Here's my 2c as a VBR developer. I'm not gonna touch on any internal/organizational questions, just technical side of things.
While I can agree with some of your arguments (mainly regarding SQL server express limitations and general lack of edition constraints in Postgres), and while I think Postgres is a great DBMS with some fantastic programmability features, here's a few thoughts on what kind of challenges we'd have adding postgres support:
Adding Postgres as a second supported DBMS alongside SQL Server is an option that can be benificial for some customers, but it would not achieve anything on your list (" less differences in the field, less support costs due to db specific issues, supported backups and procedures, optimizations done for all"), rather the opposite. We'd end up supporting 2 different DBMS with slightly but annoyingly different SQL syntax (so essentially having to write and test all DB-side code twise), their respective performance quirks, possible corner cases in C# client support etc.
So the only way to achieve anything in terms of configuration uniformity and development/testing speed would be to swap SQL Server for Postgres and drop support for SQL Server altogether.
Dropping SQL server support - leaving aside the fact that SQL Server is a fantastic DB and many of our developers have a lot of experience with it etc. etc. - is hardly an option. Besides the amount of DB-specific code we'd need to rewrite and retest, migrating critical software for 250k+ customers between DBMSs sounds like a recipe for a lot of problems and confusion for customers: consider that many of our bigger customers host Veeam DBs on a dedicated fully lisenced SQL Servers alongside other production/infra DBs, have a dedicated SQL Server DBA and so on. At certain scale, hosting a DB on a VBR server may become a bottleneck, so these customers would now need to set up and tune their own Postgres servers.
I obviously can't speak for our Product Management team, but from development point of view adding support for Postgres not a small undertaking that would require very serious reasons to commit to.
On a bright side, SQL Server Standard edition is not really expensive, and it's all you really need for VBR even at scale.
Besides, Microsoft lately has been doing a lot of good stuff in terms of licensing, enabling many Enterprise-only features in all editions starting with SQL 2016 SP1 (https://blogs.msdn.microsoft.com/sqlrel ... -released/
). We don't use many of those featues yet, but may start in future, now that they have become much more available to everybody.