First thing to understand is that asynch USB protocol does not implement a retransmit function of USB data packets - if a USB packet is corrupted at the receiving end, there's no retransmission - it will usually be heard as an audible glitch. This seldom occurs & the MicroRendu & it's internal Regen does nothing to change this operation or affect this protocol.
Second, Swenson's theories have nothing to do with less activity on the computer.
I'm not saying that I agree with all of Swenson's theories about this but underlying it is the premise that (fluctuating?) noise on the ground plane of a USB device (let's say USB DAC as it's more straightforward to understand) is detrimental to the SQ of the DACs analogue output. AFAIK, he didn't specify exactly the pathway & mechanism by which this noise affected the SQ. Most people assumed that it was via the ground noise on the audio clocks resulting in excess jitter.
Swenson's theory behind the regeneration of the USB signal (Regen) is that a USB receiver (in the USB audio device) that receives a better formed USB signal (better signal integrity or SI) has less processing work to do as a result of handling the better SI USB data signal & therefore introduces less noise onto the ground plane of the USB audio device affecting the audio clocks, as above. I actually believe that this is wrong, (or at least, not the full picture) based on my own experiments. If this was the case then galvanic isolation on the output signals (including ground) of the USB receiver would solve this issue & any such device would be immune to noise coming from the PC via USB cable. But this is not the case as we see with the Berkeley Alpha USB & with other USB audio devices that have isolation after the USB receiver. I have done experiments in which I isolated after the USB receiver & before the USB receiver - the before isolation was audibly better.
IMO there are two mechanisms involved, Jitter on the USB signal & noise riding on the USB signal or generated as a result of worsening USB signal integrity (SI). A USB signal with worse SI results in extra noise generation in the ground of the USB receiver and/or the USB clock & this causes jitter on the digital output signal generated by the USB receiver. This signal is often a I2S audio signal which directly feeds a DAC or SPDIF transmitter chip or in the case of an XMOS USB receiver it directly generates SPDIF and/or I2S signals. So isolating the digital signals after the USB receiver solves half the problem - the signals already have embedded jitter. So along with isolation, reclocking of these digital signals should solve the other half of the problem, right? If we have the two audio clocks on the "clean" side of the isolator, reclocking the I2S signals, it should "regenerate" the I2S signals with the clean clocks thus removing any jitter.
Each of these approaches improve the sound but isolating the USB signal coming into the USB receiver still sounds better when using this configuration. So something else is happening in the USB receiver when receiving a USB signal of weaker SI which is neither jitter related or noise related. What it is I don't yet have a handle on
The MicroRendu, using an isolated ethernet connection as input along with a Regen-style USB output & attention paid to the important areas of the power supply makes perfect sense to me