I don’t get it. Sending a query to a remote server is going to be much slower than sending the query to local database. When has Postgres not been enough on its own?
The privacy / security constraints stay the same whether you are branching or not.
Anonymizing data is the biggest part of such a workflow. Most prevalent use case that requires production data is for debugging. I guess there is some value in branching non-prod databases for feature development.
Most security teams do not allow prod data in non prod environments, anonymized or not l.