Most of your dbt models aren't in a hurry: How to stop paying for the whole DAG

Every dbt run has a critical path: the small chain of dependent models that determines when the run finishes. It's usually a fraction of the DAG, so why does the whole run cost as if every model were on it? In this session, Espresso AI CEO Ben Lerner is joined by Saravana Kumar of Authentic to unpack how data teams find and close the gap between the compute they pay for and the compute their jobs actually need. Expect real production numbers, what worked and what didn't, and three checks you can run on your own project next week.