One year of Fusion: powering the next era of data and AI

In 2025, dbt continued to evolve beyond a transformation framework, with the dbt Fusion engine taking root as a new foundation for how dbt projects are built and run. Through faster compilation, smarter execution, and richer metadata, Fusion is beginning to change the day-to-day experience of working in dbt.

That progress has been shaped through sustained development and close collaboration with the dbt community. The milestones below show how Fusion has taken shape over its first year — and set the stage for a hands-on workshop where you can see what it looks like to migrate real dbt projects to Fusion and apply these capabilities in practice.

January 14, 2025

SDF joins dbt Labs.

A major investment in a Rust-powered compiler designed for the next era of data.

What is Mom’s Flower Shop?

Elias DeFaria, Fusion Product Lead and SDF co-founder, based Fusion’s starter project on his mom’s LA flower shop, where she used simple tools to answer real business questions. The example reflects dbt Labs’ belief that data modeling should be accessible and practical for everyone.

January 6, 2025

The work begins.

Ahead of the January 14 SDF announcement, the first commit marked the start of development on the Fusion engine.

  • 8,611 commits from Jan 6, 2025 - Jan 14, 2026
  • 600+ issues addressed, improving stability and real-world reliability

May 28, 2025

Fusion and the dbt VS Code Extension enter public beta.

The dbt Fusion engine reaches public beta, alongside the public beta release of the official dbt VS Code Extension, bringing more intuitive local development and meaningful developer experience enhancements to the dbt community.

  • ~30× faster parse times
  • 104,200 VS code extension downloads to date
Download dbt VSCode extension

June 3, 2025

Powered by Fusion program launches.

Snowflake embeds Fusion directly into their platforms, bringing its speed and intelligence to more ecosystems and more data teams.

Oct 14, 2025

State-aware orchestration private preview

Fusion introduces smarter, data-aware execution: build when code or data changes, reuse results when it doesn’t. Teams gain faster pipelines, lower compute costs, and predictably fresher data.

  • EQT’s Coalesce talk shows how Fusion helps them achieve materially faster runs, measurable savings, and more frequent data delivery.
  • ~40% model reuse and 30% warehouse compute cost savings for projects using state-aware orchestration
  • Our community member, Zoltan Toth releases the dbt Fusion mini course on Udemy
Read more about state-aware orchestration

Dec 8, 2025

Adopted the ADE-bench framework.

Highlighting our commitment to transparent, rigorous, industry-standard performance benchmarking for AI. Early results show that agents complete real-world dbt tasks faster and more efficiently on Fusion — requiring fewer steps, lower compute cost, and less trial-and-error.

Watch the ADE-bench webinar

Dec 14, 2025

Stability & parity milestones on the path to GA.

Fusion now supports broad dbt Core parity, expanding the range of projects that can run on Fusion and accelerating adoption across customers and partners. This growing real-world usage continues to shape improvements to performance, developer experience, and day-to-day reliability as we move toward GA.

  • 25+ preview releases shipped
  • 450+ weekly average projects running on Fusion
  • 16 editions of Fusion Diaries with 25,120 subscribers staying up to date on the latest developments
Fusion readiness checklist

The garden grows because many hands tend it.

Thank you to our top community contributors

Top discussions

  • 1,957 members in #dbt-fusion-engine
  • 312 community-opened bug reports and feature requests
Join the community

Fusion marks its first year strengthening speed, reliability, and the open data ecosystem.

One year in, the dbt Fusion engine is already strengthening speed, reliability, and the open data ecosystem, and it's opening up new ways to think about how dbt projects are developed and operated. As teams begin to adopt Fusion, practical questions naturally follow: What changes when you migrate? What stays the same? And where do the biggest gains show up in real workflows?

That's exactly what our upcoming workshop is designed to answer. In a guided, hands-on session, you’ll walk through migrating a dbt project to Fusion, see how faster feedback and smarter execution surface in the IDE and runtime, and understand how Fusion fits into existing orchestration and deployment patterns. You’ll leave with a clearer mental model and concrete next steps for when and how to bring Fusion into your own projects.

Join the migration workshop

About these metrics: Metrics on this page are based on public GitHub activity from the dbt-fusion repository (excluding the initial binary repo commit), internal benchmarking, and anonymized customer usage data from 2025. Performance claims reflect measured results from representative projects and preview programs, with links provided where public case studies are available.

EQT

By running Fusion in production, we reduced our project runtime by 60%, which resulted in a 45% reduction in compute costs for that project. What’s more: our end user experience vastly improved.

Jonny ReichwaldVP Analytics Platform, EQT Group