InvestorsCareersClient support
TulanFabric + Agentic AI
Financial Services

Cutting Fabric capacity spend 38% while doubling workloads

A large commercial bank used Loom Ledger to attribute every capacity unit to a data product and owner, then tuned its way to major savings.

Capacity MetricsFabric Admin APIsSparkWarehouseLoom Ledger

The challenge

Fabric adoption spread quickly across risk, finance, and lending. Capacity costs tripled in a year, throttling hit month-end, and finance couldn't tell which teams drove spend.

What we did

Loom Ledger attributed consumption by workload and owner within a week. The top 40 consumers, mostly unoptimized Spark jobs and duplicated refreshes, were tuned or consolidated. Tulan rebalanced schedules around smoothing windows, introduced pause and resume automation for dev capacities, and set up chargeback by line of business.

-38%capacity spend
2xworkloads on the same budget
30 daysto first verified savings
100%of CUs attributed to owners

The outcome

Capacity spend dropped 38 percent while workload count doubled. Month-end throttling disappeared, and each line of business now owns a budget it can see.

“Ledger turned a single scary invoice into a cost per data product. Behavior changed overnight.”
Managing Director, Enterprise Data Platforms

Let's put agents to work on your Fabric estate.

Book a 45-minute working session with a Tulan architect. Bring your hardest data problem; leave with a plan.