Windows/Driver bugged on purpose so you must buy expensive routers.

I do not know the exact details of 1000BaseT or multichannel PCIe, but IMHO, it would be a braindead idea to assume that the prorogation delay would be the same on all channels and hence a single sample pulse could be used for all channels.

In the 1960s 1/2" 9 track computer magnetic tapes tried to use this concept up to 800 BPI recording density, but in order to read tapes written on an other computer, you had to adjust the read head azimuth angle for different tapes.

Later on in the 1970's the 1600 BPI (and more denser) tape drives used self clocking codes an all 9 tracks individually, so the read head azimuth was not an issue and the skew could be quite large.

These days, only system employing a huge number of parallel channels may use parallel sampling. One example is the DVB-T 8k on the air digital TV standard, in which there are about 6800 carriers, with a symbol rate about 1 kHz and hence symbol time of 1000 us, in which parallel sampling works reasonable well. However if you want to watch TV in a car or train above 100 km/h, the Doppler is quite harmful.