You're neglecting scheduled load-following. If all the wind in Europe is offline at the same time, we will know about it at least a day in advance. Dedicated peakers only need to cover the forecast errors and unexpected plant shutdowns.

Suppose you have a sustainable grid in which the baseload is wind, solar, and run-of-river hydro (and wave and tidal); scheduled load-following power is wood pellet, biochar, and large industrial consumers being paid to shift energy-intensive production around to other parts of the day/week; and peak load is hydro, pumped hydro and biogas.

Suppose further that European total consumption is on the order of 1 TW at the daily and seasonal peak, and 0.65 TW at the trough. If the largest single point of failure is 10 GW, with an 0.3 % probability of failing during any given day, and the largest forecast error for consumption in any given day has a standard deviation on the order of 20 GW, or about 2 % of peak power.

Now assume that we want to have enough peaker capacity to make blackouts only happen once every 1800 days (approximately five years) on average, and keep enough peaker power online at all time that we only need to bleed off scheduled power about once a year. That means you need about six or seven standard deviations plus the largest single point of failure. Call it 150 GW of dedicated peaker power. For all of Europe, in a fully sustainable grid. So 50 GW of peaker power from Scandinavia alone is very definitely in the right ballpark.

The real challenge is going to be coming up with sustainable scheduled load-following.

- Jake

Friends come and go. Enemies accumulate.