I've already mentioned my mega-project of interest. Large scale solar power. It would cost under $20 trillion to replace 100% of electrical production. And probably less than twice that to replace rest of the energy industry. And the year over year savings would break even after a few decades.
You wouldn't need to replace 100% with solar. Over 50% of energy produced in the US is lost to inefficiency in the first place, most of which could be eliminated through standards and infrastructure legislation.
Government can't change physics, though. The 2nd Law of Thermodynamics applies. Coal plants can generate an
average of 31% of theoretical potential energy, with a practical high of 45% that depends more on input coal quality rather than actual infrastructure itself. Legislation can't change the quality of coal that is available. Next up, the transmission. Energy transmission always involves some losses, because there's always electrical impedance. High-voltage transmission lines lose about 1% per 100km of cable, which really is not a bad loss rate and little can be done to improve it. Finally, transformers step down the transmission voltages to 120V for local household use, but those have an efficiency of 98% approximately. The transmission infrastructure is about as lossless as it can be right now, unless there's a revolutionary new approach.
In Sci-fi fantasyland, we would have lossless superconductors, of course, but that isn't going to happen. Government can't legislate against science obeying the laws of nature, and it can't much rush the remaining marginal advances in efficiency by throwing grants recklessly about.
As for massive-scale solar, that seems pretty unfeasible purely on logistics alone. America uses 3,700,000 GWh of electricity each year. A 200-MW solar farm like
Golmud in western China covers 2.2 square miles and produces 317 GWh each year. Extended to 100% of American consumption, that's 25,630 square miles of solar panels, or 23% of the land area in Nevada.
I'm not going to argue the costs and saving, but claiming that a massive solar megaproject pays for itself in a few decades is troublesome when considering the
lifetime of a solar cell is a maximum of 30-35 years, and diminishes to 80% capacity within 20 years. You have to maintain and replace the 25,630 square miles of solar panels, you know.
Meanwhile, suppose everyone stopped using gas-powered cars and used Chevy Volts with 16 kWh batteries capable of 35 miles range on full electric. There were 3 trillion vehicle-miles of
traffic last year. That's an additional 1,400,000 GWh. That's 35,400 square miles of solar cell that would cover 32% of Nevada.