I am still waiting to hear how the developers intend to solve the problem of scalability with regard to the energy use for any proof of work network. I have asked that question a couple of times on this message board, and I have also asked it in private correspondence with the team, and have yet to hear any answer.
That's because there is no answer: the dramatic increase in network hashrate strongly suggests, if not outright proves, the theory that miners will deploy hashrate up to the limit of their available power, so "more efficient" miners like ASICs don't lower energy use, they just increase the network hashrate (and therefore the difficulty). It's a situation where nobody but the ASIC miner manufacturers win.