I'm going to put up some numbers to illustrate how ridiculously infeasible a hydrogen based transportation system is.
From
here daily US gasoline consumption is 383.3 million gallons a day. There's about 125400 BTUs of energy in each gallon so over a year that works out to 17.5 Quads (quadrillion BTUs) or so. Car engines are about 25% efficient so that's about 4.4 Quads worth of work.
Current fuel cells are about 60-70% efficient, call it 65%. For 4.4 Quads of work that's 6.7 Quads of Hydrogen. Liquifying H2 which is the only practical way to transport it takes about 40% of the energy content of the H2, so now we're up to 9.5 Quads. H2 evaporates at 4% a day so add that in and let's round it off to 10 Quads. Current electrolysis methods are about 30% efficient so that's about 33 Quads. Congratulations, you've nearly doubled the US energy consumption for transportation, and upped the overall consumption by close to 20%.
Keep in mind that the US uses about 100 Quads worth of energy every year, with about 7.2 Quads from nuclear. IIRC, there's about 100 nuke stations in the US, going all nuclear to make the hydrogen works out to something like 450 new nuclear generating stations. A couple billion for each station and that's almost a trillion bucks right there, then add in the hydrogen production plants, the compression and/or liquifaction plants, the fill-up stations, tanker trucks & other ways to transport the fuel, and you're looking at several trillion more. I'd ballpark a figure of $5 trillion or so if we're lucky, it's probably closer to $10 trillion which is the estimate given by auto & energy companies.
Going with full-electric charge-up cars isn't nearly as bad. Let's assume the electric motor along with the power grid is about 90% efficient which is actually a bit better than average. The 4.4 Quads calculated earlier becomes 4.9, which works out to 68 new nuke plants, call it 70. Assuming of course that battery problems can be solved, and then there's the problem of upgrading the entire national power grid from the generating station all the way to every home and business. You'll need at least a 240V 50A outlet to charge a car in a reasonable amount of time. Assuming the battery capacity problem is solved it'll be a heck of a lot cheaper than going to hydrogen. Battery issues are a lot less problematic than producing, transporting, handling, and storing a highly flammable cryogenic gas.
Moving on to renewable energy such as wind and solar. The US has an installed generating capacity of roughly a terawatt, or a million megawatts. Keep this number in mind.
The total solar energy falling on a square metre of ground is about 1kW, however solar panels are only about 20% efficient at best (unless you're running CIA or NASA satellites with unlimited budgets) so that works out to 200W/m2 in the sunny Nevada desert. 200MW/km2, about 5km2/GW, or 5000 square kilometres to power the US grid in a best case scenario. After accounting for clouds, bird poop, dust, and so forth, you'll need at least twice that area. That's an area of land larger than Yellowstone National Park. Building & maintaining all those panels is less realistic than nuclear fusion.
With wind, let's see, average wind turbines these days are about 3MW each. That's oh, 335,000 or so wind turbines, assuming the wind blows perfectly round the clock. In real life there's a 3:1 reduction factor (ie. a 100MW site will produce an average of 33MW or so), so that's about a million wind turbines. Expect to see a massive shortage of carbon fibre, aluminum, and copper if a mass wind turbine program is instituted.
The problem with renewables is very simple, power density; it takes ridiculously huge amounts of land & space to generate significant amounts of power. A nuclear plant can be built on a hundred acres of land and generate several gigawatts of electricity. Build a nuke plant and I can power a metropolis, with wind & solar I'll have to litter countless miles of countryside to do the same. Don't let the greenies tell you otherwise, they never do the numbers breakdown.