Fivetran and dbt Labs Unveil Agent-Ready Data Tools at dbt Summit 2026

At dbt Summit 2026 in Las Vegas, Fivetran and dbt Labs announced new capabilities aimed at making enterprise data ready for AI agents. The companies released dbt v2 and dbt State as generally available, introduced Fivetran Context Layer, dbt Charts, and new dbt Wizard experiences, and outlined an open lakehouse vision for greater flexibility across storage and compute.

The push centers on Open Data Infrastructure, a vendor-neutral architecture letting organizations choose their own storage, compute, data movement, transformation, and visualization technologies. According to Chief Product Officer Anjan Kundavaram, the models, tests, and metrics customers have already built contain the context AI agents need, and the new infrastructure puts that context to work across systems.

dbt v2 is a full Rust rewrite that parses a 10,000-model project up to 10x faster than v1, while dbt State checks warehouse metadata and model SQL to decide whether to build, skip, clone, or defer each run. Virgin Media O2's Gordon Curzon reported a 25% savings on job run time and BigQuery compute costs. RxBenefits' Chris Shepherd cited a 59% reduction in warehouse costs on scheduled jobs, $8,173.23 saved in the first 60 days, and 716,000 reused models.

Fivetran Context Layer, in private beta, unifies data and metadata for LLMs and AI agents using the open source Agents Schema standard. dbt Wizard, in public preview, connects natively to a project and validates changes before they ship, with Wizard CLI, Wizard Desktop, and Explore Mode expanding its reach. dbt Charts, in public beta, defines governed BI as version-controlled YAML alongside dbt models. Lake Compute, built on DuckDB, is in private beta.

Source: dbt Labs Blog