[ougawouga]

Yes. The majority of the energy is spent maintaining the network's security. Once that security has been obtained, additional transaction volume -- assuming it's within the network's constraints -- will add little to no additional cost.

To expand, Bitcoin's security is maintained by the miners, who also expend the majority of the network's energy. Transaction volume for the most part is borne by all the network's full nodes, which use relatively little energy.

[DennisP]

Miners don't choose whether to mine based on whether the network is secure. They mine if it will be profitable to mine.

So we can expect the total cost of mining to be approximately equal to the total block rewards plus transaction fees. The more valuable Bitcoin becomes, the more valuable those rewards will be, and the more money will be spent on mining.

Some percentage of the mining cost will go to energy, and the rest to hardware. So it's possible to estimate the energy cost of mining at various price levels and times (since block rewards decrease over time).

Ignoring transaction fees, it's actually not terrible. I figured out a while back that it would be something like several gigawatts, if Bitcoin became a trillion-dollar currency in twenty years. But if transaction fees become significant it could get a lot worse. (Luckily there's some work in progress that could increase scalability enough to keep the fees low.)