Cloud Cost Calculator

A cloud bill comes down to four buckets: compute, storage, data transfer, and everything else the platform charges for. This calculator adds those four, takes a commitment discount off the total, and shows the monthly and annual figures side by side.

Every major provider sells the same trade: a promise to keep spending, in exchange for a lower unit price. Amazon Web Services calls them Reserved Instances and Savings Plans, Google Cloud calls them committed use discounts, Microsoft Azure calls them reservations. The terms differ, and so does the size of the reduction, which is why the discount here is an input rather than a constant baked into the maths.

The defaults total 3,000 a month gross. At 25% the net is 2,250 a month, or 27,000 a year, and compute makes up 66.7% of that gross. The compute share is worth watching in its own right, because it says which lever actually moves the bill.

The arithmetic is deliberately plain. What it does not model is utilisation, and utilisation is where a lot of cloud overspend hides: a commitment covering a machine that idles half the month is still billed in full. Cost management as a discipline exists largely to close that gap, and the FinOps Foundation framework and the cost optimisation pillar of the AWS Well-Architected Framework both set out the practice in detail.

Quick example

With compute of 2,000, storage of 500, bandwidth of 300, other services of 200 and a 25% commitment discount, the gross is 3,000 and the result is 2,250.00 net per month. The rows underneath show 27,000 annual, 750 saved by the discount, a 66.7% compute share, and the 3,000 gross the discount was taken from.

Which inputs matter most

The discount percentage is the only input that changes the ratio between gross and net; the four spend lines change the gross itself, and each one moves the result by its own value less the discount. At 25%, adding 100 of storage adds 75 to the net monthly and 900 to the annual figure. Because the annualisation is a flat multiplication by twelve, any error in the monthly figure is magnified twelvefold in the annual one.

What is happening under the hood

Add the four spend lines to get the gross. Multiply by one minus the discount rate to get the net. Multiply the net by twelve for the annual figure.

Worked example with realistic figures

Take a mid-scale application: compute at 4,500 a month, storage at 1,200, bandwidth at 800, and 500 across managed databases, monitoring and support. The gross is 7,000, and a negotiated 30% commitment discount takes the net to 4,900 a month, which annualises to 58,800.

Now let storage grow to 2,000, the sort of drift that arrives quietly when nobody prunes old snapshots. Gross climbs to 7,800, net to 5,460, annual to 65,520. An 800 a month storage increase costs 6,720 a year after the discount, and none of it shows up as a decision anyone made.

Common scenarios where this metric matters

Different teams reach this from different angles. A company scaling headcount or data volume uses it to project spend forward. A team weighing a move between providers uses it to line up total cost side by side. A budget holder needs a defensible annual line item. And an engineer usually arrives with a narrower question: whether one component, bandwidth during a traffic surge being the classic case, has grown out of proportion to the rest.

What the result captures and what it does not

The calculator returns the sum of the service charges after a single discount is applied, plus the annualised projection. Everything else is outside it.

Prices move, service tiers change, and usage rarely holds flat month to month; a constant monthly figure is an assumption, not an observation. One-off costs such as migration, training and initial setup are excluded, as are taxes and support-plan minimums in some billing arrangements. Efficiency metrics like cost per transaction or per active user are not modelled at all, and those are usually the numbers that tell you whether spend is healthy. The single discount rate is the largest simplification: real commitment structures vary by service, by region, by term length and by how much of the workload the commitment actually covers, so a partial reservation across a mixed estate will not reduce to one clean percentage.

Educational illustration note

This calculator is built for estimation and for learning how the pieces relate. The output reflects the inputs given and illustrates how gross spend, commitment discounts and annualisation connect. It sits alongside invoice review and cost-management tooling rather than replacing either.