So best case, 200 GW will only produce 525 TWh annually. This is equivalent to a 66 GW nuclear plant (nuclear has a 90% capacity factor).

Out of 20000 TWh electricity produced globally, that's only 2.5%.

Sure. But these days, you'd be hard pressed to build 66 GW of nuclear for $200 billion. It's costing about US$30 billion for one 3.2GW nuclear plant currently under construction in the UK (Hinkley Point C). And it's probably a lot more than that by the time you account for eventual decommissioning costs and long term waste storage.

They have 19 reactors under construction right now that will add 21GW +/-, and 50-60GW of additional nuclear energy planned for the very near-term. Their goal is to add roughly 100GW+ of new nuclear energy output in the next 12 years.

[0] https://en.wikipedia.org/wiki/Gemasolar_Thermosolar_Plant

But it was very expensive.. about $360M for 20MW (or $18B per GW). The price they give -- $200B for 200GW -- would rule out this type of plant.

It's almost certainly PV.. which is currently $1/Watt for new utility scale plants. The price matches perfectly.

But yeah, seems more likely to be PV at this investment price (assuming this will be the total for the project, which I kind of doubt, but I dont know how much costs will inflate to), would be great to get more details about what they are proposing.

[0] http://www.dailymail.co.uk/sciencetech/article-1393879/Gemas...

"1) regarding price, it's probably PV at those prices, as CSP today comes in at least 2x of PV installed capacity cost and that's taking a very aggressive estimation- although one could probably get CSP costs down to $1/watt with the design and procurement scale economies being talked about. 2) Regarding Venning's point about demand, even with significant amount of energy storage, this would create quite a bit more peak power on sunny days than their current grid demand can handle so announcement may be based on a couple (dodgy to me) assumptions: a) either they assume building it out in phases slowly over time to accompany some aggressive assumed demand growth or b) they are thinking about power exports in a region where Saudi Arabia probably doesn't have too many politically viable customers"

That doesn't sound right. According to the (better) article from Bloomberg [1], "Saudi Arabia’s electricity generation capacity ... stood at 77 gigawatts in 2016", which that article points out is a third of this new project's.

I don't work in electrical production, but I don't think this capacity will work for 10% of the world. Wikipedia [2] agrees with your assessment of ~2300GW for worldwide average production; but I think the problem is how an average production number is calculated as compared with quoted output numbers for a given installation.

EDIT: Hmm, if Saudi Arabia produces ~3.3% of the world's electricity currently, then maybe this does represent enough for a substantial percentage of the world's population, assuming the production and capacity numbers are comparable. (Deriving 3.3% from 77GW of 2311GW, which I recognize come from different years.) Anyone who knows more about this care to weigh in?

[1] https://www.bloomberg.com/news/articles/2018-03-28/saudi-ara...

[2] https://en.wikipedia.org/wiki/Electricity_generation#Product...

- The number of hours in a year.

- The percentage of nameplate capacity that the system can be expected to produce on average, e.g. the capacity factor.

There are 8760 hours in a year. A solar PV system in a sunny region may have an annualized capacity factor of 25%. (Installations in really sunny regions may achieve upward of 30%, and Saudi Arabia certainly has a lot of sun, but dust offsets some of the benefits of sunny deserts.)

200 * 8760 * 0.25 = 438000 gigawatt hours of electricity generated per year, e.g. 438 TWh, or about 2% of present annual global electricity generation of 21,000 TWh.