There will be a shortage if we try to replace coal, nuclear, and natural gas with wind and solar. I have on my desk a report from Morgan Stanley claiming that it would take 10 billion tons of steel and concrete annually to replace coal power.
By when? Next month? Obvious bullshit number is obvious bullshit. Nobody has suggested that replacing coal, nuclear, and natural gas with wind and solar is going to happen overnight. Not even quickly. Coal, nuclear, and natural gas represent large capital investments with long amortization schedules. The power companies will only shut one off short of its design lifespan in extremis, and there has been no spike in fuel cost for any of them. Quite the opposite. Gas is dirt cheap now, but most utility companies have set fees agreed with state PUCs when gas was expensive, which have not been revisited, so they're making money hand over fist on gas power generation.
Imagine that I have a dozen nuclear power plants all humming along at about 80% capacity. Now imagine I have one of those once in a century events that knocks out one of those power plants.
Why imagine, when we have actual numbers? Average capacity factor of nuclear power plants in the US for 2015 was 91.9%, the highest it has ever been. If you follow the link, you'll see that at least the top 10 plants are actually operating at capacity factors in excess of 100% in order to achieve that average. Now consider that, with the shutdown of Vermont Yankee, there are only 99 total nuclear plants in the US. Having not just 10 plants, but 10% of the plants running at over 100% capacity, where they are by definition eating into their safety margin, doesn't seem all that safe, and it means that quite a few of those 99 plants are running at much less than 91.9% capacity factor.
Of those 99 plants, the majority of them are of such an age and design that they're incapable of being throttled, so when they're operating, they're operating at 100% or above. That means out of the 365 days in a year, the average nuclear power plant was offline for 30 of those days, and for every year but 2015, it has been worse than that. So there is no margin to "crank up" to accommodate a plant going offline.
In short, nuclear power plants are just as dependent on the existence of the full grid as wind and photovoltaics are.
This schedule should mean that with a dozen plants and an expected lifespan of 50 years I can expect a new plant to come online about every four years.
Design lifespans were universally 30 years and between 1977 and 2013, there were no new plants started. The Obama administration approved construction of 4 new plants. The US will be transitioning from nuclear to solar and wind by default, simply because those plants are not being replaced fast enough. But it won't happen so fast that Morgan Stanley's nonsense number is even remotely relevant.