What happens to FKs when you've to partition/shard the db? At a certain scale, they actually hinder the inserts.
For sharding, Vitess kind of supports them; Citus fully supports them.
You’re correct that they do impact performance to an extent, but as a counter argument, if your data is incorrect, it doesn’t matter how quickly you wrote it.
Really, FKs are typically implemented internally by RDBMSes as TRIGGERs that do what you expect FKs to do, which means they really are nothing more than syntactic sugar.