Lots of interesting opinions on this.

I'm sure lots of people covet a Graebner-like instrument - but one has
to think how many  of Bach's (or anyone's) students, or the people who
were going to buy his Clavierubung, would have had anything near.
Still there ought to have been quite a few more modest German doubles
around if these publications were commercially viable. All chopped up
around 1810 I fear.

Did people change their harpsichord as often as people nowadays change
their cars? - which might be one implication of the timeline which
Davitt Moroney set out for the Graebner and the 2-manual works - i.e.
the appropriate instrument emerges just the same time as the music is
composed. Seems most likely that people would use them for a few
decades, or until out of action (pun intended).

The oboe da caccia is a one counterexample of Bach using new
technology, but specifically in performances directed by himself...

If people don't know what the Mayer looks like there is a picture here:

http://www.geocities.com/nikodamm/mayer.html.htm

unusual keyboard with an accidental in the far corner. It might well
go down to GG but takes no more space than chromatic C-c.

The story of the alleged 'Johann Nikolaus Bach' instrument :

http://www.geocities.com/nikodamm/bachhist.html.htm

reconverted from a piano-harpsichord, you get 2x8 and three rows of
jacks with different tone quality (nasal / 'normal' / luteish), which,
should you want it, could allow the pursuit of quite some different
instrumental styles - though playing the WTC on a spinet (single 8')
for 10 years I never got bored.

And on clavichord it's hardly boring either - think of the E minor
Prelude with dynamic shadings of the top line vs. chords vs. obbligato
LH, or the G minor with swells on the trills. In fact considering the
clavichord's abilities to both mimic a lute and play counterpoint...

With regard to the lute-harpsichord, the one technical hitch is no
damping, which means no control over when a note stops. Maybe Bach
cares about this sometimes, at least that would be one objective of
careful contrapuntal notation.

~~~T~~~