Back to News
dbt Labs Blog

Snowflake Summit Reveals New Path for Legacy-to-Cloud Migrations

The biggest news out of Snowflake Summit this year wasn't just another product launch—it was a fundamental shift in how data teams should think about moving off legacy systems. Fivetran and dbt Labs...

The biggest news out of Snowflake Summit this year wasn't just another product launch—it was a fundamental shift in how data teams should think about moving off legacy systems. Fivetran and dbt Labs have officially merged, and for organizations wrestling with on-premise databases or outdated warehouses, this changes the migration calculus.

Here's what matters for migration specialists: the combined company is betting on an open, portable data foundation. That means your migration strategy doesn't have to lock you into a single cloud provider. Fivetran handles the extraction—pulling from legacy sources like Oracle, SQL Server, or Teradata—while dbt transforms the data once it lands in Snowflake, BigQuery, or Redshift.

The most technically significant announcement is dbt Core v2.0, rebuilt on the Rust-based Fusion engine. For teams migrating large legacy data sets, parse times are up to 10x faster. The old Python engine struggled with scale; the new engine understands SQL natively, catching errors before they hit the warehouse. This cuts down the painful debugging cycles that plague migrations.

Then there's dbt State, which addresses a hidden cost of migration: runaway compute. Instead of rebuilding every table on every run, it checks what's actually changed. Early adopters report 30% less warehouse spend. For a team moving from a fixed-capacity on-prem system to consumption-based cloud pricing, that's not a nice-to-have—it's survival.

Finally, dbt Wizard brings AI to analytics engineering, but with a twist: it respects your project's lineage and governance rules. For migration teams, this means less time correcting AI hallucinations and more time mapping legacy schemas to modern models.

The bottom line? This merger signals that migration isn't just about copying data—it's about rethinking how pipelines operate in the cloud. The tools now exist to make that transition faster, cheaper, and less error-prone.