Graviton vs x86 Migration Payback Estimator
Estimate the simple payback of moving an x86 workload to Graviton using your current spend, a price discount and a benchmarked throughput gain. No live cloud prices or AWS instance recommendations are included.
Result
Simple payback
Estimated Graviton monthly compute spend
Estimated monthly compute savings
Estimated spend reduction
How to use this tool
- Enter current monthly x86 compute spend for the workload and billing scope you want to compare.
- Enter the Graviton price discount and throughput gain measured for the same workload and service target.
- Add one-time engineering, migration and testing cost.
- Review the estimated equivalent monthly spend, savings and simple payback; use your own current pricing and benchmark evidence.
Assumptions
- Throughput gain scales required capacity linearly and applies to the whole workload. Real instance sizes, integer fleet counts, memory, network, storage, licensing and service limits can change the result.
- Price discount and throughput gain are user-supplied assumptions; AWS prices vary by region, instance family, operating system, purchase option and date. No pricing feed is used.
- This is a simple compute-cost estimate, not investment advice or a recommendation to migrate. It excludes migration risk, ongoing engineering, taxes, support and non-compute charges.
Technical details: what the fields mean
- Baseline monthly x86 compute spend
x86MonthlyComputeSpendGbp - Current monthly compute-only spend for the workload, expressed in GBP. Exclude unrelated services unless you also include their comparable Graviton cost.
- Graviton price discount vs x86
gravitonPriceDiscountPercent - Percent lower unit price for the comparable Graviton configuration versus x86; enter a discount, not a raw price ratio.
- Measured Graviton throughput gain
measuredThroughputGainPercent - Throughput increase measured on Graviton for the same workload and performance target. A 15% gain is entered as 15.
- One-time migration and testing cost
oneTimeMigrationCostGbp - One-time labor, testing and migration costs in GBP; this calculator does not estimate them for you.
Methodology
Equivalent Graviton spend for the same workload = x86 monthly spend × (1 − price discount) ÷ (1 + measured throughput gain). Monthly savings = x86 spend − equivalent Graviton spend. Simple payback = one-time migration cost ÷ positive monthly savings; if savings are zero or negative, no payback is reported.
Worked example
Measured 15% throughput gain and 20% lower comparable price
Inputs: Baseline monthly x86 compute spend: 4000 GBP/month, Graviton price discount vs x86: 20%, Measured Graviton throughput gain: 15%, One-time migration and testing cost: 12000 GBP
Result: Equivalent Graviton compute spend is about £2,782.61/month; estimated savings are £1,217.39/month. Simple payback is about 9.9 months, before excluded costs.
Practical guidance
- Benchmark representative traffic, memory and latency on comparable x86 and Graviton configurations; do not substitute peak CPU benchmarks for application throughput.
- Use current pricing for the same region, purchase option, operating system and instance capacity, and include migration/testing effort in the one-time cost.
Common mistakes
- Combining a unit-price discount with a throughput gain from different instance families or different benchmark conditions.
- Treating a simple payback period as proof of future savings while omitting memory, network, licensing or operational costs.
FAQs
Does this use current AWS prices?
No. Enter the price discount you observe for comparable current configurations; the calculator does not fetch or store AWS prices.
How should I measure throughput gain?
Benchmark the same application workload at the same latency or service target on comparable x86 and Graviton configurations, then enter the measured percentage change.
Last reviewed: 2026-10-18
A linear, user-input estimate of steady-state compute cost only. It is not a full AWS bill comparison, migration plan, reliability estimate or investment recommendation.
This tool provides an indicative estimate, based on the assumptions stated above.
Theme and tool preferences, along with saved and recently used lists, are stored only in this browser.