The all-stock merger between Fivetran and dbt Labs, announced last October, is now official. The combined entity will operate under the Fivetran + dbt Labs banner, with George Fraser as CEO and Tristan Handy as President. The stated goal: build the data infrastructure layer for autonomous AI agents.

From a systems architecture perspective, the bet is that agentic workloads—continuous, parallel, machine-speed consumers—place fundamentally different demands on the data stack than human analysts. Latency tolerances shrink; the blast radius of stale or untested data grows. The merger targets this by pairing Fivetran’s managed ingestion and continuous sync engine with dbt’s transformation, testing, and semantic layer. Together, they aim to provide a governed pipeline that moves data from source to agent-ready context without manual intervention.

The first joint releases reflect this architectural focus. dbt Core v2.0 (alpha) open-sources the Fusion engine runtime under Apache 2.0, including a locally installable distribution that unlocks full platform features via terminal login. dbt State introduces a caching layer that rebuilds only changed assets, targeting a 30% reduction in warehouse compute costs. dbt Wizard offers context-aware SQL generation grounded in project lineage, tests, and metrics. The Agents Schema standard proposes a single warehouse schema as a shared context layer for agents, storing metric definitions and lineage in plain SQL tables, compatible with any engine or ingestion tool.

Early customer references from Zendesk, Inova Health, and DocuSign cite reduced time-to-trusted-data and improved scalability for analytics and AI workflows. The combined platform will be demoed at Snowflake Summit 2026.

Source: dbt Labs Blog