I thought SetJmp must be talking about the "wall wart" power transformer which are typically very wasteful. I've not been able to find authoritative data on what the dissipation of the device is but I've gotten some anecdotal reports of 2 watts. The output of the power brick for my Motorola SB5100 cable modem is 12V at 750mA or 9W. Neither 2 watts nor 9 watts is going to produce that much heat...
But then I found a report commissioned by Motorola to determine whether the 2210 was likely to burn the skin of people touching it. The case hot-spots range in temperature from 40C to 50C+! I can't even imagine how they get such a trivial device to be so damn hot. The conclusion of the report was that it wasn't quite hot enough to actually burn people (according to ASTM protocol of "can you touch it for a minute").
If 2 watts is really the dissipation, that's pretty low power and the output of the wall wart power supply sets an upper bound on how much power the device can draw. But you can tell just how wasteful the wall warts are by looking at the difference between the input and output. That cable modem power brick that has a 9W output draws 0.3A at 120V from the wall which is a 60W pull to yield 9W to the device for a whopping 85% inefficiency. For small devices, the power brick will often be the waste generators. Although not definitive, here is someone's empirical tests showing that even the wall warts aren't that important relatively speaking. Then again, George Bush II called them "energy vampires" so take your pick.