AWS Reserved Instances and Savings Plans: A Comprehensive Guide

Introduction

Many AWS organizations have significant savings opportunities just a few clicks away — yet they either don't know these options exist or hesitate to commit out of fear of locking in costs they can't reverse. AWS offers two primary commitment-based discount mechanisms to help reduce cloud costs: Reserved Instances (RIs) and Savings Plans (SPs). Both require a 1-year or 3-year commitment in exchange for discounts of up to 75% compared to On-Demand pricing. With the right understanding of how these models work, the tools available to analyze your usage, and a clear strategy for right-sizing before purchasing, achieving meaningful savings becomes straightforward and low-risk. This guide breaks down the differences, coverage, optimal use cases, and best practices to help you confidently evaluate, purchase, and manage your commitments.

---

Savings Plans

Savings Plans are a flexible pricing model offering lower prices compared to On-Demand pricing in exchange for a specific usage commitment (measured in $/hour) for a 1-year or 3-year period. AWS currently offers four types of Savings Plans:

1. Compute Savings Plans

• Most flexible option — applies automatically across EC2, Lambda, and Fargate usage
• Savings of up to 66% off On-Demand rates
• Works across any instance family, size, OS, tenancy, or Region
• Ideal for dynamic workloads that may shift across services or regions

2. EC2 Instance Savings Plans

• Commitment to a specific instance family within a chosen Region
• Savings of up to 72% off On-Demand rates
• Flexibility to change instance size and OS within the same family
• Best for predictable EC2 workloads in a known region

3. Database Savings Plans (NEW — Launched December 2025)

• Newest addition — announced at AWS re:Invent 2025
• Savings of up to 35% compared to On-Demand pricing
• Covers 10 AWS database services:
  • Amazon Aurora
  • Amazon RDS
  • Amazon DynamoDB
  • Amazon ElastiCache (including Valkey)
  • Amazon DocumentDB (with MongoDB compatibility)
  • Amazon Neptune
  • Amazon Keyspaces (for Apache Cassandra)
  • Amazon Timestream
  • AWS Database Migration Service (AWS DMS)
  • Amazon OpenSearch Service
• Applies regardless of engine, instance family, size, deployment option, or Region
• Available as a 1-year term with No Upfront payment
• Works with existing pricing programs and discount agreements for additional savings
• Applies to both serverless and provisioned instances (Gen 7+ only) — older generation instances are not eligible for Database Savings Plans; customers must be running on the latest instance generations (e.g., db.r7g, db.m7g, etc.) to benefit
• Cannot be combined with Reserved Instances or DynamoDB reserved capacity on the same workload (but can be used alongside them for different workloads)

For detailed pricing information across all supported database services, see: Database Savings Plans Pricing

4. SageMaker Savings Plans

• Specific to machine learning workloads
• Applies to SageMaker instance usage across any Region
• Automatic application across eligible ML instances

---

Reserved Instances (RIs)

Reserved Instances provide a significant discount (up to 75%) compared to On-Demand pricing in exchange for a commitment to a specific instance configuration for a 1-year or 3-year term. AWS offers two types:

Standard Reserved Instances

• Highest discount potential — up to 75% off On-Demand
• Fixed instance attributes (family, size, region)
• Best for steady-state workloads
• Can be sold in the AWS Marketplace

Convertible Reserved Instances

• More flexibility to change instance attributes
• Lower discount — up to 54% off On-Demand
• Cannot be sold in the Marketplace
• Allows modification of instance family, OS, and tenancy

Services That Support Reserved Instances

• Amazon EC2 — compute instances
• Amazon RDS — relational database instances (up to 69% savings)
• Amazon ElastiCache — in-memory caching
• Amazon Redshift — data warehousing
• Amazon OpenSearch Service — search and analytics
• Amazon DynamoDB — reserved capacity
• Amazon MemoryDB — in-memory database

---

Commitment Options

Both RIs and Savings Plans share similar commitment structures:

Feature	Reserved Instances	Savings Plans
Term Length	1-year or 3-year	1-year or 3-year (Database SP: 1-year only)
Payment Options	All Upfront, Partial Upfront, No Upfront	All Upfront, Partial Upfront, No Upfront
Commitment Basis	Specific instance configuration	Hourly spend commitment ($/hour)
Maximum Savings	Up to 75% (Standard RI)	Up to 72% (EC2 Instance SP)
Flexibility	Limited (Standard) to Moderate (Convertible)	High (Compute SP) to Moderate (EC2 Instance SP)
Capacity Reservation	Yes (zonal RIs)	No
Resale	Yes (Standard RIs via Marketplace)	No

Payment Options Explained

• No Upfront — Billed as a monthly fee; lowest barrier to entry
• Partial Upfront — Initial payment of at least half the commitment, with monthly fees for the remainder
• All Upfront — Entire commitment charged in one payment; provides the highest discount

Term Considerations

• 1-year terms — Allow customers to evaluate benefits while maintaining flexibility
• 3-year terms — Provide significantly higher savings (up to 100% additional discount depending on service and region); best for stable, predictable workloads with consistent usage patterns over 12+ months

---

Savings Plans and RIs Across AWS Organizations (Multi-Account)

When using AWS Organizations with consolidated billing, Savings Plans and Reserved Instances apply across all linked accounts in the organization by default:

• Savings Plans and RIs purchased in any account within the organization are shared across all member accounts
• The discount is automatically applied to eligible usage in any linked account, maximizing utilization
• Sharing can be disabled at the payer account level if you need to restrict Savings Plans or RIs to specific accounts
• This means a single, well-sized Savings Plan purchase can cover workloads distributed across your entire organization
• Consider purchasing Savings Plans from the management (payer) account for simplicity, or from individual accounts if you need cost allocation granularity

For details on how to enable or disable discount sharing across accounts, see: Turning off shared Reserved Instances and Savings Plans

---

Key Considerations

1. Cannot be cancelled — RIs and Savings Plans are commitment-based models that cannot be cancelled once purchased. Recommendations should be based on consistent usage patterns with the goal of achieving 100% utilization throughout the term.

2. Database Savings Plans vs. Database RIs — These cannot be combined on the same workload. Choose one based on your needs: Database Savings Plans offer more flexibility (engine, family, region agnostic), while RIs may offer deeper discounts for very stable, specific configurations.

3. Automatic application — Savings Plans automatically apply to eligible usage, reducing management overhead. RIs require more active management and matching to specific instances.

4. Stacking discounts — Database Savings Plans work with existing pricing programs and discount agreements for additional savings.

5. Coverage analysis — Regularly review utilization and coverage to ensure commitments align with actual usage patterns.

6. Right-sizing first — Before purchasing commitments, ensure instances are right-sized to avoid locking in discounts on oversized resources.

7. Latest generation requirement for Database Savings Plans — Database Savings Plans only apply to Gen 7+ instances (both provisioned and serverless). Customers running older generation database instances (e.g., db.r5, db.r6g) will need to upgrade to the latest generation before they can benefit from Database Savings Plans. This is an important consideration when planning migrations and cost optimization strategies.

---

Utilization vs. Coverage — Understanding the Difference

These two metrics are often confused but measure fundamentally different things:

Metric	What It Measures	Question It Answers
Utilization	How much of your purchased commitment is being consumed	"Am I wasting money on unused commitments?"
Coverage	How much of your eligible usage is covered by commitments	"Am I leaving savings on the table by running too much On-Demand?"

Utilization

• Formula: (Used commitment / Total commitment) × 100
• Target: 100% — meaning every dollar you committed is being applied to eligible usage
• Low utilization indicates you over-purchased or your workloads have decreased since buying
• Example: You committed $10/hour but only $7/hour of eligible usage exists → 70% utilization (you're paying for $3/hour you don't need)

Coverage

• Formula: (Spend covered by commitments / Total eligible spend) × 100
• Target: As close to 100% as practical — meaning minimal On-Demand spend remains
• Low coverage indicates you have eligible usage that could benefit from additional commitments
• Example: Your total eligible spend is $20/hour but only $17/hour is covered by commitments → 85% coverage ($3/hour is running at On-Demand rates)

The Balance

The goal is to achieve high utilization AND high coverage simultaneously. If you optimize only for coverage (buying more), utilization may drop. If you optimize only for utilization (buying less), coverage suffers. Use both reports together to find the sweet spot.

---

How to Purchase Savings Plans

Savings Plans can be purchased directly through the AWS Console via the Billing and Cost Management → Savings Plans → Purchase Savings Plans page. The workflow allows you to:

• Select the Savings Plan type (Compute, EC2 Instance, Database, or SageMaker)
• Choose your term length (1-year or 3-year) and payment option (All Upfront, Partial Upfront, or No Upfront)
• Specify your hourly commitment amount
• Review estimated savings before confirming the purchase
• Add multiple Savings Plans to a cart for batch purchasing

For a step-by-step walkthrough of purchasing Savings Plans via the console, see this official AWS video: How to Purchase Savings Plans

---

Tools for Analysis and Recommendations

AWS provides several tools to help analyze usage patterns and identify commitment opportunities:

AWS Cost Explorer — Recommendations Page

• Provides recommendations for additional RI and Savings Plans purchases based on historical usage patterns
• Analyzes 7, 30, or 60 days of historical usage
• Suggests appropriate commitment types, duration, and instance configurations
• Provides estimated savings calculations
• Allows comparison of different term lengths and payment options
• Access via: AWS Console → Cost Management → Cost Explorer → Recommendations

Cost Optimization Hub

• Quantifies and aggregates estimated savings across all recommendation types
• Takes your specific AWS discounts (RIs and Savings Plans) into consideration
• Deduplicates amongst Savings Plans and Reserved Instances recommendations
• Defaults to commitment options offering the highest overall savings (prioritizing Compute Savings Plans for flexibility)
• Typically favors 3-year All Upfront options (customizable in preferences)
• Helps measure cost efficiency and establish consistent savings benchmarks
• Enables easy comparison and prioritization of recommendations
• ⚠️ Note: Database Savings Plans recommendations do not currently appear in Cost Optimization Hub. To view and evaluate Database Savings Plan opportunities, navigate directly to the Savings Plans recommendations page in Cost Explorer (AWS Console → Cost Management → Cost Explorer → Recommendations → Savings Plans). This is the only place where Database Savings Plan purchase recommendations are surfaced.

Purchase Analyzer

The Purchase Analyzer lets you simulate and estimate the cost, coverage, and utilization impact of your next Savings Plan purchase — before you commit any money.

How It Works:

1. Define your analysis — Specify the Savings Plan type, term, payment option, and hourly commitment amount you're considering
2. Run the analysis — The tool evaluates your historical usage data against the proposed commitment
3. Review results across three dimensions:
   • Cost tab — Visualizes how the new commitment layers on top of your existing Savings Plans, showing the remaining estimated On-Demand spend after purchase
   • Coverage tab — Shows how much of your eligible usage would be covered with the additional commitment
   • Utilization tab — Projects how well the new commitment would be utilized based on your usage patterns

Key Metrics Provided:

• Savings Plan recommendation — The hourly commitment amount analyzed (e.g., $5.417/hour)
• Estimated savings — Projected monthly dollar savings (e.g., $453.19/month)
• Estimated savings (%) — Percentage reduction compared to On-Demand (e.g., 9%)

Why Use It:

• Risk-free simulation — Test different commitment amounts without purchasing
• Visual gap analysis — See exactly how much On-Demand spend remains uncovered
• Informed decisions — Compare multiple scenarios before committing
• Seamless purchase path — Once satisfied, click "Add Savings Plan to cart" to proceed directly to purchase

For a detailed walkthrough on how to use the Purchase Analyzer, see this official AWS video: How to Use the Purchase Analyzer

• Access via: AWS Console → Billing and Cost Management → Savings Plans → Purchase Analyzer

Savings Plans Coverage Report

• Tracks how well your Savings Plans commitments cover your eligible usage over time
• Displays key metrics at a glance: On-Demand spend not covered, Average coverage percentage, and Potential monthly savings vs On-Demand
• Provides a daily or monthly coverage graph showing coverage trends, making it easy to identify gaps or declining coverage
• Allows setting a coverage target (e.g., 100%) with a visual target line on the chart for quick comparison
• Helps identify periods where On-Demand spend is increasing due to insufficient commitment coverage
• Exportable as CSV for custom analysis and reporting
• Use this report to determine when additional Savings Plans purchases are needed to maintain optimal coverage
• Access via: AWS Console → Billing and Cost Management → Savings Plans → Coverage Report

Additional Tools

• AWS Budgets — Set alerts for commitment utilization thresholds
• Savings Plans Utilization Reports — Track how well your existing commitments are being consumed
• AWS Cost and Usage Report (CUR) — Detailed billing data for custom analysis
• Cost Anomaly Detection — Identify unexpected changes in spending patterns
• Cost Efficiency Metric (via Cost Optimization Hub) — Requires AWS Compute Optimizer to be enabled; provides a consistent benchmark to measure and track cost optimization progress across accounts and organizations. Learn more: Measuring Cloud Cost Efficiency with the New Cost Efficiency Metric by AWS

---

Recommendations and Best Practices

Choosing the Right Commitment Type

Scenario	Recommended Option
Workloads shift across regions or services frequently	Compute Savings Plans
Steady EC2 workloads in a known region and instance family	EC2 Instance Savings Plans
Multiple database services across engines and regions	Database Savings Plans
Single, stable RDS engine with a fixed configuration	Database Reserved Instances (may offer deeper discount)
ML workloads on SageMaker	SageMaker Savings Plans
Need capacity reservation guarantee	Zonal Reserved Instances

Migrating from Reserved Instances to Savings Plans

If you currently rely on Reserved Instances and want to transition to the more flexible Savings Plans model:

1. Don't cancel existing RIs — Let them expire naturally at the end of their term
2. Layer Savings Plans on top — Purchase Savings Plans to cover any gaps not addressed by active RIs
3. Replace expiring RIs with Savings Plans — As each RI expires, evaluate whether a Savings Plan provides better flexibility for that workload
4. Run both in parallel — It's perfectly valid to maintain RIs for stable, predictable workloads while using Savings Plans for dynamic ones
5. Use Cost Explorer — Monitor the "Savings Plans vs. RI" recommendation comparison to identify the best replacement strategy

Common Mistakes to Avoid

1. Buying before right-sizing — Purchasing commitments on oversized instances locks in waste. Always right-size first, then commit.
2. Over-committing based on peak usage — Base your commitment on your sustained baseline, not peak demand. Spikes should remain On-Demand or use Spot.
3. Ignoring planned changes — If workloads are being migrated, decommissioned, or significantly restructured within the commitment term, factor that into your purchase decision.
4. Committing to 3-year terms too early — Start with 1-year terms to validate usage patterns before locking in longer commitments.
5. Neglecting utilization monitoring — After purchase, regularly review utilization reports to catch declining usage early.

When NOT to Buy Commitments

Not every workload benefits from commitments. Avoid purchasing RIs or Savings Plans for:

• Short-lived projects — Workloads running less than the commitment term
• Workloads being migrated off AWS — If you're planning to move services to another provider or on-premises
• Highly variable environments — Dev/test environments with unpredictable schedules (consider Spot Instances instead)
• Workloads under active optimization — If you're still right-sizing or re-architecting, wait until usage stabilizes
• Planned decommissions — Services scheduled for retirement within the commitment period

---

Summary

Savings Plan / RI Type	Max Savings	Flexibility	Services Covered
Compute Savings Plans	Up to 66%	Highest	EC2, Lambda, Fargate
EC2 Instance Savings Plans	Up to 72%	Moderate	EC2 (specific family/region)
Database Savings Plans	Up to 35%	High	Aurora, RDS, DynamoDB, ElastiCache, DocumentDB, Neptune, Keyspaces, Timestream, DMS, OpenSearch
SageMaker Savings Plans	Up to 64%	High	SageMaker
Standard RIs	Up to 75%	Low	EC2, RDS, ElastiCache, Redshift, OpenSearch, DynamoDB, MemoryDB
Convertible RIs	Up to 54%	Moderate	EC2

---

Conclusion

Choosing the right mix of Reserved Instances and Savings Plans depends on your workload characteristics, flexibility requirements, and financial planning preferences. The newly launched Database Savings Plans (December 2025) simplify database cost optimization by providing a single, flexible commitment that spans multiple database services without locking you into specific engines or instance types. Use AWS Cost Explorer Recommendations, the Purchase Analyzer, and Cost Optimization Hub to identify the optimal commitment strategy for your environment.