In the H2CAR paper, figures such as 239 billion kg/year of hydrogen from 58,000 km2 of solar PV panels are tossed off rather casually. These numbers bear deeper analysis than they receive. For instance, 58,000 km2 of panels could be made by assembling an array of about 46 billion BP SX 170B PV panels (at roughly 1.26 m² each). At a future cost of $2/Wpeak, this array would cost about $15.7 trillion; today's cost would be closer to $40 trillion. Clearly we're not going to do this.
Another example of the disconnect between the researchers and reality is their proposed quantity and method of hydrogen production. Their most optimistic (smallest) quantity of hydrogen required is 239 billion kg/year, which they propose to produce from renewable electricity via electrolysis. The quantity of electricity required (at 100% efficiency, no less) is a staggering 9810 billion kWh/year2; this is nearly 2.5 times current annual US electric production. (Worse than that, it's roughly 6-10 times what it would take to power all ground transport directly with electricity3.) Even if produced from nuclear energy by a thermochemical process of 50% efficiency, this rate of hydrogen production would require nuclear plants equivalent to more than 8 times today's capacity
Here was my original article on H2CAR