Wattage 4 or 8 ohms?

Not being entirely sure by what is meant by 'peaks increasing expotentially' and for the benefit of others facing a decision about how much power is needed, consider this as a rough guide.

Assuming you anticipate that the average sound pressure level of your music will have peaks 15db higher than the average SPL, not an unreasonable assumption, I think (nor is it absolute), if you are listening to pop, jazz, & classical music.
Assume your speakers have an efficiency of 86db's.

Assume you have a meduim sized room where I think an 85db SPL would be considered by most to be fairly loud but still within the ability of the room to respond without overloading. The following watts would be needed to producte the indicated SPL. This little chart contains no consideration of current needs, impedence values etc.

1 wt - 86db
2 wt - 89 db
4 4t - 92 db
8 wt - 95 db
16 wt - 98 db
32 wt - 101 db
64 wt - 104 db
128 wt - 107 db
256 wt - 110 db
512 wt - 113 db
1024 wt - 116 db

IF, and only IF, you found my 85db average SPL reasonable, and you add 15 db to that, you would only need 32wts. Now if you want 3 or even 6 dbs of headroom you would need no more that 60 or 120 wts (BTW all numbers are rough/rounded and meant for illustration).

FWIW.

I've got no argument with you'r table. One can easily see that there is quite an increase in power from 86dB to 104dB. IMHO, and this might be more so with ss than tubes, that one would probably be better off staying away from an amplifiers ceiling of power delivery to keep the sound as clean as possible. One way to accomplish that is to have extra power at hand.

The reason to have a more powerful amplifier is not to play the music louder, but to play the music more accurately. The soundstage "opens up" because the peaks can get louder while the valleys stay quieter. The music also sounds more real because, again, the peaks can get louder. The reserve headroom in the amplifier makes all the difference. The sharp, high peak of a whip crack or even just a smack can require near infinite power to duplicate.

My viewpoint is that playing the music either louder or quieter than the original session is a type of distortion- at least that's what I tell my wife when she wants me to turn it down.

I'm sorry for getting a bit off topic here, but in some instances a higher powered amp might sound better even at volumes that could be handled by a lower powered amplifier. In some class A/AB amplifiers designs the bias of class A before going into class AB increases proportionaly to the total power output..