That a good point. But a huge part of the GDP increase is an increase in efficiency, i.e., you can do more with less energy. A 2019 computer would cost a totall

That a good point. But a huge part of the GDP increase is an increase in efficiency, i.e., you can do more with less energy. A 2019 computer would cost a totally different price in 2009, and would be infeasible expensive in 1999. But if you compare technologies within one year, the price tag gives you a very good idea how much energy was used for producing each one. Of course, there are scale effects at play, so an R&D cost (energy) might only occur once, but is then distributed within all units produced. This will very well be reflected in the price, once something enter mass market. The same will happen with photovoltaic cells: They will continue to get cheaper (and more efficient) and at one point it will make economic sense to put them everywhere. And at this point it is clear that they are net energy producers, where currently it is still debatable, at least in more northern regions.