The marketing at the time was pushing that the class G topology was enabling the a49 and p49 to run in Class A up to 50w. This has been discussed a few times and the general opinion was it was just marketing.
The problem with these sorts of statements is, it is very difficult to verify, through measurement, whether an amplifier is running in Class A mode whilst operating and, if so, for how much of its useable power output this remains true. It is marketing that is very difficult to verify or refute.
I accept that Class G has very clear design parameters that would suggest the amplifier very likely is running in Class A mode during its lowest few watts.
The real test is how does it
sound. I would not say there is any obvious advantage to Class G that makes it a substantially better amplifier topology than others in use. I have heard excellent Class A, AB, B, D designs.
As I say, the problem with switched rails is design complexity which makes long-term servicing a bit of a challenge.
I actively look for the simplest design possible now when choosing kit. This generally points away from microprocessors, lifestyle products, signal processing, and anything except the most basic circuit layout. A very good example is the Kef LS50 Wireless. It is an outstanding sounding product, easy to use, good looking, and very convenient. However, once the software is no longer supported, or once one of the various miniature, proprietary circuits dies, you are left with a very expensive door stop.