Solar power is highly variable, and highly intermittent. It's intermittent and variable even if you pair it with wind power: it's true that the sun shine tends to shine most brightly when the wind is fairly calm—and vice versa—but the negative correlation is far from perfect.
This means that, if we want to rely more and more heavily on solar (and wind) to provide our electricity we need either to find some way of storing the energy when the sun is shining, get used to the lights going off occasionally, or boot up and run down conventional plants when renewables stop providing.
We don't want blackouts. We don't want to turn conventional plants off and on. So the obvious solution is storage. But people don't quite realise the scale of storage we'd need. Two recent articles illustrate this nicely. Euan Mearns, at Energy Matters, gives an idea of just how much storage we'd need if solar provided even a Hinkley Point C size chunk of UK generation.
And Matt Ridley, in the Times (and reproduced on his blog) gives us an idea of how much this might cost:
Yes, but we would not use car batteries; we would use bigger units, and more efficient and newer lithium-ion batteries. All right, let’s buy Tesla Powerwalls instead. We would need 160 million of them to cover a day’s consumption, or 3.3 billion to cover a week when we’ve electrified heat and transport too. They retail for $3,000, so that’s about £8 trillion. For a system that would only rarely be needed in full. Maybe we could get a discount.
Pay heed: solar will eventually be our key power source, especially given all the huge technological improvements we're seeing ever year. It's getting cheaper and cheaper. But until we can do battery storage cheaply, it's practically useless.