On your typical Marshall "plexi"
.....the Presence cap is 600 Volts.
I did not think there was anywhere near that kind of voltage...
Do you really need a 600 Volt cap in that position.?
No, but you probably already have a 400v or 600v cap in that value on hand.
Worst case: 100w amp, 16Ω speaker load. Voltage = √(100w*16Ω) = √1600 = 40v RMS at the speaker terminals (and across the entire feedback loop). 40v RMS * 1.414 = ~57v peak.
I looked at 0.1uF caps from a major distributor, and the price difference between 100v and 400v caps wasn't much, and sometimes the 400v cap was cheaper (because it's a higher-volume part). If you're an amp builder and make 10,000 amps, you might need 80,000 of the 400-600v caps in that value but only 10,000 of the low-voltage presence cap. The higher voltage part will be cheaper because of your volume discount, even more so when you bump up the feedback loop to use a too-high-voltage cap.
So it's just convenience and lower cost that drives the choice. If you're building 1 amp, you can use whatever you want. Choose a voltage rating at least equal to the peak voltage of double your expected clean-output RMS voltage using the method I showed above. No sense in under-rating the cap (the a.c. voltage rating is always well below the d.c. voltage rating to account for peak voltages).