That just bothers me the more I think about it. Let's take an EL84, since that seems to be the favorite "Class A" tube. We need some assumptions to make this walkthrough easier. Let's say the bias voltage is -11V for 90% idle dissipation, and -15V will drive the tube into cutoff. This means that a grid signal of +/-4V will cause the tube to reach cutoff and make your push-pull amp cross into Class AB. Now, making the bias voltage more negative will cool the tube bias, so let's say Mesa achieves 50% idle by setting the bias supply to -13V. Cutoff will still be reached at -15V, so now it only takes a signal of +/-2V to reach that point. If anything, that makes this amp even closer to Class AB, not Class A as it's even more likely that it can reach this signal level.