Fuel cells, something I will someday have to think about in great detail and we may actually cover in the final year of my Motorsport Technology degree. While they may sound like the saving grace of the fuel world, there's actually a great number of problems.
First is the infrastructure - it doesn't exist. Go to remote places in Russia or Africa and you will find people pumping petrol by hand. Now remove the petrol, because currently there's no hydrogen supplies anyway. All of the supplies would have to be established and not just on a piecemeal basis like one filling station per state or county - a small city would probably need a third of the number of petrol stations in "hydrogen stations" for people to take notice and actually be able to fill up their tanks.
Second is cost. How much does a petrol tank cost to manufacture? In the numbers they are made, we are talking less than a dollar to cut/mould two metal sheets and weld them together using a press and robot welder. Process probably takes 45 seconds. How much would it cost to make a fuel cell? No idea, but with it being a new technology it would be a heck of a lot higher simply because setup costs will be factored into the first few hundred thousand (at least) and we aren't talking a new set of dies or a new forge - we are talking entire new factories, with machines costing in the region of $5m. Those setup costs will put off even the biggest manufacturer without even having to do a full feasibility study. Remember that the more it costs to make something, the more you as the customer will have to pay for it, and customers are only willing to pay so much.
Next up is design - remember the weird and wonderful designs from the 1950's when everyone thought we'd be using nuclear powered cars by now, or jet engines? We'd have that kind of problem again until designers finally worked out how to get the best out of the new technology. Chances are we'd end up using a mid-engined, aerodynamically sound, lightweight box on wheels - nobody wants to be driving a
Smart car now, do they? The fuel cell will actually make the cars smaller and even more lightweight - a transmission replaced by a rotary potentiometer (read: variable resistor) at the top of your accelerator pedal gets rid of that lot and puts a very simple loop of wire in its place. I suppose it could reduce the cost a bit, but not much in the grand scheme of replacing probably 2/3rds of the manufacturing hardware...
Then we have safety. Has anyone actually done impact tests on these cells or the cars they're in? If so, how does it react to a 200mph impact (2 cars doing 100mph in opposing directions colliding head on), or to a side-swipe? How safe is it if someone torches your car at night, and is it liable to go BOOM and take out your garage, house and half the street?
Finally, for now at least, there is the psychological factor - are you going to buy one of these fuel-cell cars or are you going to stay away from this new-fangled technology? Personally I'm quite happy with internal combustion. Might be dirty, but I know I can rip an engine to pieces and find out what's wrong with it without taking it to a garage. Similarly, I can take a milling machine to the block or a Dremel to the head and make it go a heck of a lot faster. Can you do that with a fuel cell? Not if you value your life!!!