My background is neuro not physics. I’ve assumed that at this point the upper limit on bandwidth was related the the physical speaker. Hearing aids use balanced armature receivers and if you look at the spec-sheets as you increase the power (and size) of the receiver you lose high-frequency bandwidth. I can’t imagine that limit is artificial.
I believe that hardware advances are tremendously more expensive than software advances. I have no idea to what degree the hearing aid industry is chasing that upper bandwidth. There have certainly been improvements in feedback management lately, which may be the first step when chasing stable high frequency gain. On the other hand, a huge proportion of users may not have useable hearing above 10 kHz.
As for changing fitting rationales–it just changes the gain, so you can do the same manually with the levers. NL2 and DSL are the two most common independent fitting rationales and have slightly different ideologies about managing loudness versus audibility (in DSL high and low frequency gain tends to be higher and compression lower). The widex proprietary rationale will basically be NL2 with some tweaks, usually reduction in gain to support first user acceptance given X years of auditory deprivation impacting their tolerance for the frequencies they’ve been missing. I’ve heard of some musician audiologists who still like CamFit, but I don’t think that’s offered in the Widex (or any) software at this point nor is any research centre still working to improve it, so it’s rather a ‘historical’ rationale at this point.