It occurs to the casual observer that if your power conditioner is doing it’s job, and the power cable is equal in size to the circuit cable, no change is necessary. Likewise, there is no advantage in swapping component cables either, as their purported capabilities have been rendered moot by the power conditioner.
Twisting the Romex in an attempt to improve common mode rejection is not a bad idea, but implementing it in most construction (e.g. through studs or along top plates) would be an electricians nightmare. My homes have neither crawlspace nor attics.
Power starving the amp is an amusing concept, unfortunately not supported by Ohm’s Law or actual amplifier requirements. A hypothetical: a 200 W/ch amp operating in Class AB at 60% efficiency draws a peak of ~700W (200+200 /.6 =666.7). Instantaneous peaks of 1 KW are possible if the amp has that much headroom. And those peaks are typically handled by the amps power supply capacitors - that’s their job. Either way, well within the capabilities of a single 15A (1800W) or 20 A (2400W) circuit
Assuming the speakers can handle 200 W (23dBW) and 300W peaks, and have an efficiency of 90dB, that gives a peak output of 114.5 dB (90 + 23 +1.5) Coincidentally, that’s about what most 2X8" or (1X8" and 1X10" such as a Wilson Alexia V) floorstanding speakers are rated for and capable of. Most stand mounts, KEF LS-50s, et al, that number is around 10dB lower. So even when pushed to their limits, the circuit demands are well within the capabilities of a single 20A circuit. The idea that a DAC or preamp or streamer have great dynamic power requirements simply shows a lack of understanding of their basic operations.
Now, using that same basic model I could come up with a scenario using Kilowatt monoblocks driving some megaspeaker that requires 2 dedicated 20A circuits, and if that applies to you, great. Ohm’s Law also still applies.
Finally, the recommendation of multiple earth grounds is not only wrong, it is dangerous and a violation of the NEC. Even slight differences in resistance between different grounds would create a voltage differential, line noise and potentially dangerous ground loops in and around the structure. Just don’t.