Why Pulumi matters for platform teams choosing between Terraform and CloudFormation
Platform teams in 2026 face a choice between Terraform, OpenTofu, and Pulumi. Pulumi offers advantages for full-stack engineers by using general-purpose languages like TypeScript and Python to apply software engineering patterns to infrastructure.
The 2026 IaC landscape
Platform teams in 2026 choose between HashiCorp Terraform, the OpenTofu fork, and Pulumi. The industry landscape changed when HashiCorp moved its products to the Business Source License in August 2023. This change prompted the creation of OpenTofu under the Linux Foundation. Pulumi added native support for Terraform and HCL in January 2026, so engineers run existing Terraform modules within a Pulumi program. Terraform maintains roughly 4,800 providers in 2026, while Pulumi manages closer to 1,800. Terraform was launched in 2014 and uses HashiCorp Configuration Language (HCL) or JSON to define infrastructure. Its providers translate code into API calls for platforms like AWS, Azure, or Google Cloud. Terraform state acts as the source of truth by mapping code to real-world resources. The Terraform workflow involves running terraform init to prepare the directory, terraform plan to preview changes, and terraform apply to execute them. AI-native engineering teams are smaller and more full-stack. The engineer who wrote a Stripe webhook often provisions the Lambda that handles it. This reality makes the context-switching friction between TypeScript and HCL a real productivity tax.
Programming logic versus DSLs
Pulumi uses general-purpose languages like TypeScript, Python, Go, and C#. This approach lets engineers apply software engineering patterns to infrastructure. Using TypeScript provides compile-time type checking that catches configuration errors before deployment. A typo in an HCL for_each block fails at terraform apply time, but the same mistake in TypeScript fails during compilation. You can import any library from npm or PyPI to handle complex logic. When a Lambda handler imports a Zod schema for its payload, your Pulumi program imports that same schema to generate API Gateway request validators, which ensures your contract remains consistent across both the runtime and the infrastructure.
Engineers write unit tests using Jest, pytest, or the Go testing package. These tests run in milliseconds in CI. You can mock providers to validate logic without provisioning real resources. To manage complex patterns, you write a class and instantiate it multiple times. For example, a MicroserviceStack class can create 12 microservices with identical VPC, ECR, IAM roles, and CloudWatch log groups. In HCL, you write a module and call it 12 times, but this lacks inheritance. You must handle the Output type carefully because resource properties are Output types, not plain values. You use pulumi.interpolate or the .apply method to compose these values. Using a loop to create buckets is more direct than the gymnastics required for HCL’s for_each. Terraform focuses on setting up infrastructure, while Ansible primarily manages system configuration. Terraform uses HCL, whereas Ansible uses YAML. Terraform manages infrastructure changes using plans and state, but Ansible executes tasks immediately without state tracking.
| Feature | Pulumi | Terraform | AWS CDK |
|---|---|---|---|
| Language Support | TypeScript, Python, Go, C#, Java, YAML | HCL (DSL only) | TypeScript, Python, Java, C#, Go |
| Multi-Cloud | Native (160+ providers) | Native (3000+ providers) | AWS only |
| State Management | Pulumi Cloud, S3, GCS, Azure Blob | Terraform Cloud, S3, Consul, Local | CloudFormation (AWS-managed) |
State management and production stability
Pulumi Cloud provides a managed service with built-in encryption, RBAC, and audit logs. Small teams use the free Individual tier, while the Team tier costs around $100 per month. Organizations with strict compliance needs use self-managed backends like AWS S3 or Azure Blob. These backends require a versioned bucket and a KMS key for encryption. You should never commit unencrypted .env files to a repository. For Azure users, Terraform provides providers like AzureRM, AzAPI, AzureAD, AzureDevops, and AzureStack to manage resources.
State file conflicts occur when two engineers run pulumi up on the same stack simultaneously. The lock prevents this, but stale locks happen if a laptop closes mid-deploy. In these cases, you run pulumi cancel to fix the issue. Secrets deserve special attention because anything you export lands in state. Pulumi encrypts secret values before storing them in state, but values marked secret remain readable to anyone with stack admin access. You should use pulumi.secret() to wrap exports that contain sensitive values.
Teams lacking programming proficiency or those managing simple CRUD resources find Pulumi’s ergonomic benefits outweighed by Terraform’s ecosystem maturity. Will the gap between HCL’s simplicity and Pulumi’s power eventually close? You should evaluate your team’s actual skills and your infrastructure’s complexity trajectory over the next two years, not just today’s needs.