We look at why Azure’s egress fees are no longer the sharpest lock-in mechanism, and walk through the specific managed services (Azure Functions, Cosmos DB, Service Bus, Logic Apps, Azure AD B2C / Entra External ID) that actually make leaving Azure hard, including the one place Azure is more open than either AWS or GCP, and the one place it’s arguably worse.

Google Cloud’s Real Lock-In Lives in Spanner and Firestore, Not Egress Fees

Jul 15, 2026

We look at why Google Cloud’s egress fees are no longer the sharpest lock-in mechanism, and walk through the specific managed services (Cloud Functions, Firestore, Cloud Spanner, Pub/Sub, Cloud Workflows, Identity Platform) that actually make leaving Google Cloud hard, including where GCP’s lock-in profile is genuinely different from AWS’s.

The Real AWS Lock-In Is Managed Services, Not Egress

Jul 13, 2026

We look at why AWS egress fees are no longer the lock-in mechanism people think they are, and walk through the specific managed services (Lambda, DynamoDB, Step Functions, EventBridge, SQS/SNS, Cognito, API Gateway) that actually make leaving AWS hard.

How to Prevent Private Cloud Migration Delays

Jul 08, 2026

Planning a private cloud project? Organizing well from the start can prevent expensive and time-consuming delays. Our guide explains why hosted private cloud projects stall across migration and day two operations and shows how to prevent delays with better planning, architecture, ownership, and operational readiness.

Top 8 Reasons Companies Leave Public Cloud in 2026

Jul 06, 2026

A skimmable breakdown of the main business and technical drivers pushing companies from public cloud to hosted private cloud, covering cost control, compliance, performance, and operational control.