Inverters are far better than anything rotating-mass-based for control over waveform. At least if you want to, that is. You just take a (possibly GPS-disciplined) quartz oscillator and have no issues hitting 50ppm frequency stability at lower phase noise than can get through the system at 50/60Hz.

They can essentially do this instantly, i.e., substantially sub-wavelength of your line frequency (a PWM of 10 kHz isn't even high for an inverter, and 100 kHz is well within what's feasible for large systems operating at a couple hundred volts; 10 kHz would equate a light-speed delay of 30km, 100 kHz of just 3km, which gets into the territory of relativistic w.r.t. medium voltage rings in cities).