AWS Graviton Savings Calculator
Estimate what you save moving compute to Graviton (Arm) — the ~20% list-price cut across EC2, RDS, ElastiCache, and more, plus optional price-performance headroom.
- No signup
- Instant results
- ~2 min
Rates reviewed
Your Graviton-eligible compute
us-east-1list rates, pricing as of 2026-07-09. Source: AWS public pricing.
Graviton instances list at roughly 20% less than the equivalent x86 size, and AWS cites up to 40% better price-performance on many workloads. Enter the spend that could run on Arm.
Monthly compute spend on Graviton-capable services
EC2, RDS, Aurora, ElastiCache, OpenSearch, and Lambda all offer Graviton. Exclude spend already on Graviton.
How much of it can port to Arm?
Most Linux and container workloads port cleanly. x86-only binaries, some vendor agents, and legacy dependencies do not.
Who This Tool Is For
CTOs, platform engineers, and FinOps leads running x86 EC2, RDS, or container fleets who keep hearing "just move to Graviton" without a number attached. If compute is a top line item, this sizes the prize before you scope the work.
Why We Built This Tool
Graviton (AWS's Arm processors) is one of the few cost levers that cuts the rate without cutting capacity — roughly 20% cheaper list price for the same instance size, and often better price-performance on top. The blocker is rarely cost; it's knowing how much of your workload ports and what the migration takes. This calculator sizes both.
What Problem It Solves
- Unquantified upside. "Move to Graviton" is advice, not a number. This turns it into a monthly and annual figure.
- Portability reality. Not everything ports — the portability input keeps the estimate honest.
- Conservative vs aggressive. Separate the defensible 20% price cut from workload-dependent performance upside.
- Stacks with commitments. Graviton savings compound with Savings Plans — model the base here first.
Pair with the EC2 and Savings Plans calculators, or roll it into the whole-bill calculator.
Frequently Asked Questions
How much cheaper is Graviton than x86?
For the same instance size, Graviton (Arm) On-Demand list prices run roughly 20% below the comparable x86 (Intel/AMD) instances across EC2, RDS, Aurora, ElastiCache, and OpenSearch. AWS also cites up to 40% better price-performance on many workloads, which can let you run fewer or smaller instances — that upside is workload-dependent, which is why this tool separates it from the flat price cut.
Which AWS services support Graviton?
EC2 (many families with a "g" suffix like m7g, c7g, r7g), Lambda (arm64), Fargate, RDS and Aurora, ElastiCache, OpenSearch, MemoryDB, and more. Managed services are the easiest wins because there is no code to rebuild — you change the instance class and restart in a maintenance window.
What does the portability percentage mean?
It is the share of your Graviton-eligible spend that can actually run on Arm. Most Linux and containerized workloads port cleanly, but x86-only binaries, some proprietary vendor agents, and native code without arm64 builds do not. Set it to reflect how much of your stack is genuinely portable — 100% is rare in practice.
Is the performance-headroom scenario safe to put in a budget?
Treat the conservative "price only" number as your defensible floor for a budget. The aggressive scenario assumes better price-performance lets you shed some capacity, which is real on many workloads but must be proven with benchmarks. Use it to size the opportunity, not to commit a number.
Do Savings Plans work with Graviton?
Yes. Compute Savings Plans apply to Graviton instances just like x86. The best-practice order is migrate to Graviton first to lower the rate, then apply a Savings Plan on the reduced baseline so you are not committing to spend you are about to eliminate.