So a fairly small database still.

SQL data files max out at 16 TB, not a few GB. You split across drives (separate IO paths) if you're having IO contention or have inadequate IO throughput, not because the file has reached some arbitrary (smallish) size.

And if you are splitting for IO load purposes, adding a new file is the worst thing you could do, as due to proportional fill, all writes would go to the new file initially.