Seeing this article made me a bit curious about this question.
Just doing the basic math. I'm not factoring in any complex calculations. Just doing straight up multiplication and division.
According to the EIA, the United States consumed a total of 89.456 Billion Quadrillion BTU's over the first 11 months of 2014. Estimating for December, that means roughly 100 for the full year.
Converting to typical SI units, that is about 29.3 million Gigawatt-hours.
Using the standard of the Topaz Solar Farm, one of the largest solar farms currently in existence, which generates about 1000 Gigawatt-hours per year, it would require about 29,300 similar solar farms.
The Topaz Solar farm uses 9 million photovoltaic cells, meaning that many farms would need about 264 billion solar panels photovoltaic cells. Assuming one solar panel holds 40 photovoltaic cells means about 6.6 billion solar panels.
The Topaz Solar farm cost 2.5 Billion dollars, meaning it would cost 73 trillion dollars to build 29,300 of them.
Now, how does that compare to other forms of energy generation?
Comparing that to the Susquehanna Steam Electric Station, one of the largest nuclear power plants in the US, Susquehanna produces about 18,000 GWh of energy per year, or about 18 times as much as Topaz, but it cost 8 billion (2007) dollars to construct, or about 3.2 times as expensive.
The Grand Coulee Dam, the largest hydroelectric plant in the United States, produces about 20,000 GWh of energy per year, or about 20 times as much as Topaz, and cost about 5.5 billion (1998 dollars) for its current electrical generation, or about twice as expensive.
Alta Wind Energy Center, the largest wind farm in the US, generates about 1690 GWh per year, and has so far cost 1.85 billion.
One of the advantages of renewable energies that often go unheralded is their ability to be implemented on a smaller scale, relatively closer to their points of usage. One of the costs that don't often get factored in is the amount of energy lost due to transport. Because electricity often has to be transported from highly centralized power stations and distribution points, more energy is often lost in transport than used by residential and commercial consumers. Small-scale wind and solar farms that can be placed relatively close to residential and commercial properties would be much more efficient, since so much less energy is lost between the generators to the consumers.
At the end of the day, people may look at the sheer numbers involved in such projects and scoff, but it is often good to remind them of how these costs do not just go away when considering less-renewable, less-sustainable means. And it is often good to be able to do some back-of-the-envelope type calculations, just to show that as large as the numbers can get, they are still well within the means of possibility.
For example, imagine taking 1% out of our 600 billion annual defense expenditures and investing that in renewable energy generation. It's not that far out of our grasp as people may think.