How much do savings plans save? Commitment discounts on AWS compute

Quick answer

AWS savings plans discount compute by committing to a steady hourly spend for one or three years, and the saving ranges from roughly 20 percent up to about 72 percent versus on-demand depending on the plan type, term, and payment option. Three-year all-upfront commitments save the most; one-year no-upfront saves less but stays flexible. Compute Savings Plans are the flexible choice (they apply across instance families, regions, and even Lambda and Fargate); EC2 Instance Savings Plans discount more but lock you to a family. The real saving depends on how much of your steady usage you cover without over-committing.

Savings plans are AWS's main commitment-discount mechanism for compute. Instead of paying on-demand rates, you commit to a fixed amount of compute spend per hour for one or three years, and in exchange AWS discounts that usage. The discount can reach about 72 percent, but that headline number is the best case, what you actually save depends on several choices.

What drives the discount

Choice	Effect on savings
Term (1 vs 3 year)	3 year saves more, less flexible
Payment (all/partial/no upfront)	More upfront, bigger discount
Plan type (Compute vs EC2 Instance)	EC2 Instance saves more, less flexible
Coverage	Cover steady usage, not peaks

A three-year, all-upfront EC2 Instance Savings Plan is where the roughly 72 percent lives. A one-year, no-upfront Compute Savings Plan might save closer to 20 to 30 percent but keeps you flexible and requires no cash outlay. The discount is a reward for certainty: the more you commit and prepay, the more AWS shaves off. This is the same tradeoff behind GCP committed use discounts.

Compute vs EC2 Instance Savings Plans

Compute Savings Plans are flexible: your commitment applies across instance families, sizes, regions, and even Lambda and Fargate, so it keeps discounting even as your architecture changes. EC2 Instance Savings Plans give a deeper discount but lock the commitment to a specific instance family in a region. Most teams favor Compute plans for the flexibility, accepting a slightly smaller discount, because a plan that keeps applying as you evolve beats a bigger discount you might strand. Understanding how EC2 is priced makes the choice clearer.

The coverage tradeoff

The catch is that you pay for the commitment whether or not you use it, so over-committing wastes money on capacity you do not consume. The goal is to cover your steady baseline (the compute that runs 24/7) and leave variable peaks on on-demand or spot. Look at your minimum consistent usage over recent months and commit to that floor, not your average or peak. Coverage and utilization are core FinOps KPIs: high coverage of steady usage with high utilization of the commitment is the sweet spot.

How to get the most out of savings plans

Analyze your steady-state usage, commit to the baseline you are confident will persist, and layer plans over time rather than committing everything at once. Combine savings plans with Graviton, since commitment discounts apply on top of the roughly 20 percent Graviton price advantage, and use spot for interruptible work outside the commitment. Review coverage and utilization regularly, and forecast future usage (see forecasting cloud costs) before signing a three-year term. Done right, savings plans are among the highest-ROI cost moves on AWS: a large discount for committing to usage you were going to run anyway.

FAQ

How much do AWS savings plans save?

Roughly 20 percent up to about 72 percent versus on-demand, depending on the plan type, term, and payment option. A three-year, all-upfront EC2 Instance Savings Plan reaches the top of that range, while a one-year, no-upfront Compute Savings Plan saves less but stays flexible and requires no cash outlay. The actual saving also depends on how well you match the commitment to your steady usage.

What is the difference between Compute and EC2 Instance Savings Plans?

Compute Savings Plans are flexible: the commitment applies across instance families, sizes, regions, and even Lambda and Fargate, so it keeps discounting as your architecture changes, at a slightly smaller discount. EC2 Instance Savings Plans give a deeper discount but lock the commitment to a specific instance family in a region. Most teams favor Compute plans for flexibility.

How do I get the maximum savings plan discount?

Choose a three-year term with all-upfront payment and, for steady usage in a fixed family, an EC2 Instance Savings Plan, which together reach the roughly 72 percent range. The tradeoff is less flexibility and a large cash commitment, so weigh the deeper discount against the risk that your usage or architecture changes over three years.

What happens if I do not use my savings plan commitment?

You still pay for it. A savings plan commits you to a fixed hourly spend whether or not you use that much compute, so over-committing wastes money on capacity you do not consume. That is why you should commit to your steady baseline (the compute running 24/7) rather than your average or peak, leaving variable usage on on-demand or spot.

Can I combine savings plans with other discounts?

Yes. Savings plan discounts apply on top of the roughly 20 percent Graviton price advantage, so running Graviton instances under a savings plan stacks both. You can also use spot instances for interruptible work outside the commitment. Layering plans over time and combining them with Graviton and spot is how teams push their effective compute discount higher.

How does C3X help with savings plan decisions?

C3X prices your Terraform against a live catalog before deploy, giving you the on-demand cost of steady infrastructure so you can see what is worth committing to. Knowing the baseline cost of what you run makes it easier to size savings plan commitments to your real steady usage rather than guessing from the bill.

What to do next

Know your steady compute cost before you commit. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.