If I might be indulged in a very long post on fuel cell/battery fuel cell hybrid vehicles based on the recent DOE reports projections of $50kw for fuel cells when mass production is reached:
Running the numbers the best way to take cost and weight out of a BEV is by the addition of a fuel stack.
If the target for a BEV is perhaps twice the range of a Leaf, then the likely improvements in battery technology pretty well cancel out, so you still end up with quite a heavy and expensive car about the same as the Leaf but with better range - I have allowed a two-fold improvement in weight and cost.
So you could swap a ~50kwh battery for a ~12kwh one in a hybrid, and replace that with a fuel stack of around 25kw which even with the hydrogen storage tank and exhaust system and so on would weigh a fraction as much as the ~38kwh battery weight you have saved.
You will end up with something much lighter and nimbler than the Leaf, and with better range than even the 200 miles I have allowed for 'Leaf II'.
That would really take advantage of the light weight of moving to carbon fibre and so on, instead of loosing the weight advantages in the battery pack.
The cost the references are giving, $50kw for the fuel cell system, would if realised still make this cheaper than adding the extra 38kwh of battery, plus losses incurred by the extra weight, in the battery alternative.

Hydrogen infrastructure costs would also be much reduced, as most folk would only need to visit a petrol station occasionally, and the alleged efficiency losses from the use of hydrogen instead of batteries would be moot as short journeys would be done on batteries.
Local pollution would remain zero whether batteries or the fuel cell were in use, whilst the life expectancy of the fuel cell would greatly increase.

The car would also be a lot more fun to drive than a BEV.
From the DOE report
http://hydrogendoedev.nrel.gov/annual_progress10.html
the introduction page 7:
'The cost of a hydrogen-fueled 80-kWe fuel cell power system projected to high volume production (500,000 units/year) has been estimated to be $51/kW (assuming 2010 technology), as shown in Figure 2.2 Cost reduction was a result of simplified architecture and reduction in stack component costs through ongoing R&D efforts. The cost of the fuel cell stack has been estimated to be $25/kW (assuming 2010 technology).'

OK, I am caught with my hand in the cookie jar, I rounded it down by $1!
OTOH, by the time we hit mass production we won't be using 2010 technology.
I'd guess that we will hit mass production around 2017.

What a lot of folk here have not done is distinguish between inherent cost and the cost of low volume production.
There were several issues in cost, the biggest of which was the use of precious metals.
Note that I say 'was', as the precious metals cost has now been reduced enough to allow the costings given.
We now have in our sights though further reductions, so that the amount of precious metals used can be no more than that in a catalytic converter.

Fuel cell stacks scale, so that the 80kw stack they are quoting for will be proportional to a 25kw stack, near enough.

The savings from a pure fuel cell car in building a hybrid would be around 25/80 of the stack cost, so you are talking around 1/3rd of the cost, although the price for the hydrogen storage will remain constant, as that is there for when you want to go long distance.

The savings against a BEV amount to around 12/50kwh, or around 25% of the battery cost, plus whatever you save overall by weight reduction, plus the hydrogen tank cost which looks from this to be around $2500 and to weigh around 70kg or so:
http://www.hydrogen.energy.gov/pdfs/review10/mn008_leavitt_2010_o_web.pdf

Now it is not immediately clear if the system costs for the 80kw fuel cell stack includes the hydrogen tank, so I will simply assume that it doesn't, which is unfavourable to my case.

Costing for the bits that are different and assuming $200kwh for batteries, you come to 12kwh of batteries plus 25kw of fuel cell stack plus the hydrogen tank for the plug in fuel cell hybrid, which is $2400 + $1250 + $2500 = $6150

For the 80kw fuel cell car: $4000 + $2500 = $6500

For the 50kwh BEV: $10,000

The advantage of the plug in over the fuel cell car is that it would not require the same amount of infrastructure and most of the time you could simply plug in at home, which would also extend the life of the fuel cell stack and allow the use of home solar etc for those who like such things for most of most people's mileage.

The advantage over the BEV is that the weight would be much reduced, and I have not counted the cost savings from synergies due to this, and range limitations would be eliminated.
The whole car is going to be lighter and nicer to drive.

Weight saving is 38kwh battery at perhaps 200wh/kg ( I have used around twice present system weight ) + 190g, less 70kg for the hydrogen tank = 120kg, less the fuel stack weight, which runs at about 1kg/kw, so 25kg

So you are talking about a ~100kg weight saving, which on something like the BMW Megacity is a lot of weight.

Of course if you plug in different figures you get different results, but the conclusions I have come to appear to be robust, as you save money even if you assume a very challenging $150kwh, and weight even at 250wh/kg

It would take a lithium air battery to change them, and we haven't got a clue at present on how to do that for production vehicles.

Two further thoughts:
The addition of a fuel cell stack solves the dilemma of providing for those that can't charge at home, without creating a very expensive infrastructure to do so, and without needing to build two different types of car.
The price against their using a pure fuel cell car is pretty much a wash: (12kwh @ $200kwh = $2400, vs 80kw fuel cell - 25kw = 55kw @ $50kw = $2750)
And folk if they went to the shops or wherever could still charge up using any electric points available, but only when convenient rather than a matter of necessity.

Secondly as we move to larger battery sizes and greater ranges the problem of the speed of charge gets worse.
A 50kwh battery would be impossible to recharge on a 110 volt circuit in any reasonable time, and difficult on a 220 volt if fully depleted - OK, fans such as Tesla drivers can work around it, but it takes several hours.
Of course if we are talking about larger heavier vehicles with even bigger batteries the problem gets worse.
Perhaps we could fast charge, but on a 50kwh battery the c.30 minutes for the fast charge to 80% on the Leaf goes up to a much less convenient hour, and if you go to 75kwh then the problem gets even worse.
If you keep fast charging you will also decrease the life of your battery.

There are no such problems with the c.12kwh that would be needed for a plug in fuel cell hybrid, which could comfortably recharge overnight even on 110Volt, and for which a stop at the shops and an hours charge on 220 Volt would provide useful additional range.

I would seek to argue that a fuel cell plug in hybrid is an inherently superior technology to BEV cars for anything other than purely urban use, providing fuel cells can be had at reasonable cost.
From the DOE figures it appears that they can.

Introducing pure fuel cells in larger vehicles providing 80kw of power will obviously drive down prices much faster than going to hybrids initially, and the economics are better for large vehicles.

What they mean by "zero emission".