On this page
AWS#
These ranges reflect AWS Enterprise Discount Program and Private Pricing Agreement contracts benchmarked by our team across 2024 to 2026. "Effective discount" combines EDP headline + PPA service-level + commitment-layer savings.
| Annual Commitment | EDP Typical | Achievable With Leverage | Notes |
|---|---|---|---|
| $500K to $1M | 3 to 8% | 8 to 14% | Minimum EDP threshold ($1M typical); RI/SP layering is the dominant lever. |
| $1M to $5M | 8 to 14% | 14 to 22% | Strategic-account engagement; PPA becomes accessible with Azure RFP. |
| $5M to $15M | 12 to 20% | 22 to 32% | Sweet spot, full strategic-account team, PPA on EC2/S3/egress unlocks. |
| $15M to $50M | 18 to 26% | 28 to 38% | Executive relationship; multi-service PPA + 5-year commitment + SI partnership. |
| $50M+ annual | 22 to 32% | 34 to 45% | Top-tier strategic; custom MSA, egress carve-outs, committed-use services. |
Headline EDP discount is only part of the economics. A 20% EDP with 100% MACC, uncapped egress, and standard rate-card exposure is economically worse than a 14% EDP with 85% MACC, PPA-discounted egress, CPU/GPU commitment carve-outs, and rate-card protection. Across a 3-year term, the gap commonly exceeds 15 to 20% of total cloud spend.
EDP discounts scale with commitment size. A 1-year $1M commitment typically earns 5 to 10%; a 3-year $5M commitment earns 13 to 18%; a 3-year $25M+ commitment earns 20 to 28%; and a 5-year $50M+ commitment with Azure or GCP competitive pressure reaches 25 to 35%. Discounts apply on top of RI/Savings Plan rates and can be combined with PPA custom pricing.
A PPA is a custom discount layered onto specific services, typically EC2, S3, data transfer, RDS, negotiated based on committed usage at the service level. PPAs can produce another 10 to 25% on top of EDP discount for strategic accounts, but they require a credible alternative deployment (Azure, GCP, or on-prem) to unlock. PPAs are AWS's answer when a customer threatens to migrate workloads.
Layer all three. Compute Savings Plans (up to 66% off on-demand) are the workhorse for steady-state compute. Reserved Instances for predictable database workloads. Spot Instances for fault-tolerant batch workloads (70 to 90% off). Avoid standard RI lock-in on workloads that might migrate regions or instance families, Compute Savings Plans offer equivalent discount with flexibility.
Yes. AWS defaults to 100% MACC, meaning you commit to a fixed monthly spend floor. Negotiate down to 90% or 85% MACC with quarterly flexibility, this prevents over-committing when a product launch slips or a workload migrates. On commitments above $10M, AWS will accept seasonal flexibility and quarterly true-up windows on strategic deals.
— amazon-web-services-aws-discount-negotiation
Microsoft Azure#
These ranges reflect Azure MACC and EA contracts benchmarked by our team across 2024 to 2026. "Effective discount" combines MACC headline + AHB + RI/SP + bundle economics.
| Annual Azure Spend | MACC Typical | Effective With Leverage | Notes |
|---|---|---|---|
| $500K to $1M | 3 to 7% | 10 to 18% | MACC at threshold; AHB + RI/SP are dominant levers at this tier. |
| $1M to $5M | 8 to 14% | 18 to 28% | EA bundle with M365/Dynamics unlocks; AWS RFP begins to move real points. |
| $5M to $15M | 14 to 22% | 26 to 36% | Sweet spot, full strategic-account team, AHB audit compulsory. |
| $15M to $50M | 20 to 28% | 32 to 42% | Executive relationship; multi-product EA + 5-year MACC + SI partnership. |
| $50M+ annual | 24 to 34% | 38 to 48% | Top-tier strategic; custom MSA, service-level carve-outs, Fast Track credits. |
Headline MACC discount is only part of the economics. A 22% MACC with AHB not applied, unused RI commitments, and portfolio-bundle obscurity is economically worse than a 16% MACC with AHB fully applied, 3-year RIs on stable workloads, and line-item EA transparency. Across a 3-year term, the gap routinely exceeds 18 to 25% of total Azure spend.
Azure MACC discounts scale with commitment size and term. A 1-year $1M MACC typically earns 4 to 8%; a 3-year $5M MACC earns 12 to 18%; a 3-year $20M+ MACC earns 20 to 28%; and a 5-year $50M+ MACC with AWS RFP pressure reaches 28 to 38%. Azure Hybrid Benefit stacks on top for additional 35 to 55% savings on Windows and SQL workloads.
Azure Hybrid Benefit (AHB) applies existing Windows Server and SQL Server licenses with Software Assurance to Azure workloads, saving 40 to 55% on compute costs for those workloads. This is on top of MACC discount and Reserved Instance commitments. AHB is one of the largest structural savings levers in Azure, and it is legitimately under-used, many buyers leave AHB unapplied and pay full Azure rates for Windows/SQL.
Yes, and you should. Bundling Azure MACC with Microsoft 365 E5, Dynamics 365, and Power Platform in a single Enterprise Agreement unlocks 5 to 12 additional discount points on each component. Microsoft field comp strongly rewards multi-product EAs. The trap: never let Microsoft bundle at the portfolio level in a way that blocks line-item benchmarking, insist on decomposed pricing for every product.
For stable long-term workloads, yes. Azure RIs offer 40 to 72% off pay-as-you-go for 1 or 3-year commitments, and can stack on MACC. But RIs lock you to specific VM families and regions, so use Savings Plans for Compute on workloads that might migrate and reserve RIs for truly stable database and compute commitments.
— microsoft-azure-discount-negotiation
Google Cloud#
These ranges reflect GCP custom-contract and CUD negotiations benchmarked by our team across 2024 to 2026. "Effective discount" combines custom-contract headline + CUD layer + sustained-use + bundle economics.
| Annual GCP Spend | Custom Contract Typical | Effective With Leverage | Notes |
|---|---|---|---|
| $500K to $1M | 4 to 10% | 12 to 22% | Custom contracts accessible; CUDs + sustained-use dominate savings. |
| $1M to $5M | 10 to 18% | 22 to 32% | Strategic-account engagement; AWS/Azure RFP unlocks real points. |
| $5M to $15M | 18 to 26% | 30 to 42% | Sweet spot, GCP aggressively competes on headline discount. |
| $15M to $50M | 24 to 34% | 38 to 48% | Executive relationship; BigQuery Editions + Vertex AI commitments stack. |
| $50M+ annual | 30 to 40% | 42 to 55% | Top-tier strategic; custom MSA, egress carve-outs, PSO credits. |
GCP's aggressive headline discount creates a real risk: buyers focus on the discount number and miss softer contract terms. A 35% GCP discount with uncapped egress, no commitment-reduction rights, and standard sustained-use rules is often economically comparable to a 25% AWS EDP with egress PPA, MACC flexibility, and reduction rights. Headline is not the whole picture, structure matters.
GCP custom-contract discounts scale aggressively with commitment because Google actively targets AWS and Azure displacement. A 1-year $1M commitment typically earns 6 to 12%; a 3-year $5M commitment earns 18 to 26%; a 3-year $25M+ commitment earns 28 to 38%; and a 5-year $50M+ commitment with AWS RFP reaches 35 to 48%. GCP is consistently the most aggressive on headline discount to win strategic workloads.
CUDs apply to Compute Engine, Cloud SQL, and Cloud Spanner with 1-year or 3-year commitments. Resource-based CUDs (tied to specific VM families) offer 37 to 57% discount. Spend-based CUDs (flexible across VM families) offer 20 to 46% discount. Spend-based CUDs are usually better than resource-based for workloads that might evolve. CUDs stack on custom-contract discount for effective savings of 35 to 55% below pay-as-you-go.
For enterprises running more than ~500 TB of query volume monthly or with steady BigQuery workloads, BigQuery Editions (flat-rate slots) deliver 30 to 60% savings versus on-demand query pricing. But on-demand is better for bursty, low-volume workloads. Model both carefully against actual query patterns, Google will often push flat-rate because it commits revenue, but it can be more expensive for low-volume use cases.
Yes. GCP offers meaningful discount on Google Cloud Marketplace purchases counting against committed spend (CUDs and custom contracts). This is under-utilized by most customers. BYOL (Bring Your Own License) for Oracle, Microsoft SQL, and Red Hat can save 30 to 50% versus Google-provided license bundles on applicable workloads.
— google-cloud-platform-gcp-discount-negotiation
Databricks#
These ranges reflect Databricks deals benchmarked across 2024 to 2026. "Achievable with leverage" assumes written Snowflake or Microsoft Fabric proposals, multi-workload DBU commitments, Unity Catalog adoption, and January (Databricks Q4) close.
| Deal Profile | Typical Discount | Achievable With Leverage | Notes |
|---|---|---|---|
| On-demand DBUs, no commitment | 0 to 5% | 0 to 8% | Essentially undiscounted. DBU commitment is the gateway to discount. |
| Single-workload, $250K to $750K/year | 15 to 25% | 25 to 35% | Mid-market tier. Written alternative essential to push above headline. |
| Multi-workload, $750K to $2M/year | 25 to 35% | 35 to 45% | SQL plus ML plus Workflows. Cross-workload fungibility required. |
| Multi-workload + governance, $2M+/year | 35 to 45% | 45 to 55% | Unity Catalog and Delta Sharing bundled at signing. |
| Snowflake displacement, $1M+/year | 40 to 50% | 50 to 60% | Strategic displacement pricing. Pre-IPO velocity amplifies depth. |
| Renewal without leverage | 0 to 5% off prior | N/A | Auto-renewal uplift 5 to 8%. Workload expansion drives the real increase. |
The compound lever most buyers miss: Databricks treats workload expansion, serverless migration, governance tooling, and renewal uplift as separate concessions from headline DBU discount. Optimizing one at the expense of the others often delivers worse 3-year total cost. Model effective cost across the full term including projected workload mix shifts and serverless adoption.
Databricks list pricing supports 25 to 55% discounts on DBU commitments for Fortune 500 buyers with credible alternatives. Benchmarked deals show median 35% off list on 3-year DBU commitments above $1M/year, rising to 45 to 55% with written Snowflake or Microsoft Fabric proposals, multi-workload commitments spanning SQL, ML, and workflows, Enterprise-tier commitments, and pre-IPO timing leverage. On-demand DBU pricing is essentially undiscounted, committed-use contracts are the gateway to any discount at all.
Size commitment at the mid-point of projected year-one DBU consumption with 25 to 30% headroom, not at peak projected year-three usage. Databricks rolls over unused commitment quarterly within term, so moderate over-commitment is protected. Over-commitment to hit a discount tier is the most common Databricks buyer mistake, the headroom premium typically outweighs the incremental discount for commitments above 150% of realistic year-one consumption. Negotiate quarterly rollover, annual re-baselining, and cross-workload DBU fungibility.
Moderate on base rate, aggressive on workload expansion. Standard renewal uplift is 5 to 8% on per-DBU pricing, but workload expansion during term, SQL workloads, ML training, Workflows, new users, is where effective renewal cost compounds. Fortune 500 customers with active ML programs routinely see 40 to 70% effective renewal increases driven primarily by workload expansion at list-rate overage. Cap per-DBU uplift at CPI or 3%, pre-negotiate overage at committed-tier pricing, and secure workload-cross-fungibility.
Yes, when your use case requires multiple workload types, and only with cross-workload DBU fungibility. Databricks bundles SQL, ML, and Workflows at 10 to 20 points of additional discount vs. single-workload commitments. But without cross-workload fungibility, committed DBUs from one workload cannot offset overage in another. Negotiate a unified DBU commitment pool that flows across all workloads, with rollover and re-baselining rights protecting against mix shifts during term.
— databricks-discount-negotiation
30 to 55% vs on-demandCapacity Commitment Discount Jobs Compute DBU (enterprise)
$0.15 to $0.28All-Purpose Compute DBU (enterprise)$0.30 to $0.55SQL Warehouse DBU (enterprise)$0.22 to $0.40Annual EscalationCPI or 3 to 5%Contracts Benchmarked160+Access Full Benchmark