One of the most powerful levers for reducing your AWS bill is moving away from On-Demand pricing and embracing a commitment-based discount model. AWS offers two primary options: the traditional Reserved Instances (RIs) and the more modern Savings Plans (SPs). Both can provide discounts of up to 72-75% for a 1 or 3-year commitment, but they differ fundamentally in flexibility.

The Models Explained

Reserved Instances (RIs): The Legacy Choice

RIs are a commitment to a specific instance configuration in a specific region.

• Standard RIs: Offer the highest discount (up to 75%) but are the most rigid. You commit to a specific instance family, size, OS, and region. Ideal for extremely stable workloads.

• Convertible RIs: Offer a lower discount but allow you to exchange your commitment for a different instance type, providing some flexibility.

• Capacity Reservation: A key feature of Standard RIs is that they can provide a capacity reservation, guaranteeing you can launch your instance when you need it.

Savings Plans (SPs): The Modern, Flexible Choice

Introduced to address the rigidity of RIs, Savings Plans are a commitment to a certain amount of spend ($/hour).

• Compute Savings Plans: This is the most flexible option. You commit to a spend (e.g., $10/hour), and that discount automatically applies to any EC2 instance usage, regardless of family, size, or region. It also applies to AWS Fargate and AWS Lambda usage.

• EC2 Instance Savings Plans: A middle ground. You commit to a specific instance family in a specific region and get a higher discount than a Compute SP, but less flexibility.

Head-to-Head Comparison

Scenarios: When to Choose Which Model

The best choice depends on your workload's predictability.

Choose Reserved Instances if:

• Your workload is extremely stable and predictable (e.g., a large monolithic application).

• You absolutely need a capacity reservation for a mission-critical application.

• You need discounts on non-compute services like RDS or Redshift.

Choose Savings Plans if:

• Your workloads are dynamic or evolving (the case for most modern applications).

• You use containers (Fargate) or serverless (Lambda).

• You want to simplify management, as SPs "just work" automatically.

The Hybrid Approach: The Best of Both Worlds

For many large organizations, the optimal strategy is a hybrid approach.

1. Use RIs for your foundational, never-changing workloads like core production databases to lock in the highest discount and capacity reservations.

2. Use a flexible layer of Compute Savings Plans to cover the rest of your dynamic compute workloads.

Conclusion

While Reserved Instances still have a place, Compute Savings Plans have become the default and recommended choice for the vast majority of AWS users. The small premium paid in a slightly lower discount is almost always worth the immense gain in flexibility, allowing your engineering teams to innovate without being financially penalized by rigid, outdated commitments.