back
54 comments
Some more background here: https://www.greentechmedia.com/articles/read/saudi-solar-bid...

PV has become a lot cheaper in recent years, and Saudi Arabia has plenty of sun, but these bids are so low that they may require behind-the-scenes support. Or they might not come to fruition. India has recently had similar problems where bidders tried to guess how fast hardware costs would continue to fall when bidding on future projects, and guessed too optimistically.

https://www.greentechmedia.com/articles/read/indias-renewabl...

It's a "feature", too optimistic bidders always get the bid as they give the best price. That's part of the reason this sort of bids go wrong so often.
Correct, initial bids rarely match the final cost in this part of the world. They also are highly dependent on the level of detail and soundness of the specification/scope that was bid for.
> these bids are so low that they may require behind-the-scenes support

I thought the gap between the lowest and next-lowest gap was interesting: $0.01786 vs $0.02342. Made me wonder what the lowest bidder's advantage is.

Having worked on 2 multi billion dollar projects in Saudi and UAE the price the bids are in very rarely is the final total cost of a contract or purchase order. Ongoing changes, corrections, under preforming suppliers that have underestimated the complexities and cost are always present. The client will need to come up with ways to pay these additional cost if he does not want to sit on a half built plant. It is not uncommon that contractors go bust under this price pressure and another one with a more realistic cost will come in to save the day. Problems like this often involve Saudi or UAE government or royalty as the risk of loosing face is hight. Things then get smoothed over with a cup of tea behind closed doors.
Those same officials who will smooth over the future problems are happy with the press today. SA is coming off today as a herald of the green future. That's good press regardless of eventual outcomes.

(Remember also that SA is currently involved in a messy war against a neighbour. We must assume that any and all news is being carefully managed.)

I had the exact same thought. Two ideas:

1. They're well capitalized enough that they think the actual cost to produce is going to drop significantly. Maybe they're planning on a 50 year lifecycle for this solar plant and thinking: we'll we'll lose money now, but when we have to replace all these panels in a decade, at that point they'll cost nothing and for the remainder of this project we'll make bank (though I'm not even sure if that's how this works)

2. They're somehow financially co-mingled with a solar cell producer and have a significant hidden cost advantage.

Or a mixture of both most likely. It might be they’re already there, too.
I would be extremely wary when the lowest bidder is that much lower than all the other competitors. See: legions of failed highway projects in the US.
> Saudi Arabia has plenty of sun

This could be a problem as above 25c panels can lose efficiency so temps in the middle east could drop panel efficiency quite a bit.

Given a constant level of illumination, solar cells perform better at lower temperatures. But for higher total generation it's better to install panels in hot-and-sunny regions than cool-and-cloudy ones, even if the panels in the latter region convert a higher percentage of light to electricity. Cold-and-sunny conditions are harder to come by, but on (e.g.) a clear winter day at noon, with cold air temperature and plenty of sun, panels can perform beyond their nameplate ratings.
So where is the most efficient place for solar panels, given that the solar panels are heated by the sun, they could be coated to reflected IR. A cooler climate could help heat be convected away. High altitude deserts like the Atacama might be ideal, but there's less atmosphere for convective heat transfer.

I do love an optimisation challenge.

Sorry - I must be missing something.

300 MW = 300,000 kW

6 hours of generating capacity per day, 365 days in the year, gives 2190 hours of generation per year. 2,190 hours * 300,000kW * 0.02USD gives an annual cost to Saudi Arabia of $13.14M a year, for 300MW of installed capacity.

A modern panel is around 250W, so they are installing and running 1.2M panels, for ~$11/year per panel, including grid connection and transformation costs? Insane!

The wonders one can achieve in a country that still uses slaves...
Is the United States doing this? I know we have literal slavery in prisons.
Utility panels are quite often 350W these days. But yes, still insane.
As a complete noob - Is there a place I can find data on panel prices? I was just reading this article: http://www.livemint.com/Industry/sMC62YoWv4LybaPZnKivMM/Indi...

and the prices seem very very low.

I got my pricing from AliBaba - commercial rates seem to be $0.3/Watt to $0.4/Watt - so roughly $100USD for a 250W panel
Why 6 hours? Not disputing it, just assuming I'm missing something.
The math starts to make a lot more sense at 8 hours a day and 350W panels, that's for sure.
6 hours was just a ballpark to compensate for weak sunlight early/late in the day, bad weather, some panels getting damaged, loss due to dust, etc.

Not sure what the real rate of generation would be, and I'm happy if someone with actual solar experience can clarify my estimate.

One use of cheap power is to do very energy intensive operations. Egypt powered fertiliser plants from the Aswan dam, and Iceland uses hydropower to smelt aluminium.

Saudi Arabia is building a huge aluminium smelter, I had thought it would be solar powered, but they would need far, far more power, and of course it couldn't run at night.

http://www.bechtel.com/projects/ras-al-khair-aluminum-smelte...

So I assume this is gas fired power, which seems strange.

Natural gas is effectively free for the Gulf states as it’s a byproduct of oil drilling.

Compared to oil it’s harder to store and transport, and there aren’t any gas pipelines from the Gulf to neighbouring areas (unlike say Russia). If you fly over the Gulf at night, you’ll see the sea lit up from oil rigs just burning it off, so it makes sense to use it to fuel heavy industry.

The primary constituent of natural gas is methane, which has a much worse greenhouse effect than CO2, so it’s ‘greener’ to burn it rather than just releasing it into the atmosphere.

The Ma’aden Aluminium refinery and smelter complex is the world’s largest, lowest cost fully integrated aluminum facility.[1] The smelter as well as the alumina refiner have been operational since 2014.[2] [1] http://www.maaden.com.sa/en/business/aluminium [2] https://www.alcoa.com/saudi-arabia/en/default.asp
It's interesting that we don't favour the idea of something only running in the day when transitioning to renewables. Would it be such a bad thing and would there be other operations they could happen during the night ?

I know it might mean less money or whatever, but the universe doesn't seem to run on money so it's weird logic sacrifice the environment over something like operational uptime.

There are some processes where equipment lifetime is more determined by total usage than by start-stop cycles. Those processes can more or less seamlessly transition to intermittent renewables as long as the time value of money isn't very high. Given persistently low inflation rates despite years of QE, I think we might be there already. EDIT: actually, even if the processes are tolerant of frequent stop-start, the associated labor scheduling may not be. Not knowing when you'll have a shift scheduled sucks for workers. Not knowing if your regularly-scheduled workers will have favorable weather to do useful work when they show up at the plant would suck for employers.

Other processes are intolerant of complete shutdown/startup cycles. Modern aluminum smelting is one of them. If everything freezes up in the production cells because of too little power for too long, it causes damage and it takes a long time to start up again. To make aluminum smelting more cycling-tolerant would require rethinking the whole plant design. Even then the cost savings from cheap but non-dispatchable renewables may not be enough to compete with existing plants powered with dispatchable electricity.

The Great Wikipedia[0] says "power must not be interrupted for more than 4-5 hours, since the pots have to be repaired at significant cost if the liquid metal solidifies."

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

Here's a crazy idea - can't you smelt aluminium using a giant magnifying glass?
It is quite telling that this solar project is taking place in Saudi Arabia, the worlds most profitable producer of fossil fuels. You know, the fossil fuel companies all see where things are going. It is only American conservatives who are in deep denial.
> It is quite telling that this solar project is taking place in Saudi Arabia, the worlds most profitable producer of fossil fuels.

Is it? I always thought of it as smart planning. They are taking the profits from a limited resource and diversifying.

Norway has been doing this for decades, albeit not directly with PV.

In my comment I am agreeing with you, it's smart planning. The fact that renewables are going to replace fossil fuels is so obvious that even fossil fuel producers, who you would expect to deny it, are instead acknowledging it.
Not just Americans, Australians too.

    > US$1.786 cents
There's got to be a better way of writing that
Since that's just flat out wrong, yes. It's either $0.01786 or 1.786¢ (if you need to specify the US, then definitely the first with either “USD” or “US $” in place of the plain “$”.)
It would be more logical to to put units at the end, as the SI norms say.

eg: USD 24k is a bit stupid, whereas 24 kUSD is more clear. It's even worse with things like 1/(USD 800/kWp) which gives things like:.8kWp/USD 1 ?

The article title is flat out misleading too.
USD0.01786
How much of this is due to the near-slave working conditions and compensation that imported manual laborers in Saudi Arabia have to deal with?
Not much. Even in the relatively high-labor-cost United States, only 12% of utility scale solar project costs are attributable to installation labor. See table 14 on page 56 of this report:

https://www.nrel.gov/docs/fy17osti/68925.pdf

Thanks for this, I didn't know it had fallen so fast!
I can imagine that if proof-of-work remains a prevailing feature of popular blockchain systems, setting up e.g. bitcoin mining operation off near-free solar would be a nice passive income stream.

How fast can modern ASIC hardware pay for itself if mining?

I already operate a small scale operation near my hometown in India which is almost self sufficient (powered by wind and solar). For deficit, we have connection to the grid but it is rarely used save for rainy season.
Very interesting. Can you please provide more details on your setup. I would love to know more as I am interested in doing something similar.
Power is largely fungible, though. It doesn't merely need to pay for itself eventually, it needs to pay better than simply selling the power.
It's not a completely crazy idea. Building transmission infrastructure to connect remote solar farms to load centers can be a significant part of the project cost and timeline. Crypto-mining on a detached solar farm could export its financial value over a low-capacity wireless data link instead of heavy wires and towers. It might have more siting flexibility than projects that plan to deliver electricity to existing load centers. Of course it would just intensify the red queen's race among the crypto-miners.
the less fossil fuel they use themselves, the more they can sell.