What Azure Reserved Instances Are (and Are Not)
An Azure Reserved Instance (RI) is a pricing commitment — not a reserved resource. When you purchase a 1-year or 3-year RI for a specific VM family and region, Microsoft applies a discount to your Azure bill for matching compute usage. You are not reserving capacity; you are committing to pay for a quantity of compute hours at a discounted rate, and Microsoft automatically applies the reservation benefit to matching usage. This article is part of our Microsoft Enterprise Agreement Negotiation: Definitive Guide.
RIs are distinct from Azure Savings Plans, which provide a commitment-based discount across multiple compute types and regions in exchange for a fixed hourly spend commitment. Both mechanisms deliver significant savings — the choice between them depends on your workload characteristics and flexibility requirements, which we cover below.
RIs vs Azure Savings Plans: Which to Use
Microsoft offers two commitment-based discount mechanisms for Azure compute:
Free Guide
Microsoft EA Negotiation Tactics
How Fortune 500 buyers slash Microsoft EA costs — true-up traps, ELP rules, and renewal leverage.
Reserved Instances
RIs are tied to a specific VM family, region, and optionally operating system. They deliver the highest discount rates — typically 40–60% for 1-year and 3-year terms. The trade-off is specificity: a D-series RI in East US does not apply to an F-series workload in West Europe. If your workloads migrate, resize, or change region, the RI may not match. Instance size flexibility (within the same VM family and region) helps — an RI for a Standard_D4s_v3 can apply to multiple smaller D-series instances totalling the same vCPU count — but cross-family flexibility does not exist.
Azure Compute Savings Plans
Savings Plans are a newer mechanism that commits to a fixed hourly spend (e.g., $10/hour) across any Azure compute type, region, and operating system. The discount rate is lower than RIs (typically 15–35%) but the flexibility is much higher. Savings Plans are the right choice for workloads that are actively migrating between VM families or regions, or for Dev/Test environments with highly variable resource profiles.
The Combined Strategy
The most cost-effective approach for large Azure estates uses both: RIs for stable, predictable workloads (production databases, fixed-size application servers, permanent infrastructure); Savings Plans for variable workloads, development environments, and workloads in active migration or modernisation programmes. Azure Cost Management provides a blended recommendation that optimises the allocation between the two mechanisms based on your actual usage patterns.
Azure Hybrid Benefit stacking: Azure Hybrid Benefit allows organisations with Windows Server and SQL Server licences covered by active Software Assurance (available through the EA) to apply those licences to Azure VMs, eliminating the Windows licence component of the Azure compute cost. Stacked with a 3-year RI, Azure Hybrid Benefit can deliver total savings of 65–72% against pay-as-you-go Windows/SQL pricing. This is the highest available Azure compute discount mechanism and is only accessible through the EA. See our Azure Hybrid Benefit: Max Savings guide for the full methodology.
Stay Ahead of Vendors
Get Negotiation Intel in Your Inbox
Monthly briefings on vendor pricing changes, audit trends, and contract tactics. Unsubscribe any time.
No spam. No vendor affiliations. Buyer-side only.
RI Sizing: The Most Common Mistakes
Over-commitment on RI sizing is the most frequent Azure cost optimisation failure. Microsoft's Azure Portal RI recommendations are a starting point, not an endpoint — they are generated from recent usage patterns that may not reflect future workload plans.
Mistake 1: Using Current Usage Without Accounting for Planned Changes
If your organisation is planning a workload migration, application rationalisation, or cloud-to-cloud migration during the RI commitment term, current usage is not the right baseline for RI sizing. RIs for workloads that will be decommissioned, migrated, or significantly resized during the term deliver poor return. Model your workload plans for the next 12–36 months before committing.
Mistake 2: Buying 3-Year RIs for All Workloads
The 3-year RI delivers the highest discount, but only for workloads you are confident will run unchanged for the full three years. For any workload under active review, a 1-year RI provides meaningful savings with a shorter lock-in period. A blended portfolio — 3-year for core infrastructure, 1-year for application workloads, Savings Plans for development — is typically the right structure for large enterprises.
Mistake 3: Purchasing Through the Self-Service Portal Without Negotiation
RIs purchased through the Azure Portal self-service are priced at Microsoft's standard rates. RIs purchased as part of an EA negotiation — particularly when combined with a MACC (Microsoft Azure Consumption Commitment) — receive additional discounts of 5–15% beyond portal pricing. See our Azure Committed Spend: Negotiation Guide for how MACC affects RI pricing.
Negotiating RIs in Your EA
Azure RI procurement should be part of your EA negotiation strategy, not a post-renewal self-service decision. Key negotiation points:
Include RI Value in MACC Sizing
When negotiating your MACC (Azure Monetary Commitment), the total annual value of your committed RI purchases contributes to your MACC tier. Higher MACC tiers typically unlock additional Azure consumption discounts. Structuring your RI commitments as part of the MACC negotiation maximises tier-based benefits across your entire Azure estate — not just the RI-covered compute.
Negotiate RI Cancellation and Exchange Rights
Microsoft's standard RI terms allow cancellation with a 12% early termination fee and exchanges to a different RI of equal or greater value. In EA negotiations, some enterprise customers have obtained enhanced exchange rights — broader instance size flexibility, cross-region exchanges — as part of a larger Azure commitment package. These terms are not published but are achievable under competitive pressure.
Use AWS and GCP as Competitive Leverage
A documented workload migration analysis showing equivalent AWS or GCP compute at a lower effective cost — including reserved pricing comparisons — is a powerful lever in Azure RI negotiations. Microsoft's account team has authority to offer Azure-wide discounts (not just on RIs) when a credible cloud migration threat is presented. This is particularly effective for data and AI workloads where Azure, AWS, and GCP compete directly.
RI Governance: Preventing Waste
Purchased RIs that do not match active usage represent paid capacity earning no return. RI governance — monitoring reservation utilisation and managing the RI portfolio over time — is as important as the initial sizing decision.
- Review RI utilisation monthly via Azure Cost Management — target utilisation above 85% for all active reservations
- Set up Azure Reservations utilisation alerts to catch under-utilised RIs before they represent significant waste
- Exchange or cancel under-utilised RIs when workload changes make the original commitment unviable — the cost of early exchange is typically less than the cost of carrying a mismatched reservation
- Assign RI purchase authority to a central FinOps function — decentralised RI purchasing across multiple teams frequently results in overlapping commitments and gaps
Further Reading
- Microsoft Enterprise Agreement Negotiation: Definitive Guide — the full EA commercial framework
- Azure Committed Spend: Negotiation Guide — MACC sizing and discount optimisation
- Azure Hybrid Benefit: Max Savings — stacking Hybrid Benefit with RIs
- Microsoft EA Renewal: 15 Tactics That Work — EA renewal negotiation tactics
- Right-Size Your Microsoft Licence Estate — full Microsoft licence optimisation
- Microsoft EA Negotiation Guide (Free PDF) — downloadable reference
- Microsoft Advisory Service — IT Negotiations' full EA and Azure engagement capability
- Case Study: $8.4M Microsoft EA Savings