one big pro is that you do not need to copy data to the dwh. also your transformations and analytics queries reside on the same platform (databricks).
If databricks can cover all the requirementsm compared to snowflakem is hard to tell. Probably there are functionalities that are available in one but not in the other and vice versa.
Costwise it is necessary to analyse your current query consumption on snowflake, and then use a cost calculator to see what it would cost on databricks (serverless/classic dwh).
We decided to get rid of Synapse (which is similar to snowflake but way crappier) and use sql serverless instead.
We calculated that costwise it would be a small saving, but effortwise and maintenancewise a big win.
