I doubt that there is any difference between the 20 A 6-FET and the 25 Amp other than the shunt.
What is the weakest link in the controller as far as taking more current is concerned? I thought it was the FETs, but a glance at the spec sheet makes me think they can take more current than 20A distributed amongst 6 of them... Can anyone tell me if I'm flat out wrong?
I think you are asking the wrong question. You really can't divorce the controller form the system and I would again ask, what are you trying to accomplish?
The rough analogy is Current is "torque" and Volts is speed(I know, very simplistic).
Assuming there is enough current for the motor to reach it's no-load speed, adding more Amps has diminishing returns.
Generally, more Amps are needed as Volts go up to allow the motor to reach the higher no-load speeds.
In answer to your specific question, I would guess the Infineons would be limited to the physical size of the shunt(s) and possibly the heat sheading ability of the controller, the phase wires and their connections.
Running too high of a speed motor forces the controller to run at Max Amps and this cond. is exacerbated anytime the motor speed falls to below 1/2 of the no-load speed(Climbing hills).
I have done all these things together and melted the phase wire connectors. A good thing as these acted like fuses, before I started to melt windings.
Bottom line, the gains you are looking for will come with Volts and the Current only in the sense of what is needed maintain the higher Volts.
This is what I have found to be true with sm. geared motors and i suspect a latge DD motor may react differently. If we keep this thread alive, I'm sure someone will chime in with more real World experience with large DD motors/controllers.