Ben nailed the issue in his reply, but I'm going to expand on it.

I'm not sure buffering the grid with batteries is cheaper than stationary SOFC fuel cells.

The point is that the power-handling capacity of a national fleet of (PH)EV's dwarfs the generation connected to the grid.  This allows a large number of plugged-in vehicles to compensate for wide swings between immediate electric supply and demand without adding a cent to capital costs or fuel bills.

Consider the US light-duty vehicle fleet, roughly 200 million strong.  If we conservatively postulate that each one has an engine of 100 horsepower (75 kW), the total power is 7.5 terawatts.  Total nameplate grid-connected generation in the USA is on the order of 1 terawatt.  If the US vehicle fleet could be plugged into the grid as generators, they could back up the entire world (for as long as their fuel held out).

(PH)EV's must have batteries with enough power capacity to perform acceleration and regenerative braking.  Depending on the design, they may be able to charge and discharge at rates from 10 kW to 100 kW.  This is considerably greater than the power which can be handled over a likely charging connection (220 V 30 A, or 6.6 kW) so the connection is the limiting factor.

If we assume a fleet of 200 million vehicles, 80% plugged in at any time, and 6.6 kW apiece, that is 1.056 terawatts of power-handling capacity attached to the grid.  If they are taking power at an average of 120 GW, that is 120 GW of spinning reserve you don't need because you can turn off their chargers briefly (reacting much faster than any turbine) and make up the lost energy later.  You can run most of the chargers on low until the incoming cold front hits the wind farm, and then crank them up to absorb the surge in power (allowing wind to supply much more than 20% of total grid demand).  Given a huge mass of demand which mostly needs a certain number of gigawatt-hours (plus or minus) between now and the next rush hour but isn't sensitive about exactly when or how fast, the grid manager would be able to do his job much more efficiently and more cheaply than is possible today because so much of the hardware for vehicles is available at no additional cost.

Didn't mean to write a book, sorry.

If the US vehicle fleet could be plugged into the grid as generators, they could back up the entire world (for as long as their fuel held out).

Wow! That is impressive.
What a pity it could back it up only for half an hour.

What breaks grids is lack of spare capacity in an emergency. That's one of the things Poet was implying in his post. Distributed, non-central, spare capacity is the best form of hardened infrastructure you can have.

Another problem with grids tied to renewable energy is underutilization at peak periods where the energy simply cannot be used. If PHEVs become a means of storing this over production, then your need to build other expensive storage is reduced.

I know he was referring to peak load. Usually, half an hour would be sufficient. Besides supplying all of the electricity for half an hour, it could also supply the missing 10% of the electricity for 5 hours.

Or alternatively, I could chose not to connect to the grid, and in case of outage run my car as a generator for all of my house. And you bet that then I will really save the electricity, when I know that I have only a few hours worth of energy in the batteries, and I still want to drive somewhere.

Just the way he put the global numbers looked so enhusiastically, as if we could supply the world from our car batteries :-)))

I thought it was really funny.