Its not a hard boundary because for arbitrarily-small distances the corresponding pressure differences would be arbitrarily-small to.
Trying to understand physics intuitively is sometimes not a good idea. An adiabatic atmosphere would taper off this way and it is more realistic than the isothermal case described by the exponential curve without boundary. Explaining why the real atmosphere has no upper limit is not as easy as you seem to believe. The pressure gradient depends on the temperature profile and that's a quite complex topic in the upper atmosphere.