Making new accellerator cards for Amigas in 2015 that have LESS RAM than an old 1990s Amiga is lame from the consumer's perspective. Amigas already had 256MB of ram waaaaay back in the previous century.
This is a very simple to understand question of price.
Designing a small CPU accelerator with 1 or 2 memory chips on board
and a limited number of PCD layers in one taks that takes a certain time to do and to test.
Desiging a CPU accelerator which has many times more memroy chips, more PCD layer and or even DIMM sockets where you will need to test various DIMMS is a project which takes much longer.
The first project might need 3 month of pure working time.
A dedicated hoppy developer like Igor Majsta can do this using his Winter and Summer holidays and the weekends in between them. This means such a project could be done in parallel to having a normal job.
The second project will take longer, many month longer.
This means it will not only take longer but also take more money to do.
Now the AMIGA market is not really big.
Does it make sense for an developer to take 6 month unpaid holiday from work
to be able finish to accellerator in time?
Can this lost income be compensated by selling Amiga CPU card with more memory?
My guess is that the most sensible solution is to focus on developing a good working, well tested CPU card for a good price.
