Bitcoin does use a lot of electricity, exactly how much is hard to judge, most of the shock headlines seem to trace their way back to Alex De Vries who uses some questionable assumptions. He seems to calculate the energy use by taking the price of BitCoin and guessing how much profit is being made by miners and how much they pay for electricity.
Nevertheless it is safe to assume that the energy use is significant, even at half or a tenth of Digiconomist estimates it is still quite large. It is broadly true that the energy use of Bitcoin is related to the value of Bitcoin, via the mining reward.
The arguments against the energy use that are based upon the merits of Bitcoin are not valid. If Bitcoin is useless and doomed it stands to reason that the energy consumed is a waste, but that is true of anything. To argue specifically against the energy consumption you have to make the case that it is not worth the cost even if Bitcoin is a success.
If you assume that the costs and profit for mining are more-or-less static then the mining network will use roughly the same amount of energy if Bitcoin doubles in value every reward halving. That flattens out to about a 20% increase in value every year. To accept that the energy use will continue at this rate you would have to also accept that Bitcoin will increase in value at that rate. To date Bitcoin _has_ increased in value quicker than that, which is why the energy use has grown so high. Hardly anyone would argue that the increase in value will be sustained perpetually. If just holds its value the mining reward and power consumption will reduce accordingly. If Bitcoin ends up worth pennies, the consumption will be minima.
Furthermore, Bitcoin mining adds options to _how_ electricity is generated and consumed. Being able to convert electricity into money not only acts as a stimulus for cheaper energy generation but as a motivator for building new capacity. We are already at a point where Solar and Wind are becoming the choice not only for Environmental but also Economic reasons. Any new power plant needs customers, those customers take time to appear. If a power plant sold excess generation at cost to mining facilities the investment required to build a new plant is greatly reduced. This presents the possibility of developing nations introducing power generation to areas that currently have little or none.
If Bitcoin did end up stimulating the increase of carbon neutral capacity and efficiency, then people in the future might end up making the argument that Bitcoin ended up carbon negative.