They don't mention per-unit cost anywhere, but they're comparing its size to 300-500kW turbines. If memory serves, those are a few hundred thousand a piece? So maybe we can assume these will be ~$300k...?
Supposedly, minimum health and sanitation is 20 litres per person per day. Assuming these last the 30 years stated on Eole's website, and consistently output 1000 litres/day, and not counting any maintenance costs, $200/person/year.
Can the areas that need these things the most afford $200/person/year?
A large UAE desalination plant for comparison : natural-gas fired CCGT piped to an offshore gas field, 1,430 MW electricity, and 380 million liters desalinated water/day (~$1 million/day or ~$300 million/year), presumably from waste heat.
Average per-capita water consumption in the UAE is 550 liters/day  (so ~$400-$600/year?) -- actually among the highest in the world, comparable to the US. The point probably being that their GDP/per capita is also comparable to the US, so the desalination cost isn't that big of a disincentive.
A few other middle eastern, and numerous African countries, with GDPs so small you'll think there was a typo, both desert and not, have serious clean water shortages.
Of course these things will get cheaper, and you don't need to pay $4 billion, you can set up smaller numbers of them.
If that plant can last 30 years as well, discounting any maintenance, that's $4.8/person/year.
And, tangentially, desalination plants need salt water, so only works near the coast. The wind-turbine would work anywhere, regardless of the distance to the coast.
Getting to the point on costs: it's supposed to be around $2/m^3 in the UAE for desalination, so ~$400/capita/year. It would be around $15/capita/year at 20 L/year, but that kind of consumption wouldn't logically exist side-by-side with billion-dollar desalination plants (if your society is developed enough to have built things like that, it's probably wealthy enough not to be counting fractions-of-a-cent aka. liters of tap water.)
Melbourne is a non-desert city in a Western, water-conscious country, that in 2008 was nearing the end of a 10-year drought with water reserves starting to approach the 'panic' levels. The city was on ever-tightening water restrictions and water consciousness was pretty much part of daily conversation. In the end, three and a half million people were using 277 litres per day per capita. What you've forgotten is that many people need industry, and industry needs a lot of water.
I've been trying to find comparisons for other cities. I came across http://www.circleofblue.org/waternews/2010/world/the-price-o... which shows just how hard a comparison that is.
That is, it shows that Boston and Milwaukee use less water per capita than Santa Fe. They get more rain so there's probably less use of water for outdoor use. Boston only used 150 L per day per capita?
That's still a lot more than 20L/d/c.
A curious anecdote - in Melbourne in the '50s and '60s, it was actually against the law to have a private water tank, and this wasn't changed until the '90s, from memory. My mother installed an illegal water tank in the late '80s. It sounds crazy now, but Back In The Day, water tanks weren't well sealed, and Melbourne had a massive mosquito problem. Fast forward a few decades, mosquitos are no longer a problem but drought is, so now you get incentives to install tanks...
I had assumed that that comparison graph used the same measure for all cities. Its Santa Fe value matches that described in http://www.santafenm.gov/index.aspx?NID=168 : "The City of Santa Fe computes its per capita water use as the daily average of annual total water diversions from all sources of supply, less bulk water deliveries to the County of Santa Fe and Las Campanas, divided by the estimated customer population served." That's what you were using for Melbourne, but it seems that that's not what the Boston plot was using.
But Santa Fe has little in the way of industry. "Of the 8,500 acre-feet
billed, the single family residential sector used 4,530 acre-feet (53%), the multifamily sector used 878 acre-feet (10%), and the commercial sector used 2,743 acre-feet (32%). Irrigation use accounted for 330 acre-feet (4%)."
At 5408 acre-feet per year, with a service population of 80,000 people, that's about 230 liters per day of residential use per person, which is still more than Boston's residential use. I think that's attributable to how Boston gets much more rain; 41.5"/1050mm vs. Santa Fe's 13.8"/350mm.
How much rain did Melbourne get during those 10 drought years? It looks like around 650mm, so you can see how Melbourne is a somewhat wetter place than Santa Fe. Santa Fe encourages rain barrels, and I can understand how they would also be effective in Melbourne.
(I also don't know how private well usage affects these numbers.)
(Edit: Ooops! Mixed up Melbourne and Sydney - now corrected.)
One of the articles linked to indicates that this article's statements are misleading. The turbine is "the size of a standard 300-500kW model", but generates 30kW of excess capacity used for pumping and purification.
"efficient desalinization" + building pipelines + building pumping stations + maintenence on infrastructure
for 30 years?
This would obviously need a holding tank of some sort, but in the areas water is really scarce in practice, people don't so much care about getting it delivered to their home. This could be the equivalent of going to a well or stream to fill some containers.