I vaguely recall the Scott Manley video (longtime follower of his channel) - wasn’t he assuming that the chips can be run hot, and because the T^4 factor it results in much better radiative cooling?

I don’t know if the “chips can be run hot” assumption is a fair one. It results in significant degradation of the chips life.

At minimum it needs a redesign of the current set of leading chips which are designed for a specific temperature range. Redesigning at a different operational temperature is not a walk in the park.