The dbt Fusion engine shows up at 2025 Databricks Data + AI Summit

This week, we joined over 20,000 data professionals at the 2025 Databricks Data + AI Conference in San Francisco. We wanted to show how dbt Labs and Databricks are working together to improve the quality, trust, and cost of data and AI development. It was a week packed with product announcements, breakout sessions, booth conversations and a whole bunch of fun.

Just in time for Databricks Summit, we announced beta availability of the dbt Fusion engine for anyone using Databricks. We’re continuing to bring the speed and cost savings of our new engine to Databricks. For our Databricks users, this means:

• Significant performance gains: Project parsing and compilation is as much as 30x faster than with dbt Core, dramatically reducing turnaround time from code change to insight.
• Developer intelligence: Fusion’s native SQL comprehension provides real-time feedback and error checking as you write SQL, catching mistakes and providing suggestions instantly—without executing queries against your warehouse. The engine understands Databricks SQL and provides all the same benefits to this ecosystem.
• Cost control: Fusion’s state-aware orchestration understands what’s changed in your data and code, so only necessary models run. This eliminates redundant workloads and helps organizations reduce cloud compute spending by 10% or more.
• Better governance (coming later this year): Precise column-level lineage and richer metadata will soon provide enhanced data governance, making it easier to manage risk, enforce policies, and lay the foundation for safe AI applications downstream.

Databricks’ new capabilities

This week Databricks launched Lakebase, their separation of storage and compute to bring together OLTP data structures for fast processing, while supporting new and emerging AI workloads.Its bones are built on Postgres, and we have a long history of supporting Postgres. dbt Labs is proud to be a launch partner for this innovation. Check out the announcement blog that includes a quote from our CPO, Ryan Segar.

Riot Games was a riot

Marco Garcia from Riot Games joined us this week in San Francisco to share how Riot Games is building reliable, scalable data products with dbt on top of Databricks. They turned to dbt to streamline their development, increase collaboration and leverage built-in testing to deliver trusted data to the business.

Check out Riot Games' presentation at Coalesce 2024 to share how their data platform team paired with analytics engineering, machine learning, and insights teams to integrate Databricks Data Intelligence Platform and dbt Cloud to significantly mature its data capabilities.

We also had a ton of fun

On Tuesday night, we were joined by our partners Thoughtspot, Indicium, and Atlan along with 300 of our closest friends for a relaxed evening at Thriller Arcade. Well, it was relaxed unless you joined the Skee-Ball tournament. Congrats to the winners. Thank you to our sponsors for making a memorable night. I’ll always remember beating my boss, Clarke, at Skee-Ball.

What’s next for dbt and Databricks?

We can’t wait to see what you all take advantage of with Fusion at Databricks. Try it, get used to the blazing fast speed, and tell us what you want next. If you want to see how to deliver AI to your company - join us for our webinar on June 25 & 26 to learn more.

And lastly, join us in October for Coalesce where you can join thousands of enthusiastic data pros and leaders at the analytics engineering event of the year.