Savings Plans

Savings plans are a flexible form of commitment discount. Instead of committing to a specific instance type in a specific region, as with standard reserved instances, a customer commits to a dollar amount of compute spending per hour, such as $50 an hour, for one or three years. Any eligible usage up to that amount is billed at discounted rates, and usage above it is billed at on-demand rates. Because the commitment is expressed in spending rather than specific resources, it can apply across instance families, sizes and, in some cases, regions and services.

AWS introduced savings plans in 2019, and Microsoft Azure offers a similar savings plan for compute. For many organizations, savings plans have become the main way to commit to compute usage, since they deliver substantial discounts with less risk of being stranded when workloads change. For IT financial management, savings plans are a central tool for rate optimization, balancing discount depth against flexibility.

Definition of Savings Plans

A savings plan is a billing commitment to a consistent hourly spend on eligible compute services over a one-year or three-year term, in exchange for lower rates. The commitment is measured in dollars per hour of discounted spend.

Each hour, eligible usage is charged at the savings plan rates until the commitment is used up. If usage falls below the commitment, the customer still pays the full committed amount for that hour.

Types of Savings Plans

AWS offers several types. Compute Savings Plans apply to EC2 instances in any family, size, operating system or region, as well as to AWS Fargate and AWS Lambda, offering the most flexibility with discounts that AWS describes as up to 66%. EC2 Instance Savings Plans apply to a chosen instance family in a chosen region, offering larger discounts of up to 72% in exchange for less flexibility.

AWS also offers savings plans for specific services, such as machine learning workloads. Microsoft Azure's savings plan for compute applies across many compute services and regions.

Payment options, such as all upfront, partial upfront or no upfront, are similar to those for reserved instances, with higher upfront payments generally earning larger discounts.

Savings Plans vs Reserved Instances

Standard reserved instances commit to specific instance attributes and can offer the deepest discounts, but they can become stranded when workloads change instance type or move to other services.

Compute Savings Plans apply automatically across instance families, regions and some serverless services, so they continue to deliver savings as architectures evolve. This makes them easier to use fully over a long term.

Many organizations combine the two, using savings plans for most compute commitments and reserved instances where they offer benefits that savings plans do not, such as capacity reservations or coverage for services that savings plans do not include.

Importance of Savings Plans in IT Financial Management

Savings plans make it easier to capture commitment discounts without detailed predictions of which instance types will be used. This lowers the barrier to rate optimization.

They also reduce the risk of unused commitments. Commitments tied to specific resources can become stranded when those resources change, but a spending commitment can follow usage to new instance types.

Because the commitment follows spending rather than specific resources, it remains useful through migrations, modernization and changes in instance families.

Broader Coverage With Less Effort

A single savings plan purchase can cover usage across many accounts, teams and instance types. This simplifies purchasing for the central FinOps team and helps raise commitment coverage quickly.

Engineers remain free to choose the most suitable instance types, including newer and more efficient ones, without losing the discount. This supports usage optimization alongside rate optimization.

The result is often a higher effective savings rate than an organization can achieve with reserved instances alone.

Committed Spending and Budgets

A savings plan creates a fixed hourly commitment for the term. If usage falls, for example because of optimization or reduced demand, the organization still pays the committed amount, and commitment utilization drops.

Finance should treat savings plan commitments as committed spending in budgets and forecasts, and amortize upfront payments over the term. This makes the cost of the commitment visible in each month it applies.

Coordination between the FinOps team and engineering is important, so that major optimization efforts or migrations are reflected in future commitment decisions.