dbt Core vs dbt Cloud After v2: Licence and Support Options

dbt (data build tool) now comes in four forms, and the name dbt Core no longer covers the newest one. Since 15 September 2026, dbt v1 keeps the dbt Core name, dbt v2 ships as dbt OSS and as the full dbt binary, and the hosted product is the dbt platform. Each has its own licence and support window, which matters if you run dbt in production or need someone to support it. AceMQ offers independent dbt support and assessment; this article sets out the facts from dbt Labs’ licensing FAQ and versions page, read on 4 November 2026.

The four forms of dbt and their licences

What changed in 2026

On 1 June 2026 dbt Labs moved the code that had been published under the Elastic License v2 into the dbt repository under Apache 2.0, and introduced the dbt Product Licensing Agreement for the proprietary binary. On 15 September 2026, with dbt v2 generally available, it settled the names: the full distribution is dbt, the Apache 2.0 subset is dbt OSS, and dbt Core is reserved for v1. Older articles call the hosted product dbt Cloud; dbt Labs’ current documentation calls it the dbt platform.

dbt Core vs dbt Cloud: what the hosted product adds

dbt itself is free and needs no account: dbt Labs says you do not need a dbt platform account to use dbt or its premium local features. What a local install does not give you is a place to run it. Production needs a scheduler, a pull-request check, secrets handling, alerts and somewhere to host documentation. Teams build that around Apache Airflow, GitHub Actions or cron, or buy it as part of the dbt platform, where dbt Labs lists orchestration among the hosted services. The choice is about who runs the surrounding machinery, not about what a model can do.

Who can support each form

Support is a licence question as well as a skills question.

• dbt Core v1 and dbt OSS. Apache 2.0 permits anyone to use, copy, redistribute or modify the code, provided licence notices, attribution and trademarks are preserved.
• The full dbt binary. Self-managed use is permitted. The redistribution terms apply to anyone who offers dbt to third parties as a hosted or managed service, who must let their end users use the full capabilities of dbt, including dbt platform features gated behind login, through standard documented sign-in flows. dbt Labs’ FAQ names a consultant setting up dbt for a client’s internal use as permitted, provided each client’s deployment is standalone and under the client’s control.
• The dbt platform. Support runs through dbt Labs under your agreement with it.

This is a summary of what dbt Labs publishes, not legal advice. For an unusual case, dbt Labs invites licensing questions at license@dbtlabs.com.

Which one to run

• Stay on v1.12 while it is in active support if the project is stable and your adapters and packages are pinned. dbt Labs expects critical fixes for the final 1.x minor for several years and has set no cutoff date.
• Move to v2 when you want a line that keeps receiving new minor versions after 2027 or the full binary’s features. dbt follows semantic versioning, so a new major version can break projects and deprecated functionality stops working. Treat it as an upgrade project.
• Choose the dbt platform when you would rather buy scheduling and the editor than run them yourself.

What to do now

• Run dbt --version in every project and write down the form and the minor.
• If you are on v1.11, plan the move to v1.12 or v2 before 18 December 2026, when critical support ends.
• If you use the dbt platform on v1.3 to v1.7, note that dbt Labs removes those versions from the platform on 31 January 2027.
• If your policy requires an Apache 2.0 runtime, install dbt OSS and not the full dbt binary.
• Upgrade packages and adapters on a copy of the schema first, and run old and new side by side before cutting over.

A written assessment of your projects covers these points in one report. See the AceMQ dbt page for scope, and Apache Airflow consulting if the scheduler is the part that needs work. AceMQ does not write dbt models or build pipelines.

• dbt Labs: dbt Licensing FAQ
• dbt Developer Hub: About dbt versions