Once you have that constraint, it means you will either have the same network latency when writing to SQLite (if it is fronted by some lightweight proxy), or have a lot more frequent failover of SQLite (if it is running embedded within the app, thus following the app's deployment schedule).
I suppose if someone decides to deploy Postgres/MySQL replicas as a sidecar, then it will be the same as what you will end up with?