A New Chapter in Analytics Engineering: dbt Labs Introduces the Fusion Engine and Official VS Code Extension

On May 28, dbt Labs made a major announcement during their annual Launch Showcase: the public beta release of the new dbt Fusion engine, now available for eligible Snowflake projects. They also introduced the official dbt VS Code Extension, also in public beta.

Together, these releases mark a significant advancement in how teams build with dbt. Fusion is more than a new feature; it's a foundational shift into a new era of analytics engineering, promising enhanced speed, intelligence, and efficiency for modern data teams.

Why This Matters Now

Since the introduction of dbt Core in 2016, the data transformation landscape has evolved dramatically. dbt Labs played a pivotal role in that transformation, embedding software engineering principles into analytics workflows.

But nearly a decade later, the landscape has shifted again. The modern data stack is now standard, cloud compute is cheaper than ever, and AI continues to redefine how we interact with data. dbt Labs recognized that the expectations of data teams have evolved and so must the tools.

To meet these new demands, dbt Labs acquired SDF Labs earlier this year and integrated their technology to create the Fusion engine at the core of dbt.

Inside the Fusion Engine

dbt Fusion fundamentally reimagines how analytics code is interpreted and executed. It introduces three key innovations:

1. Lightning-Fast Performance

Built in Rust, the Fusion engine enables dbt to parse even large projects 30x faster than before. This speed allows for faster iterations, deeper focus, and quicker delivery of high-quality data products.

2. Native SQL Comprehension

Fusion brings local SQL comprehension into the workflow. Instead of merely sending SQL to the warehouse, dbt now understands the code. This enables real-time validation and compiler-like intelligence without querying the warehouse, a major leap in development capability.

3. State-Awareness

Fusion is aware of both your codebase and your warehouse state. It understands which models are truly in need of rebuilding, enabling smarter workflow automation and real-time cost optimization.

What This Means for Data Teams

The implications for data practitioners are substantial. Fusion enables:

• A More Responsive Development Experience

With native SQL understanding, dbt Fusion supports features like intelligent autocomplete, go-to-definition, inline CTE previews, and automatic refactoring, all within the IDE. These enhancements help developers stay in flow and work with greater precision.

• Built-in Cost Savings

By only running models when upstream data has changed, dbt Fusion reduces unnecessary warehouse queries. dbt Labs expects this to result in ~10% cost savings, with further improvements on the way.

• Improved Lineage and Governance

Fusion enables true column-level lineage, enhancing impact analysis and supporting compliance workflows. It lays the groundwork for future features like PII tagging and policy enforcement, key components of any robust AI and data governance strategy.

Now Available: dbt VS Code Extension

To support developers working locally, dbt Labs also launched the official dbt VS Code Extension, purpose-built for the Fusion engine. This extension brings Fusion’s full power into local environments like VS Code or Cursor, offering a seamless and enhanced development experience.

It’s worth noting: the extension is built to work only with the Fusion engine and does not support dbt Core due to the technological dependencies required for these enhanced features.

What’s Available Today

Here’s what’s already live from dbt Labs:

✅ Fusion engine: Available in public beta for eligible Snowflake projects. A rollout notification will appear in the dbt UI for eligible users.

✅ VS Code Extension: Now available via the VS Code Marketplace and ready for local development with Fusion.

✅ State-aware Automation: Automatically enabled for Enterprise plan users on Fusion, offering optimized job runs and warehouse savings.

How to Get Started

If your project is eligible, dbt Labs has made it easy to begin the migration within the dbt UI. If you are not eligible yet, preparation resources are available while support continues to expand.

Non-customers can also begin experimenting today: install dbt Fusion locally, try the VS Code Extension, or explore the Fusion Quickstart, which includes a sample project.

What’s Coming Next

dbt Labs has made it clear: Fusion is the future, and this is just the beginning. The roadmap includes:

• Expanded support for more data platforms

• Column-aware CI/CD for even smarter orchestration

• Built-in data governance, including tagging and policy controls

Beyond the technical upgrades, dbt Labs has signaled that the longer-term implications of Fusion, particularly its role in AI and infrastructure, are significant. For a deeper look into where they’re headed, their CEO, Tristan Handy, has shared more insights in a dedicated post.

dbt Labs has once again raised the bar in analytics engineering. As teams like ours at Datum Labs build cutting-edge data workflows, these innovations open up new possibilities, from faster development to stronger governance. We are watching the evolution of dbt Fusion closely, and we are excited about the possibilities it introduces for the future of scalable, intelligent, and cost-effective data systems.