Cloud Cost Estimator — AWS vs Azure vs OCI

About this tool

Describe a workload in five numbers — instances, vCPUs and gigabytes of RAM per instance, running hours per month, and total block storage — and the calculator produces a side-by-side monthly and annual figure for AWS, Azure and Oracle Cloud. It models general-purpose Linux compute on demand plus attached block storage, and everything is computed in your browser.

The point is an order-of-magnitude comparison in a few seconds, before anyone opens a vendor calculator or asks an account team for a proposal. It is an estimate for planning, not a quote and not a guarantee: list prices change frequently and your actual invoice depends on region, instance family, commitment discounts, licensing and support tier. Use it to decide which options are worth pricing properly, not as a number to put in a budget. When the shortlist is settled, our cloud migration & administration team can size it against real workload data.

How to use it

Work from your current estate rather than a wish list, then read the three cards:

• Enter the instance count and the vCPU and RAM figures for a representative instance size.
• Set running hours: 730 for always-on, lower for environments you shut down overnight.
• Enter total block storage across all instances, in gigabytes.
• Compare providers — the gap between them matters more than any single figure.
• Confirm the shortlist in the official AWS, Azure and OCI calculators before committing.

Common questions

How accurate is this estimate?

Treat it as a planning figure, not a quote or a guarantee. It applies flat per-vCPU, per-GB-RAM and per-GB-storage rates to whatever you enter. Real invoices depend on region, instance family, committed-use or reserved discounts, licensing and support tier, and providers change their published rates regularly. Confirm anything you plan to commit to in the vendor's own calculator.

What is not included in the figure?

Compute and block storage only. Network egress, load balancers, public IPs, object storage, managed database and Kubernetes services, snapshots and backups, provisioned IOPS, Windows or Oracle licensing, and paid support plans are all excluded. On a real production estate those line items are often a large share of the bill.

Why is the default 730 hours per month?

730 is the average number of hours in a month across a year, so it represents an instance left running continuously. For non-production environments that are shut down outside office hours, use a lower number — roughly 260 hours for twelve hours a day on weekdays only.