...

dbt Fusion FAQ: All You Need to Know About the New dbt Engine

Márton Hubay
5 June 2025
Read: 6 min

dbt Labs has introduced dbt Fusion, a new engine designed to modernise how we work with dbt projects. You might be deciding whether it is time to try Fusion, or simply looking to understand what is changing. This dbt Fusion FAQ article addresses the most common questions.

Which data platforms does dbt Fusion support?

In the coming weeks, dbt Labs will roll out support for Snowflake, Databricks, BigQuery, and Redshift (in that order). Athena, Postgres, Spark, and Trino are also expected to be supported in future.

Which dbt packages does Fusion work with?

The most widely used dbt packages (dbt_utils, audit_helper, dbt_external_tables, dbt_project_evaluator) are already compatible with Fusion. Some external packages may not work immediately, but dbt Labs intends to collaborate with package maintainers to ensure compatibility.

My project contains a functionality deprecated in dbt Core v1.10. Can I use Fusion?

No, this is currently blocked due to deprecated functionality. You will need to address this using the dbt-autofix script or the workflow available in dbt Studio.

My project contains Python models or uses a not-yet-supported adapter, can I use Fusion?

No, not at the moment. dbt Labs is still working on support for these features. As a temporary workaround, you will need to remove any Python models from your project.

Can I try Fusion if some features I use are not supported?

Yes. If your project parses successfully, unsupported models can be skipped. This allows much of your DAG to be tested using Fusion, so you do not need to make an all-or-nothing choice.

What features are not supported at this time?

Python models

These are completely unsupported for now. Fusion cannot parse them or extract ref() calls or config() values.

The reason is to avoid building models out of DAG order.

The good news is that support is planned. You can follow progress in the issue tracker.

Custom and advanced materialisations

Fusion currently supports:

  • Standard materialisations: table, view, incremental, ephemeral
  • Materialisations used by snapshots, seeds, and tests

Not yet supported:

  • microbatch strategy for incremental models
  • Iceberg tables
  • Materialised views or dynamic tables
  • Stored test failures
  • Custom materialisations

Workaround

Use CLI filters to exclude unsupported materialisations. For example:

dbt build --exclude config.materialized:my_custom_mat

dbt build --exclude config.incremental_strategy:microbatch

How complete is the current feature set?

There are over 400 documented dbt features. Fusion already supports approximately two thirds of these.

The remaining features are actively tracked with open issues and milestones in the dbt Fusion GitHub repository.

What notable features are still in development?

Some key features currently in progress include:

  • Exposures;
  • A new stable logging system;
  • A redesigned local documentation experience to replace dbt Docs;
  • Programmatic invocations;
  • Model governance, including contracts, constraints, access control, and deprecation dates;
  • Advanced CLI commands such as dbt clone, state:modified.subselector, --empty, and others.

Will all features behave exactly as they do in dbt Core?

Not necessarily.

Some features will use Fusion’s SQL comprehension to offer improved or redesigned experiences. Others, such as logging and events, rely heavily on Python and may be completely reworked to better fit Fusion’s architecture.

More information is available in dbt’s documentation.

Licensing

Full information about dbt Fusion licensing is available here.

Is the licence for dbt Core changing?

No.

Will dbt Labs continue to support and maintain dbt Core?

Yes.

Do I need to be a paying dbt customer to use Fusion?

No.

Do I need to be a paying dbt customer to access some enterprise features of Fusion?

Yes.

What licence do I need for dbt Fusion?

Fusion includes source-available (ELv2), open-source (Apache 2), and proprietary components. You may choose to use only the source-available code if you prefer.

I am a dbt Core user. What do I need to know about Fusion?

Provided you are not offering Fusion as a hosted or managed service to a third party, you can consider using Fusion in almost the same way as you use dbt Core today.

You may use Fusion internally within your own business, free of charge and without restrictions. This includes supplying transformed data to other customers, as long as the transformed data itself is the product.

Is Fusion replacing dbt Core?

No.

The dbt framework consists of a language and an engine.

The dbt language (the code you write in your dbt project) has become an industry standard. dbt Labs has developed this language over the past decade through dbt Core, with contributions and input from the community.

The dbt engine is the underlying technology that compiles dbt projects, executes transformation graphs, and produces metadata. Fusion is the next-generation engine for dbt. It is entirely new software, written in Rust rather than Python, which is used by dbt Core. From its initial release, Fusion will be available under the Elastic Licence v2 (ELv2), allowing most existing dbt Core users to upgrade to Fusion as their engine.

dbt Labs remains committed to maintaining and expanding the dbt language across both Core and Fusion engines. You can expect new dbt Core versions (such as v1.11, v1.12, and so on) that support new language features and fix issues to maintain compatibility. dbt Labs will also continue to evaluate and integrate contributions from the community.

The dbt Fusion engine is the foundation for future investment and innovation in dbt’s engine. dbt Labs aims to minimise disruption for existing users while enabling as many as possible to upgrade to Fusion.

Fusion will eventually support the full dbt language and the majority of existing dbt projects. It is expected to deliver a better development and deployment experience for all users.

You do not need to be a paying dbt customer to use Fusion.

I am self-hosting Airflow to run dbt Core for my own business. Can I start running Fusion instead?

Yes.

I am an end user of a hosted or managed service that provides dbt Core to me today. Can I expect to immediately upgrade to Fusion?

No. The hosted or managed service provider must negotiate a licensing agreement with dbt Labs before offering Fusion as a hosted or managed service. In the meantime, you may continue using dbt Core or try Fusion independently.

dbt fusion FAQ about components and licences

I am an end user of a hosted or managed service that provides dbt Core to me today. If I install Fusion instead, am I violating the licence?

No. As a user of a hosted or managed service rather than its provider, you are not in breach of the licence.

If you are a provider of a hosted or managed service intending to run Fusion, or if you have any other questions not addressed here, please contact dbt Labs at license@dbtlabs.com.

My team is using dbt to develop, transform, and test data assets in the cloud DW. Those assets are powering real-time dashboards for colleagues and embedded product experiences for customers. Is this allowed?

Yes, this is allowed. We strongly support the “dbt way” of analytical work, often referred to as the ADLC, and believe that upgrading to Fusion will help your team carry out this work more effectively.

I am a contractor or consultant setting up dbt for my clients to use internally. Can I continue doing this?

Yes, this is allowed under the Elastic Licence v2 (ELv2), as you are not offering Fusion as a hosted or managed service.

I want to use Fusion to transform my customers’ data into assets that power my app. The app shows Fusion data test failures directly to customers to alert them to data quality issues in source data systems. Can I continue doing this?

Yes, this is allowed under the Elastic Licence v2 (ELv2), as your customers do not have direct access to or control over Fusion.

Are the VS Code extension and its features free to use during the beta phase only? Is the goal for Fusion to remain free and accessible in VS Code in general?

The dbt Fusion engine is free and distributed under the Elastic Licence v2 (ELv2). You may use the binary or publicly available source code for your analytics workloads at no cost and without having to engage with dbt Labs.

The dbt VS Code extension is proprietary, but it will remain free for teams of up to 15 users. The goal is to encourage broad adoption and use.

What if you have more than 15 users?

If your team exceeds 15 people, you will need a commercial licence. The VS Code extension is expected to be monetised through broad licences that allow all dbt users at your organisation to access it.

These licences will likely come with significantly lower inferred per-user costs compared to dbt Enterprise.

The aim is to make it straightforward and affordable for teams of all sizes to use the new dbt VS Code extension.

Does the Fusion Engine currently support Snowflake Dynamic Tables as a materialisation?

Not yet, but it is on the list of planned features.

You can follow progress here.

Does the Fusion engine currently support Snowflake User-Defined Functions (UDF)?

No, Fusion is currently unable to compile user-defined functions (UDFs), which define custom logic in Snowflake and other data warehouses. If you are using a UDF in a model, you will need to set static_analysis: off.

Native support for UDF definition and compilation is planned for a future release.

Read more on the launch of the dbt Fusion engine.

Looking to update to Fusion? We are here to guide the transition. Get in touch at contact@infinitelambda.com.

More on the topic

Everything we know, we are happy to share. Head to the blog to see how we leverage the tech.

10 Questions to Ask Before You Migrate off Legacy ETL
10 questions before legacy etl migration
10 Questions to Ask Before You Migrate off Legacy ETL
I spent about ten years on the buyer side at Johnson & Johnson and AstraZeneca, so I have sat through my share of migration pitches....
12 November 2026
What are Claude Dashboards and is this the death of BI?
Claude Dashboards
What are Claude Dashboards and is this the death of BI?
Anthropic released a new artefact referred to as Claude Dashboards, alongside a short introductory video that showcases how the feature works and how teams can...
15 October 2026
How to plan a predictable ETL migration
plan a predictable ETL migration
How to plan a predictable ETL migration
In the world of data migrations, boring is good. A well-planned project tends to be wonderfully uneventful, making for a predictable ETL migration from start...
9 October 2026
Patterns vs Platforms: What Google Cloud Platform Projects Teach Us about Trust and Speed
Patterns vs Platforms - What Google Cloud Platform Projects Teach Us about Trust and Speed
Patterns vs Platforms: What Google Cloud Platform Projects Teach Us about Trust and Speed
Every client engagement starts the same way: we learn as much as we can about the client so we can support them in reaching their...
2 September 2026
Infinite Lambda Wins Snowflake Partner of the Year Award for ASEAN 2026
Infinite Lambda Wins Snowflake ASEAN Partner of the Year Award 2026
Infinite Lambda Wins Snowflake Partner of the Year Award for ASEAN 2026
Infinite Lambda has won the 2026 Snowflake Partner of the Year award for ASEAN. It is Snowflake's top regional award, given to the partner that...
27 August 2026
Infinite Lambda Achieves ISO 27001 Certification
ISO 27001 certified
Infinite Lambda Achieves ISO 27001 Certification
Infinite Lambda has achieved ISO 27001 certification, the leading international standard for information security management. The certification was awarded by LRQA following an independent audit...
17 July 2026

Everything we know, we are happy to share. Head to the blog to see how we leverage the tech.