Reading posts on this forum I am coming to the conclusion that digital component design is rife with poor engineering decisions.
About 10^20 bytes of data are moved thousands of miles over the internet, error free (after error correction).
The process of converting the digital representation of our music to analog should not depend on the last few feet of Cat5/6/7/8 cable or the RJ45 connector.
The engineering problem is to present 16/24 bit words (or bits for DSD) to the DAC chip with the correct timing, and to suppress and EMI, RFI or other noise to prevent its reaching the audio circuitry. Opto-isolators (e.g.) can operate a 10MHz or better (even 150MHz) and a Schmitt trigger can clean up a pretty ragged wave into a clean edged waveform. Buffers can store the data and OCXO parts with femto-second jitter levels are available off the shelf to strobe the data into the DAC's chip(s).
I am not suggesting that this is trivial to implement but I believe that Esoteric have come, if not to perfection, so close, that the front end components have little if any effect on sound quality when using async USB.
I am not denying that changes to front end stuff (switches, cables etc.) can affect SQ, I am saying that DACs should (and in my experience can) be engineered to be essentially insensitive to noise and jitter.
I think a similar argument applies to power cords.
Interconnects and speaker cables, carrying audio signals are a different story and do affect SQ obviously.