Can the specs be changed? I'm having a hard time achieving the requested GBW for the opamp while respecting power consumption and phase margin constraints, and it is still typical corner without parasitic capacitances.

Yes! I said so in the webinar. You can't just bend the specs to match what you have in simulation (otherwise everyone would be doing it and the specs would become meaningless). But I admitted that the specs are largely patched together from various sources, and I'm sure there's more than one place where we inadvertently put in a spec that will be impossible to meet. In that case you just need to alert us---like you just did---and we'll review it, and if you have a reasonable point then we will adjust it. Just to be sure: You're talking about the 12MHz GBW, 70 degree phase margin, and 300uA no-load power?

Yep. I'm achieving 6 MHz with 30 pF load only, 70º, 300 uA current consumption and 3 V power supply. I still have to run MC and transient sims to get offset and distortion. Which should be the voltage offset?

Okay, the conclusion is that the spec implies two different op amps. This could be done by having different selectable stages, but Andy thinks that we are less likely to violate some patent or other if we just split the op amps into two, one high bandwidth, high power and one low bandwidth, low power. This works well considering that the op amp is currently over-subscribed; if we split that into two IP blocks then we can accept more designers.

Thanks! The high gain opamp is the high gain-bandwidth, not high DC gain. My design pass the low power one as it is. I will try to triple its power and see what happens.

Right, high gain-bandwidth. My fault for the poor wording.