The modern data lake: the governed foundation for your AI
November 10, 2026
11 a.m. ET
Virtual
Jamie Cole
Casey Karst
What Open Data Infrastructure is and why it matters for your AI plans.
Why ODI stores data in a data lake.
How landing data in open formats on S3, ADLS or GCS can save up to 28% on warehouse costs.
What a live build looks like: connect a source, land data as files with Iceberg metadata and inspect those files in cloud storage through a catalog.
You're all set, [Name].
Registration closed.
The warehouse solved a real problem 10 years ago: get data into one place, model it, let analysts query it. Now AI tools reach for more data in more ways, and a pipeline tied to one warehouse limits what you can run and how far it scales.
Open Data Infrastructure (ODI) fixes that. You store data once, in an open table format, and Snowflake, Databricks, DuckDB and other engines read from and write to the same copy. The modern data lake holds that copy, and Fivetran lands the data there.
This is session 1 of 3 in the Open Data Infrastructure series. Each session stands alone.
- Session 2 on Nov 17: How to choose the right compute engine for the workload. Save your seat.
- Session 3 on Nov 24: Build a context layer to connect AI to your data. Save your seat.
What Open Data Infrastructure is and why it matters for your AI plans.
Why ODI stores data in a data lake.
How landing data in open formats on S3, ADLS or GCS can save up to 28% on warehouse costs.
What a live build looks like: connect a source, land data as files with Iceberg metadata and inspect those files in cloud storage through a catalog.
How a catalog gives you a unified view of the data in your lake.
Jamie Cole
Casey Karst
Jamie Cole
Casey Karst
The warehouse solved a real problem 10 years ago: get data into one place, model it, let analysts query it. Now AI tools reach for more data in more ways, and a pipeline tied to one warehouse limits what you can run and how far it scales.
Open Data Infrastructure (ODI) fixes that. You store data once, in an open table format, and Snowflake, Databricks, DuckDB and other engines read from and write to the same copy. The modern data lake holds that copy, and Fivetran lands the data there.
This is session 1 of 3 in the Open Data Infrastructure series. Each session stands alone.
- Session 2 on Nov 17: How to choose the right compute engine for the workload. Save your seat.
- Session 3 on Nov 24: Build a context layer to connect AI to your data. Save your seat.