In the 21st century, Americans are ever-reliant on constant and reliable electricity. However, because scientists still haven't figured out an easy way to store electricity, all of our hourly, daily and yearly energy needs need to be estimated in advance.
Energy predictions are made using complex models, which give you an extremely accurate estimate of the total electric loads – or how much energy people are likely to use on any given day.
The problem is, these models have begun to shift with the rise of solar power.
California researchers started noticing something odd in their state's electricity load curves.
The outcome of the shift - a "duck" in the curve - is the biggest challenge that the solar energy industry currently faces in the US.
In a new video from Vox, writer David Roberts explains why this is the case.
The sun offers the most energy around midday, which means that in states like California, where solar panels have a strong presence, less and less energy is being drawn off the grid by the public at noon.
This causes the demand curve to "duck" at this particular moment.
At the same time that the sun is setting, solar production ends and - because it is nighttime - electricity use increases.
This means that power plants have to suddenly ramp up their production to meet energy demands at sunset, which is extremely difficult to do given the current infrastructure.
There's also an economic problem. Energy plants are generally only economically beneficial when they are constantly running. In other words, if you turn them off at midday, it costs a lot of money.
On the other hand, an overload of energy at midday could destroy or heavily damage the power grid.
Until researchers figure out how to store electricity easily and efficiently, solar power will continue to be wasted on an archaic energy grid.