Fractional Data Architect
Book a discovery call →

dbt v2 vs v1, What Actually Changes

dbt v2 vs v1, What Actually Changes
dbt v2 vs v1, What Actually Changes

dbt v2, the Rust engine that used to be called Fusion, is GA. Your models mostly stay as they are.

What changes day to day:

  • Speed. dbt’s own example is a 10,000-node project compiling in 17 seconds instead of 70. A 300-model project will feel it less.
  • The engine understands your SQL. It catches a wrong column or type before the warehouse does, and your editor gets real autocomplete and errors.

Read the licensing carefully. Full v2 is free to use but proprietary. “dbt OSS” is the Apache 2.0 subset, and it leaves out the SQL comprehension. The old dbt Core (v1) stays Apache 2.0, and I haven’t seen an end-of-life date for it.

Check your adapter too. Snowflake, BigQuery, Redshift and Databricks are available. Spark and ClickHouse are still in beta.

How I’d plan the switch: run v2 locally next to v1 on one branch and fix whatever the stricter SQL checks find. Settle licensing before anyone wires it into CI. For most teams that’s probably a week of work.

What’s stopping your team from trying v2? Reply with it and I’ll tell you how I’d handle it.

Written by Thomas Nys

Fractional Data Architect helping startups and scaleups build data platforms that scale.

More about Thomas Nys →

Recognise the problem? Let's talk about it.