And with the chip being cooled via convection to the inside of the case, you have to probably double that temperature difference. And add the temperature difference from the chip die to its case. And compare that to to the maximally allowed junction temperature, probably around 150°C.
So the amp is not designed to continuously run at 100 W (assuming 90% efficiency; I haven't checked what power is claimed for this amp; ditto for the amplifier chip efficiency. These are just some round numbers).
But it could comfortably deliver a few W, with 100 W peaks, which is plenty for smaller rooms. (I typically listen at around 75 dB SPL and a tenth of a Watt, in a 110 sq.ft room. But other people might use a few W average power.)
Regarding the heat transfer coefficient, just google 'heat transfer coefficient to air', or similar, and some curves should pop up.
Anyway, the whole point was that air does transfer heat. If you are trying to use an amp like this to blast 100 W sine waves, you'll be disappointed. But for listening to music with normal power distribution (i.e. average to peak ratio), and peak power demand matched to the max power the amp can put out, it should work just fine, including thermally.