"Still... it takes a lot of energy to turn water into LH2/LOX - if we're using straight LH2 in an NTR it takes 120 MJ/kg just to make the stuff from water. If that was kinetic energy the kg would be doing 15.5 km/s."

A joule is amount tiny amount of energy. A million joules isn't much energy. When plug in a hair dryer into a electrical socket, how much do you pay for a million joules of energy?
How far do you go in your car using a million joules?

A kilowatt hour is 3.6 MJ, so if you using a hair for about 15 mins you are comsuming somewhere around 1 MJ and it's costing you a few pennies worth of energy.

120 MJ is 33.3 kwh or if paying 10 cents a kwh- $3.33
A tonne at 10 cent kwh: $3333.33
100 tonnes: $333,333.33
100 tonnes at $1 per kwh: $3.3 million
100 tonnes at $10 per kwh: $33.3 million

The cost of energy has very little to do with the cost of launching something into space. Whereas the cost of energy has a lot to do with the cost of air travel. Double the cost of jet fuel [kerosene] and you will increase air travel costs by around 50%. 5 times the cost of Kerosene, Hydrogen, and LOX [most common fuel for rockets] and it would have little affect upon the cost/price of launching rockets [even though rockets use a huge amount of fuel].
The reason fuel cost matter to airlines is because it is a mature and competitive market- in other words it is incredibly efficient- though it could get much more efficient.

The launch market is no where near a mature market and vaguely competitive.
Example:
Zenit-3SL total mass: 471 000 kg
About 90% of weight is fuel [kerosene and LOX]
Say kerosene cost $3 gallon or $1 per kg and LOX is $.05 per kg. So if rocket fuel cost was $1 per kg and there was about 425,000 kg of fuel, that $425,000 in fuel/energy cost, if this were to increase by 5 times that would be 2.1 million.
The cost/price of Zenit-3SL is around 50 million or maybe it's 80 million. But one thing is absolutely unconnected to the cost is the rocket's energy/fuel cost. If a Zenit-3SL was about $10 million per launch then fuel costs would be a significant factor. And according to:
http://www.astronautix.com/lvs/zenit3sl.htm
"Launch Price $: 90.000 million. in: 1999 price dollars."
So, "Payload: 3 750 kg (8 260 lb). to a: Geosynchronous orbit trajectory."
So, 90.000 million divided by 3750 kg gives $24,000 per kg- $10,909 per lb.
If it was $10 million, it would be $2666 per kg [$1212 per lb] to Geosynchronous orbit trajectory. This is equal to around $500 per lb or less to LEO.

Anyways energy costs are not a significant factor in rocket launches. Rockets launches are a immature and not very competitive market. At the the moment there are not even any markets in space- e.g. you can not buy rocket fuel or electrical power in Space- therefore you first will need a immature and not very competitive market, before you will get to anything resembling a mature and competitive market in space. A mature and competitive market in space could occur at a "blinding speed" by which I mean could occur within a period of a decade or two.