Natural gas is used currently because it's a clean fuel, it's good for demand load because the generators can be started and stopped eaily, and because it's been cheap for the last couple of decades.

But it seems to me that you're extrapolating the current ratio of energy sources, coal, gas, oil, nuke, etc., without considering the possibility of changing that ratio over the next 40 years. Coal reserves are not close to exhaustion, and the cost of production won't rise for quite a while. So as gas and oil become more expensive, there will be a tendency to build more coal, nuclear, and wind plants, thus offsetting the increasing price of gas and oil.

It seems to me that additional consideration needs to be taken of the change in the supply curves over time, in addition to consideration of only the supply as represented by today's curves.

Also I think you severely discount the long term elasticity of energy demand, at least in the U.S. where we currently are so wasteful, as was demonstrated in the 1970s.

And there is also the question of the energy/GDP ratio, which has been shrinking for a long time; is the rate of change of this ratio dependent on the cost of energy or is it only driven by other economic changes, like improvements in productivity or new technology?

Perhaps fodder for future diaries on this interesting topic...