When using the speaker, for some purposes you want absolutely max volume - that involves getting the physical parts of the speaker moving fast and far. Yet move them slightly too fast or too far or do that too often and it'll physically break. Odd effects like resonance mean two very similar looking signals can have widely different physical effects.
The ideal threshold might depend on numerous factors, like temperature and pressure, the age of the speaker, etc. Easy to do in software, impossible in hardware.
So why not just have more margin and make the line between 'loud enough' and 'breaking apart' wider? Well to do that your speaker would need to be physically larger, and space is very much in a premium in many devices.
(Of course, you could feed always feed such an absurd amount of power that it would break even the protective circuits, but that ultimate vulnerability is just inherent in the nature of electronics).
On the other hand, very high-power, short-lived 'transients' are necessary to give proper clarity and realism to the music. Most amplifiers, even very-high power ones usully idle along at about one or two watts or less, but the extra couple of hundred watts available allow lots of 'head-room' for those high-power transients so that they don't get clipped and distorted.
A high-power amp doesn't necessarily sound very much 'louder', but it sounds very much 'cleaner'.
It's speakers inside a phone or something that are very size-performance limited and are vulnerable to this. Those same devices tend to have advanced processing of audio headed to the speakers to make it sound better and not blow up the speaker.
An old Motorola I had had over 15,000 knobs to tweak in the audio section of the service app...