Prompt
I'm building a multi-cloud setup and need one IaC workflow for AWS and Azure
Latest observation
Oct 25, 2026
OpenAI APIWeb search: off
A good way to get one IaC workflow for both AWS and Azure is to standardize on a tool that supports multiple providers and a single pipeline.
Best practical options
1) Terraform / OpenTofu
Most common choice for multi-cloud.
Why it fits:
- One language/workflow for AWS + Azure
- Large provider ecosystem
- Good for reusable modules
- Easy to plug into CI/CD
- Strong community and enterprise support
Typical setup:
- Separate provider configs for
awsandazurerm - Shared modules for patterns like network, IAM/RBAC, compute, storage
- One pipeline that runs
planandapplywith environment-specific variables
Example structure:
infra/
modules/
network/
compute/
storage/
envs/
dev/
staging/
prod/
Example providers:
provider "aws" {
region = var.aws_region
}
provider "azurerm" {
features {}
}
2) Pulumi
Also a strong option if you prefer a general-purpose language.
Why it fits:
- Use TypeScript, Python, Go, C#, etc.
- Great if your team prefers real programming constructs
- Works well across AWS and Azure
Tradeoff:
- More “software engineering” style
- Slightly less standard in infra teams than Terraform/OpenTofu
3) Crossplane
Good if you want Kubernetes-native infrastructure management.
Why it fits:
- Declarative, cloud-agnostic control plane
- Nice for platform teams running Kubernetes already
Tradeoff:
- More operational overhead
- Better as a platform abstraction layer than a simple IaC replacement
Recommended approach for most teams
If you want the simplest single workflow, use:
OpenTofu or Terraform + CI/CD
- One repo
- One pipeline
- Separate cloud providers in the same codebase
- Modules for reusable patterns
- Environment separation by folders or workspaces
- Remote state per environment
Key design principles
To keep AWS and Azure manageable in one workflow:
- Don’t force identical resources between clouds; map equivalents instead
- Build cloud-specific modules behind consistent interfaces
- Use remote state and locking
- Keep credentials separate per provider
- Standardize naming/tagging across clouds
- Use CI/CD to enforce:
- formatting
- validation
- security scans
- plan approval
- apply gates
Example CI/CD flow
- Developer opens PR
- Pipeline runs:
fmtvalidateplan
- Reviewer approves
- Merge triggers
applyto target environment
If you want the shortest answer
- Best default: Terraform / OpenTofu
- Best if you want a programming language: Pulumi
- Best if you’re Kubernetes-native: Crossplane
If you want, I can give you:
- a sample repo layout for AWS + Azure,
- a Terraform/OpenTofu starter template, or
- a decision matrix comparing Terraform vs Pulumi for your case.