Fix Your Data Platform Costs Before It’s Too Late

How to Cut Data Platform Costs on Snowflake, Databricks & BigQuery with FinOps and Autonomous AI Optimization

Data platform spend has exploded from 5% to 30–50% of the total cloud bill, and most teams have no idea where the waste is hiding. In this FinOps Weekly interview, Kunal (CEO of Unravel Data) breaks down why data FinOps is completely different from cloud FinOps, how query inefficiency silently burns millions, and how autonomous AI agents are closing the execution gap for enterprise data teams at scale.

00:00 – Intro & Guest Presentation
01:02 – Why Data Platform Spend Has Exploded to 30–50% of Cloud Bills
02:05 – Data FinOps vs. Cloud FinOps: Key Differences
03:17 – How Query Inefficiency Drives Massive Cost Waste
04:48 – The Scale Challenge: Petabytes, Millions of Queries & Non-Expert Users
06:55 – Visibility in Business Context: Attributing Costs to Users & Workloads
08:27 – Detecting Inefficiencies & Prioritizing by Financial Impact
10:22 – Why Developers Are Not Incentivized to Optimize Cost
11:07 – Performance and Cost as Two Sides of the Same Coin
12:13 – Explaining Problems in Plain English to Close the Knowledge Gap
14:03 – The FinOps Execution Gap: From Insight to Action
15:02 – Autonomous Enforcement: Human-in-the-Loop vs. Fully Autonomous
16:25 – The Watchdog: Validating Fixes and Protecting SLAs
17:34 – AI-Powered Code Rewriting and Why LLMs Alone Are Not Enough
19:22 – Integrating with GitHub for Shift-Left Cost Prevention
20:39 – Infrastructure Optimization: Dynamic Cluster Sizing in Snowflake & Databricks
22:30 – Interactive vs. Job Clusters: A Hidden 4x Cost Mistake
24:03 – AI Workloads on Data Platforms: The Next Cost Explosion
27:51 – Persona-Based Autonomous Agents for FinOps, DataOps & Quality
31:29 – Building the Full-Stack Data FinOps Product
32:16 – Wrap-Up & Final Thoughts