I think it is assumed, otherwise there will be cathode degeneration, which will greatly increase the signal voltage at the grids required for full output.Can I assume this would also necessitate two separate bypass caps?
Bear in mind that when you want to run KT66 in Cathode Bias each tube should have it's own Rk/Ck combination as per KT66 Data Sheet.
Marshall is doing that in their Astoria Series (anyone has schematic for the Classic for me?), two nice 10W Welwyn Enamel Resistors accompanied by 2 uncomfortably close capacitors..
Ive just gone through this process with an EL84 amp. In the end I had to gator-clip in a few different R values to get the ~97% dissipation as recommended by Rob Robinette.
I'd recommend you do the same, just start with some ballpark R value and take the measurements as per Robinette's site. Keep gatoring until you get it.
For the bypass cap I just went for whatever 'big' 50V cap I had on hand which was 100uF which I think is way higher than necessary. You could probably go way lower, but you'll need to do the math to make sure its sufficiently large so that none of the frequencies are being attenuated.
Note that the g2 voltage there is down at 300V.At 450v the Rk ( cathode resistor value) is listed at 500 ohm