Servers described in a file

Infrastructure as code means your servers, networks, databases and everything else are described in files kept in version control, and created by running a tool rather than by clicking in a console.

What it replaces. A person configuring things by hand, remembering roughly what they did, and being the only one who could do it again.

What that costs you today. You cannot recreate an environment reliably. Staging differs from production in ways nobody knows. A change cannot be reviewed before it happens. And when the person who built it leaves, the knowledge leaves with them, which is the risk most organisations here are actually carrying.

What it gives you. An identical staging environment in minutes. Infrastructure changes reviewed like code, with a record of who changed what and why. A disaster recovery plan that is a file rather than a memory. And the ability to destroy an environment when you are not using it, which for a test environment is a substantial saving.

The honest test of whether you have it. Could you lose your entire production environment and rebuild it from what is in version control, with no human recalling anything? If the answer involves somebody remembering something, that memory is a document not yet written.

The cost. Learning the tool, and the discipline not to make ad hoc changes afterwards, which is the harder half. One manual change in a console and your files no longer describe reality.