For signal integrity issues increasing the link power can often overcome some of the issues caused by longer traces and connectors in the line - while less of an issue for desktop devices, then that goes against the ideal of a low-powered device with limited cooling. Doubly so as it's hard to re-clock in the timescales needed for intermittent use power saving, so will be using that extra power when idle.
I suspect the earlier comment about "Half the performance with CAMM" is likely at iso-power, but that might still be a pretty big dealbreaker.
More power is to overcome switching losses and parasitic reactances. You can increase drive strength up to a limit to overcome this, but a slight clock reduction will make things work at the same power.
CPU's and GPU's reclock extremely fast to my knowledge, but what we're talking about isn't dynamic reclocking, just limiting the max clock as suitable to the system design.
We already see this when we have laptop silicon that run faster clocks when using soldered RAM compared to when the same silicon is using socketed counterparts.
That this wasn't an option probably mean that they're either far too close to the limit, or unwilling to allow a design that runs below max speed.
I suspect the earlier comment about "Half the performance with CAMM" is likely at iso-power, but that might still be a pretty big dealbreaker.