The next wave of data infrastructure is in the Lake

The modern data stack got you here. It won't get you to where AI agents need to operate.

Vendor lock-in caps what your data can do: AI agents can only see what one platform exposes, adding new tools means another integration project your team doesn’t have time for, and switching compute vendors means tearing out the floor. Open Data Infrastructure (ODI) is the next evolution: not a rip-and-replace, but a leveling-up for the age of AI.

The ODI foundation starts with landing data in open formats like Iceberg, in object storage you control, so you own your data outright. No closed architecture can hold it hostage. That ownership is what makes everything downstream flexible: the right compute engine for any job, tools you can swap without rebuilding pipelines, one source of truth every team and agent can trust.

You'll leave this session with a framework for building open, portable, and AI-ready infrastructure with Fivetran + dbt.