You often see articles like “100% of new capacity built in location X this year was renewable“, but what does that mean?
Let’s explore the German wind power build-out as an example. If you look at this graph from 2014, you will notice the actual capacity that the turbines are rated for is the upper line. The lower, small series of spikes is the actual power produced by the wind turbines. It occurs at effectively random times and so does not follow demand, and it is roughly 20% of the theoretical maximum total of the turbines.
This means that if you build a 10 MW wind turbine, over the year you should get an average of 2 MW out of it – and, as I mentioned, randomly. Power grids can handle a little of this kind of power, but there are serious limits what you can put on the grid. Ignoring that though, it means you need about 1500, 10 MW turbines (capacity would be 15000 MW, actual output 3000 MW) to equal a single good size nuclear plant in output. However, having the power reliably produced from the nuclear plant means that it is predictable and useful.
Predictable and useful means you can reliably displace other sources with it. We burn a lot of coal and gas because we know when it is burned it will reliably produce energy, and cheaply (at least for the cost of the fuel anyway). Nuclear power can replace coal because it will generate energy reliably, continuously, and cheaply.
Note also the original example quote from the top said “built in year X”. This means if you had 10,000 MW of power infrastructure, and you built a total of 100 MW more, and it was all wind, you would increase the total capacity by 1%, and even less impressive, the total power by 0.2% (since the wind power only produces 20% of its rated capacity).