How much energy and money does it take to produce the nanotubes and power plant and how does it compare to the electricity and money that the nanotube power plant produces over its life?

This is very interesting especially when read along with today's (2012.3.13) story about Lockheed's graphene-based fresh water reverse osmosis breakthrough : http://news.yahoo.com/pentagon-weapons-maker-finds-method-cheap-clean-water-051529904--finance.html
The reverse osmosis process will produce some high salt concentration waste water along with its fresh water and the Statkraft system can be used to generate the electricity to power the reverse osmosis pumps. Now that's what you call a win-win :)

I am a techie and this is great, if it works on a commercial scale it will have fantastic ramifications world wide both from an electrical generation perspective and a human food perspective.
If it works electricity will be renewable and cheap. Seafood will also be rare and expensive. Remember where the fresh water and the salt water mix is now call a marine estuary or a mangrove swamp. In the future it will be a power plant. Marine estuaries like mangroves are the nurseries of the oceans. No places on the planet are as important for the health of the oceans and we are now contemplating turning them into power plants. These power plants WILL destroy the mangroves and the estuaries.
Seafood is the last relatively cheap source of high quality, extremely tasty, high volume protein left on this crowded planet. All you need is a good boat, net and sonar to make a good living. You don't need to buy the land, pay land taxes, plant the crop, fertilize it, protect it from pests you just buy the combine and harvest.
We really need to think about what we are doing. When we do the cost benefit analysis on this technology through an environment assessment in whatever country has the money to build one of these plants please let us not assume that the loss to the marine ecosystem will be negligible and be made up for by the mangroves in Mexico or Bangladesh as these countries will be looking at this as a panacea too.
Up until I read this I thought the only really serious problem mankind faced was converting all our good cropland to asphalt.

"The global osmotic, or salinity gradient, power capacity, which is concentrated at the mouths of rivers, is estimated by Statkraft to be in the region of 1,600 to 1,700 TWh annually."
You often get "estimates" of the total solar power of the planet which envision every square meter of land being a solar panel (i.e. abstract total power which means nothing as far as real potential goes but makes it sound really promising). I wouldn't be surprised if the global osmotic capacity was based on every ounce of water entering the sea being run through the process. I hope this technology does well but without subsidies from governments who are easily swayed by such "estimates".

30 mWh per year. Thirty milliwatt-hours / year / sq.m. Really? The average household in the United States uses about 8,900 kilowatt-hours of electricity each year (easy to google). So it will require 300 sq.km of membrane to power 1 household.

Splatman, I'm thinking they meant a 30 megawatt hours per year - should the 'M' have been capitalized? Debatable, the 'k' in kWh (kilowatt hour) is not. So assuming they did mean 30 megawatt hours, then using your figures the average US household would need 8.9 sqm.

I can't imagine this ever becoming economical. This is essentially a reverse osmosis filter run in reverse (forward osmosis?) So unless your filtering the incoming fresh water down to micron levels before exposing the membrane to it, your membrane will be clogged before you know it. And filtering huge quantities of river water with fine filters would cost a fortune.

Am I missing something here, or can you have just two tanks of water: one freshwater, one saline water? why do need a river and the sea? The amount of energy available seems to be extraordinary. I wonder what these filters cost per square metre.

Devil is in the "dirty details" & I'm with Siegfried on this. the Nanorods are only a few dozen nanometers long, measurements similar to those are used to discribe distances on microchips. What often works on the small scale doesn't work well on a large scale & this is about as small as it gets. I doubt it will work very well on the large scale but I hope I'm wrong. If it does work it would be best implemented to reduce the cost of osmosis purification systems used to make drink water.

Would it make a difference if such a plant were utilized on a body of water with a higher salinity, such as the Dead Sea in Israel or the Great Salt Lake in Utah?