Agree with most of your points. Unless they're already started up (i.e., part of expensive spinning reserve) gas peaker plants can't respond fast enough to sol

Agree with most of your points. Unless they're already started up (i.e., part of expensive spinning reserve) gas peaker plants can't respond fast enough to solar PV ramp rates, which can be hundreds of MW/minute when clouds blow over a utility scale plant like Topaz (550MW). They take 10-15 minutes to come online from a cold start and sync with the grid. Gas peaker cold starts are expensive - the plants I've looked at (GE, mainly) allow some number of cold starts before the plant has to be taken down for service. But keeping them hot (spinning reserve) is expensive, too. The fix for rapid PV ramp rate is enough storage at the PV site to give ~ 15 minutes of continued nominal output when insolation drops out. This is enough time for classical grid management tools to be engaged. One driver for the equipment upgrades needed to deal with reactive power compensation, etc., is the price premium paid for "ancillary services", or power supplied at short notice for grid stability/management needs. The multiple over normal, average power pricing can be 100x depending on the circumstances. When residential inverters are smart and can respond to utility signals, then residents can participate in the ancillary services market. That will (or should, at least) pay much more than simple net metering schedules. The economics should accelerate deploying equipment that can do this, and that will mitigate the problems brought on by high grid penetration of PV generation.