Back to Insights
Platform Engineering

Infrastructure from Intent: Why Terraform is Legacy Code

Author
elitics.io Editor
Jul 12, 2026 5 min read
Infrastructure from Intent: Why Terraform is Legacy Code

For the last decade, we treated infrastructure as a separate discipline. You wrote your Node.js app, and then you wrote 500 lines of HCL (Terraform) to define VPCs, Subnets, Security Groups, and IAM roles just to deploy it. In 2026, this duplication is dead.

IfC (Infrastructure from Code)

The concept is simple: The infrastructure requirements are already implicit in your application code.

If you import a `Queue` object and call `queue.push()`, the platform should automatically provision an SQS queue or a Redis instance. You shouldn't have to declare it twice.

main.tf (Legacy)

resource "aws_sqs_queue" "q" {

name = "orders"

visibility_timeout = 60

}

resource "aws_iam_role" "..." {...}

// ... 200 more lines

api.ts (2026)

import { Queue } from '@elitics/cloud';

const orders = new Queue('orders');

// AI detects this line & provisions SQS + IAM

await orders.push(payload);

The "Compiler" for Cloud

At elitics.io, we use intelligent compilers that parse your AST (Abstract Syntax Tree). They see that you are accessing a `bucket`, so they create the S3 bucket. They see you are running a cron job, so they create the EventBridge rule.

No More Drift

Because the code *is* the infrastructure definition, it is impossible for your infrastructure to drift from your application logic. They are atomically coupled.

Stop hiring DevOps engineers to write YAML. Hire Software Engineers who build systems. Let the compiler handle the plumbing.

Enjoyed this perspective? Share it with your team.

Infrastructure from Intent: Why Terraform is Legacy Code | elitics.io Insights