40 amps of current is needed to produce 40 volts peak (100 watts RMS) across a 1 ohm dip in impedance. If this impedance is reactive the amplifier will be really unhappy.
I don't understand it. 40V across 1 ohm (or 40amps thru 1 ohm) would make 1600W.
Note: Watts RMS means something different and has no relevance. Power, obtained by multiplying RMS current and RMS voltage is AVERAGE power (equal half of peak power for sinewave). It represents DC power that would produce the same amount of heat. RMS Power does not represent anything useful.
AHB2 power supply is highly regulated SMPS capable of producing large currents. As for amp being less stable in bridged mode - AHB2 is designed to be stable since it does not have feedback in traditional sense. Of course in order to reduce distortion that low it has to have negative feedback, but this feedback in not recursive, meaning it is not zapped back to the front of the amp. Instead it drives another amp (error amplifier), with another set of power transistors, that at the very end corrects output signal.
Also, low output impedance, doubled in bridged mode, might be important for driving complex loads, but it is already very low, about 0.03ohm @1kHz. For the purpose of damping the speaker membrane it does not matter, since most of the speaker resistive impedance is in series anyway, reducing highest obtainable effective DF to 1.5