Skip to main content

dbt platform CLI migration

The dbt Fusion engine is taking over the pip install dbt namespace. If you rely on the dbt platform CLI, you need to take action to keep using it without interruption.

What's changing​

Starting September 14, 2026 pip install dbt will install the dbt Fusion engine, dbt's next-generation Rust-based engine, instead of the dbt platform CLI. This is a one-time namespace change on PyPI, not a dbt platform CLI upgrade.

If you install the dbt platform CLI using pip today and don't pin a version, your next pip install dbt or pip install --upgrade dbt installs Fusion instead. Fusion is a different engine: it runs locally against a warehouse connection you configure yourself, instead of running your commands against your dbt platform development environment. Commands like dbt cancel, dbt reattach, dbt environment, dbt sl export, and dbt sqlfluff don't exist in Fusion.

What to do​

If you use the dbt platform CLI and want to keep using it, switch to Homebrew:

brew untap dbt-labs/dbt
brew tap dbt-labs/dbt-cli
brew install dbt

If you have multiple Homebrew taps: brew install dbt-labs/dbt-cli/dbt.

If you can't switch to Homebrew right now (for example, a Docker image that only has pip available), pin your version instead:

pip install dbt==1.0.0.40.18

This keeps working indefinitely, but won't receive further dbt platform CLI updates through pip. Homebrew is the supported long-term path.

General areas of impact​

  • Dockerfiles / CI pipelines: any unpinned pip install dbt step. Search your repos for pip install dbt without a == version pin.
  • Airflow: BashOperator/KubernetesPodOperator tasks that install the CLI into the worker image at build time.
  • requirements.txt / Pipfile / pyproject.toml: any unpinned dbt entry.
  • dbt Power User (VS Code extension): it shells out to the dbt platform CLI internally. Pin your version the same way; the extension itself doesn't need any separate action.

FAQ​

  • Why is dbt Labs doing this? Fusion is the new default dbt engine going forward. Freeing up the dbt name on PyPI for it means new users get Fusion by default, matching how dbt behaves everywhere else (Homebrew, standalone install).

  • Will my pinned dbt platform CLI version stop working? No. Pinned installs keep working. They just won't get new dbt platform CLI releases through pip after September 14. For that, move to Homebrew.

  • How do I tell which one I have installed? dbt --version. The dbt platform CLI prints a version like 0.40.18. Fusion's version string needs reconfirming post-update. At the time this article was published, the binary still internally identifies as dbt-fusion X.Y.Z, which may change.

Was this page helpful?

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

0
Loading