Do you think its more likely that all economists are shills who don't understand that datacenters consume power, or that economists model things on a shorter time scale than physicists?
By this time, the current relevance of those inputs they ignore is big enough that their predictions' time horizon is somewhere on the last decades of the last century.