The paper seems to be here, although all I'm getting back right now is long load times and a generic error page: http://www.nature.com/nature/journal/vaop/ncur

The paper seems to be here, although all I'm getting back right now is long load times and a generic error page: http://www.nature.com/nature/journal/vaop/ncurrent/full/nature18593.html http://www.nature.com/nature/journal/vaop/ncurrent/full/natu... I found a copy elsewhere (won't link, piracy, etc) and it does indeed claim an estimated density of 1 megawatt per square metre. There definitely seems to be something missing here, because another paper I found (http://www.ijesd.org/papers/243-B20001.pdf http://www.ijesd.org/papers/243-B20001.pdf) claims a typical reverse osmosis plant processing seawater uses 3-10kWh to desalinate a cubic metre of seawater. They don't have any information on water consumption (which I'm really curious about), but the break-even point on energy consumption is ~333m^3/hr and, as a layman, that seems like a pretty high flow rate for how they describe this working.