The future of Terraform CDK

IBM branding and “HashiCorp, an IBM company”

• Many notice the repeated phrase “HashiCorp, an IBM company” and find it awkward, even blame‑shifting.
• Several argue it’s a standard IBM branding/“endorsed branding” pattern, akin to “X by Hilton”, driven by corporate marketing or legal, not engineers.
• An ex‑acquired employee says IBM mandates this phrasing and that people internally tend to dislike it.
• Others note third‑party articles sometimes do this for Red Hat too, but Red Hat’s own site mostly doesn’t.

Alternatives and ecosystem after CDKTF

• People weigh options: plain Terraform/OpenTofu, Pulumi, AWS CDK/CloudFormation, Crossplane, Terranix (Nix-based), Jsonnet/Dhall generating TF JSON, SST (on Pulumi), cdk8s/Yoke for Kubernetes, custom tools like terrars, or even simple templating with Ansible.
• OpenTofu is described as a drop‑in Terraform fork, sharing providers. Recent features include ephemeral values and an enabled meta‑argument, which simplifies conditional resources and migration from count.
• Some hope a CDKTF fork will emerge, ideally targeting OpenTofu; others think OpenTofu itself is the main community fork story.

HCL vs “real language” IaC

• Strong split: some argue HCL’s limited expressiveness is a feature—preventing overly clever imperative logic in a declarative domain.
• Others find HCL “garbage”: weak modularity, awkward conditionals/loops (count hacks), hard refactoring, and poor DRY compared to Python/TypeScript/Rust, etc.
• CDKTF proponents liked using mainstream languages, construct patterns, sharing code between app and infra, and leveraging Terraform providers.
• Critics say CDKTF’s design and codegen pipeline were clunky and under‑resourced; generating HCL/JSON limits what code can actually do at runtime.

Rug pulls, stability, and migration pain

• Sunsetting is seen as a “rug pull” on infra tooling, especially painful because it can affect entire estates.
• Infra engineers emphasize conservatism: migrations are lots of grunt work and avoiding downtime is hard.
• Some complain about very short notice (commit and sunset date aligned), and expect more product retirements.
• Others appreciate that at least some heads‑up and open‑source archiving (Mozilla license) is better than a pure paywall pivot.

Pulumi reception

• Opinions are mixed: a few say “stay far away” without details; many report positive multi‑year production use and good developer ergonomics.
• Benefits cited include use of familiar languages, strong workflows around previews, applies, drift checks, and testable code modules.
• Downsides mentioned: issues with certain statically typed providers, and anxiety that a higher‑level, vendor‑owned layer might one day face a similar fate to CDKTF.