This needs a step back. Nobody says all amps sound the same under all conditions. Certainly I could run my mains out of a couple of my guitar amps and ezch would sound different from teh other and from any "normal" stereo amp you'd care to name. I don't know if it's even possible to set the controls so that those amps are flat. I could run one channel off a half watt amp and get it clipping merrily and it would sound different from the other amps in the system.
That's an extreme example, I know.
But what people generally say is amps that measure the same sound the same. If I bring an 80s era Arcam to the fight and it has an audible rolloff in the high frequencies, that audible rolloff is also measurable. The amp is not "trying" to be accurate it's trying to make (at least what people then perceived as) ugly highs in CDs sound acceptable.
So generally what is happening is the tester has to make the "lesser," "cheaper" (more accurate) amp measure like the more expensive, less accurate amp. Not the other way around.
If you really want to spend thousands of dollars to get an amp that deviates from flat in a way you could perfectly replicate with an inexpensive resistor at the output of a "cheap" flat amp, knock yourself out. Just know what it is you're actually doing. But in music reproduction, amp-as-equalizer seems a less than ideal approach to me.