Replying to @null and @dansereduick, @EuphoricEuler, @MarietjeSchaake
One alternative to the "big bureaucracy" option is to have it so that if you process more than N% of a population's data subjects' data in a given year, you're in scope for the local DPA. Sort of similar to the CNIL's argument. That way you can kinda have it both ways.
♥ 2