Back to Devoured - June 04, 2026
DEVOURED
dbt Core v2 is here: still open source, now rebuilt for what's next

dbt Core v2 is here: still open source, now rebuilt for what's next

Data dbt Labs
dbt Core v2.0 open-sources the Rust-based Fusion engine runtime under Apache 2.0, promising faster parsing, Parquet artifacts, and a unified foundation with Fusion.
What: dbt Core v2.0 alpha release is now built on the same Rust-based runtime as dbt Fusion, with much of Fusion's code open-sourced under Apache 2.0. This brings significant parse time improvements, new Parquet artifacts, revamped local documentation, ADBC-powered adapter building, and simpler installations, while Fusion remains the recommended binary for most users.
Why it matters: This move signifies dbt Labs' commitment to a more performant and unified dbt framework, leveraging Rust and modern data technologies, and clarifying the roles of dbt Core (fully open-source) and dbt Fusion (enhanced binary) in the ecosystem.
Takeaway: If you use dbt, consider migrating to dbt Core v1.12 first and running `dbt parse --use-v2-parser` to prepare for the v2.0 upgrade, or install dbt Fusion for the best free CLI experience.
Deep dive
  • Unified Foundation: dbt Core v2.0 merges with the dbt Fusion engine's Rust-based runtime, placing much of the previously ELv2-licensed Fusion code under the Apache 2.0 license.
  • Key Features: Includes significant parse time improvements for large projects, a tighter language specification, new high-performance Parquet artifacts for data, a revamped local documentation experience, streamlined adapter building via ADBC and Arrow, and simpler installations by reducing Python virtual environment complexities.
  • Distributions: dbt Fusion will continue as the recommended precompiled binary for most users, offering additional free and premium features, while dbt Core v2.0 serves teams requiring fully open-source code or custom builds.
  • Strategic Shift: This change aims to consolidate development efforts on a single, more performant Rust-based engine, ensuring that all users benefit from ongoing innovations.
  • Backward Compatibility: The existing Python-based dbt Core v1.x will still be available, but future capabilities will increasingly require upgrading to v2.x.
  • Migration Path: Users are advised to upgrade to dbt v1.12 first, which includes the Fusion-powered parser for compatibility checks, before moving to v2.0.
Decoder
  • dbt (data build tool): An open-source framework that helps data analysts and engineers transform data in their warehouse by writing SQL SELECT statements.
  • dbt Core: The open-source command-line interface for dbt.
  • dbt Fusion: An enhanced version of the dbt CLI, previously under ELv2 license, that includes performance improvements and additional features, now sharing its Rust runtime with dbt Core v2.0.
  • Apache 2.0 license: A permissive free software license by the Apache Software Foundation.
  • ELv2 license: The Elastic License v2, a non-open source license often used for software where commercial offerings exist.
  • Parquet artifacts: Data stored in the Apache Parquet column-oriented data file format, known for efficient data compression and encoding schemes.
  • ADBC (Arrow Database Connectivity): A specification for a set of high-performance APIs for database connectivity using Apache Arrow for data transfer.
Original article

Full article content is not available for inline reading.

Read the original article →

Explore all 55 articles in today's digest

Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi