Starburst + Databricks

With the Databricks Data + AI Summit 2025 now firmly behind us, it’s a good moment to reflect.

The Starburst team had a great time at the conference and would like to thank everyone who attended. It was a memorable week, highlighted by social hours and demos that bridged our two worlds–Starburst and Databricks–together. 

What were our highlights of the conference? 

We hoped you’d ask. 

 

AI Summit 2025 recap

Starburst executive speakeasy

On Tuesday, Starburst hosted an exclusive speakeasy soirée at Marianne’s, joined by leaders from Citi, Expedia Group, Edward Jones, EY, Capital One, and more. The night was a special one, where we got to celebrate the customers who make everything that we do worth it. 

Agent Orbit at AI Summit

Even our astronaut-in-chief, Agent Orbit, made an appearance at the speakeasy. We’re always happy when Orbit can get involved, so look out for more cameos in the future. 

Demo: Starburst and Unity Catalog 

This year, it was exciting to see Starburst Galaxy featured at the conference, highlighting some of the innovations that Starburst has launched involving the Unity Catalog. 

On Wednesday, we hit the stage to showcase how Starburst now supports reading from and writing to Unity Catalog—a powerful capability that enables governed, cross-engine interoperability. 

 

Thanks to the Starburst and Databricks teams

A big thank you to our joint session speakers: Michelle Leon (Databricks), Tathagata Das (Databricks), and collaborators Daniel Pan (Starburst) and Jian Chen (Starburst)—for their stellar work in bringing this vision to life. And a special thanks to Starburst account teams Justin Voshell, Robert Windesheim, Kate Dietrich, and Pat Bates for their collaboration as well.

 

Starburst: Choosing how to access your data 

With that in mind, this is also a good time to reflect on something so key to the world of data–choice. 

Starburst is all about choice, specifically the choice of accessing your data where it resides. For example, we make it easy to access data stored in the Databricks Delta Lake format. 

That’s why data access is one of our core values. Alongside data collaboration and data governance, it is one of the key philosophies driving our mission to make data more useful to organizations. So whether your data is housed in Iceberg, Delta Lake, Hudi, or Hive, we’ve got you covered.

Let’s take a quick look at the impact of this for those operating with a catalog and what interoperable compute means for them. 

What is a metadata catalog? 

Metadata catalogs store the metadata needed to make your data architecture run. For this reason, they are often known as metastores. 

Data lakehouses and metadata

Metadata catalogs are particularly important for data lakehouses, which use an open table format like Apache Iceberg, Delta Lake, or Apache Hudi to store additional metadata needed to operate. This metadata is the cornerstone of all data lakehouses, and it’s all stored in a centralized location referred to as the catalog. 

Unity Catalog is one of the catalogs available. 

Compute interoperability 

With your choice of catalogs in mind, Starburst allows you to run workloads – whether it is ETL jobs or ad-hoc queries – no matter where your metadata is stored. Starburst’s interoperability with different catalog options, including Hive Metastore, AWS Glue, Starburst Galaxy metastore, Polaris, Lakekeeper, and now Unity Catalog, allows customers to choose the catalog that best fits their needs. 

The best part? It’s already here. It’s already being used by companies around the world to offer choice around compute costs. 

Unrivalled customer experience

The result is an unrivalled customer experience for compute interoperability. As a global pharmaceutical company, a mutual customer of Starburst + Databricks, put it:

“Unity Catalog’s interoperability with Starburst allows us to position UC as a centralized metadata layer while ensuring changes made in Starburst are instantly available and automatically reflected in UC… It’s a critical step toward minimizing metadata duplication, cutting storage costs, and simplifying data management.”

Bringing choice to data access and compute

Starburst is a choice-driven organization. We understand that not all your data can be centralized, and that’s okay. Using our Icehouse architecture, you can move as much or as little data as you like. The choice is always yours, allowing you to make the right decisions for your data and your organisation. 

More Starburst, more choice

And now, with interoperable compute, you can even use Starburst to run compute workloads with Unity Catalog. Once again, the choice is always yours. 

The result is a win for choice, something Starburst cares deeply about. You can think of this as offering choice at every level, namely: 

• Choice between which catalog to use
• Choice between which compute engine to use 

Excited by this? Good, so are we! We’ll have more information soon, so please stay tuned. In the meantime, you can read about the interoperability of compute in our documentation page.