Universally ideal arranger is both very hard and easy to produce.
Why hard?
Because every person have different requirements and different approach to the keyboard, depending what music you play, how you play it, how advanced of a player you are, how technically advanced you are, and many many other aspects.
For some, simple MP3 player is more than enough (even keys may be optional), there are people that only play with styles, people that use MIDI's, or the combination of those three.
Why easy?
Now we're entering 'the twilight zone' where every keyboard manufacturer is afraid of, and will never dare to make this suicidal move and create such thing, because that's gonna be the last keyboard that they'll ever sell. There where some some attempts in this direction from hand full of companies back from the late 90's like Open Labs with their NeKo and MiKo that where running Windows XP, there where few others that I can't recall right now, and currently available Studio BLADE, Kami XL and ES models up by Music Computing, but they where all focused in studio application, and still portable enough to be used on stage if you wish to.
To recap:
It's not absolutely impossible to create such Arranger keyboard from engineering point of view, for both hard and software, although such creature would rely mostly on software than hardware. I would imagine more than 2 or even 3 fairly big touchscreens and some hardware knobs, sliders and buttons, all highly configurable. Writing proprietary software for it would be the most challenging part I think, since it have to be extremely flexible and universally compatible. It would need to be able to load any kind of VST or any other kind of sounds of your choice. There a currently available processors that are more than capable for such task, and are capable to cope with the demand of driving such software. Hi-class computer processor today have as much as 16 cores, 4.0ghz per core of clock speed, and total of 32 threads. So, yes...from the hardware point of view, we have the power to fuel such thing TODAY. And if you take into account that those numbers will be doubled in no more than a year with the current rate that technology evolves, it's crazy how much power we can have in a keyboard.
Still, the chance of such beast to be created is fairly low since the BIG names in the industry that produces such class keyboards have no interest of doing so, and even if they do, the niche market is not that big and they would not make much money from it anyway, even if the price tag is very close to $10,000 or slightly more.
They have it all sorted, fellas...that's why they'd like to grind us slowly into the next model they'll produce with mere cosmetic changes and handful of new features from the previous model, and have a solid piece of money every other year...it's that simple.
So, anyone willing to write some arranger software?