I believe the focus is too strongly and exclusively aimed at the amplifier, indeed it's indicative of an underlying issue. Instead the question should be formulated: how well can a speaker be made to sound, and what are the bottlenecks to avoid in the amp-to-speaker interface to best and most effectively get there?
Improving on said interface, to begin with simply by removing the passive crossover to make way for active configuration, is a way to maximize not only the performance envelope of the amp itself, but also the drivers of a speaker. With further, specific "tailoring" from a more holistic approach the amp to speaker/driver package can be made to sound that much better, and at a much lower price, even with more amp channels necessitated for each driver segment instead of wasting power and performance envelope looking into a fiendishly heavy passive crossover load before the amp can get its handle on the driver(s).
I'm not necessarily saying a hugely expensive behemoth of an amplifier wouldn't make a difference or even improvement over a much cheaper ditto in a passive speaker context, but it's a "one size fits all"-approach that's invariably and inherently compromised, and one that will make the price of the amp spiral into stratospheric levels in the effort to try and ameliorate the negative influence of the heave load passive crossover itself.
Obviously there are great sounding passively configured amp to speaker combos out there, but from my chair you'll end up paying dearly for something that could've been done much cheaper actively. I guess it's symptomatic of an underlying tendency to revel over singular components - at least more than looking at the totality in which they're involved from a configurative perspective - and the sense of pride that comes from having paid large sums of $$$ on equipment.
