Databricks: Warehouse-Native vs. Traditional Product Analytics?

One of our customers has a data lake (Databricks). They collect every product event into it. Every event has its table in the Unity Catalog.

It would make perfect sense to run a warehouse-native product analytics solution. We currently send every event to Amplitude, but the costs are enormous and rising.

Does anybody have experience with warehouse native product analytics?