toolbox
Home / Startups & Digital Business / Spot Instance Savings Plan Calculator

Spot Instance Savings Plan Calculator

The Spot Instance Savings Plan Calculator estimates how much a spot-compute workload saves compared with running the same base runtime at an on-demand rate. It accounts for the lower spot hourly rate, extra retry hours caused by interruptions, and optional fixed costs tied specifically to the spot strategy. The output includes dollar savings, percentage savings, and the spot effective hourly cost after overhead. This helps teams test whether a headline spot discount survives real workload friction.

The calculator does not assume that spot capacity is always available or that interruption overhead will be stable. Enter rates and retry assumptions that match the instance pools and workload you are evaluating. For critical services, the economic result should be considered alongside reliability, fallback capacity, engineering complexity, and missed-deadline risk. A positive modeled savings percentage is a cost signal, not a recommendation to move an unsuitable workload to spot capacity.

Inputs

hr
$/hr
$/hr
hr
$
Result
—
Estimated spot savings rate
On-demand baseline—
Spot strategy cost—
Estimated savings—
Effective spot cost per base hour—

1. Enter base workload hours
Use the productive runtime that would be required regardless of pricing model.

2. Set the on-demand rate
Enter the comparable hourly price for the same or equivalent compute capacity.

3. Enter the average spot rate
Use the expected or observed average hourly cost for the spot capacity.

4. Add retry hours
Include additional billable spot runtime caused by interruptions and repeated work.

5. Add spot-specific fixed costs
Include checkpointing, orchestration, or other costs only when they are incremental to the spot strategy.

6. Review both savings and effective rate
The effective cost per base hour shows how interruption overhead changes the apparent spot price.

On-demand baseline = Base workload hours × On-demand rateSpot strategy cost = (Base workload hours + Retry hours) × Spot rate + Spot-specific fixed costsSavings = On-demand baseline − Spot strategy costSavings rate = Savings ÷ On-demand baseline × 100

Where:

  • Base workload hours — productive compute hours required by the workload
  • On-demand rate — hourly cost of the comparison capacity
  • Spot rate — average hourly cost of spot capacity
  • Retry hours — additional spot runtime caused by interruptions
  • Spot-specific fixed costs — incremental non-hourly costs of the spot strategy

Assumptions: The on-demand and spot capacity are equivalent for the workload, and retry hours are the main additional runtime cost introduced by spot interruptions.

What the result means

A positive percentage means the modeled spot strategy costs less than the on-demand baseline after retry time and the entered incremental fixed costs.

Reliability and delivery risk are outside the savings formula and should be evaluated separately for production or deadline-sensitive workloads.

Given

  • 4,000 base workload hours
  • $0.40 on-demand rate
  • $0.15 average spot rate
  • 320 retry hours
  • $120 spot-specific fixed cost

Calculation
On-demand baseline = 4,000 × $0.40 = $1,600
Spot strategy cost = (4,000 + 320) × $0.15 + $120 = $768
Savings = $1,600 − $768 = $832
Savings rate = $832 ÷ $1,600 × 100 = 52.00%

Result
52.00% estimated savings

After retry time and incremental spot costs, the modeled workload still costs $832 less than the on-demand baseline.

Why use base workload hours for the on-demand baseline?

It keeps the comparison focused on the productive work that must be completed. Retry hours are treated as spot-specific overhead rather than duplicated on both sides.

What if on-demand also has retry or failure overhead?

Then include only the incremental difference attributable to spot, or build separate detailed cost models for both approaches.

Can the savings rate be negative?

Yes. If spot rates, retry hours, or supporting costs are high enough, the modeled spot strategy can cost more than the on-demand baseline.

Should engineering labor be included in fixed costs?

You can include allocated labor when the analysis is intended to capture total operating cost. For a compute-only comparison, leave labor out and document that scope.

How is this different from the spot unit cost estimator?

This calculator compares total spot strategy cost with an on-demand baseline. The unit cost estimator divides spot workload cost by completed output units.