I think the problem is that the author says that no “Grid-Level inverter synchronizers” currently exist to power a grid when there is no existing grid to follow.
An easily solvable technical problem.
Perhaps satellites could be programmed to send out synchronization signals — both 60 HZ and 50 HZ are extremely slow compared to modern electronics.
Perhaps even current GPS satellites already do something that would be usable, since they are currently used to provide accurate time to computers and cell phone networks. GPS time is expressed with a resolution of 1.5 seconds (https://en.wikipedia.org/wiki/GPS_signals#Time), so it might be usable if the jitter was small enough. Maybe StarLink satellites could provide stable 300 HZ (LCM of 50 & 60) which is globally synchronized.
Given a synchronization source, it should be easy to get the DC to AC “inverters” to put out AC of the right frequency and phase. But I still wonder about filtering the AC, as most (solid-state) inverters naturally put out waveforms that only approximate sine waves (i.e., lots of nasty harmonics) and filter capacitor banks for high power are big and expensive.
None of this solves the “reliable power” problem, of course.