In a major reversal, the world bank is backing mega dams (2024) | Hacker News Reader

In a major reversal, the world bank is backing mega dams (2024)

e360.yale.edu

45 pointsprmph92 comments

Not mentioned in the article (as it is not being funded by the world bank):

China building world's largest hydropower dam in Tibet (reuters.com)

  * Hydro project located on Yarlung Zangbo in Tibet
  * Project to dwarf Three Gorges Dam on Yangtze River
  * Start of construction fuels surge in engineering, related shares in stock market
  * India, Bangladesh have expressed concern about the dam's impact

https://news.ycombinator.com/item?id=44631938

https://www.reuters.com/sustainability/climate-energy/china-...

Pretty sure that's hydrological warfare.

Dammed if you do, damned if you don't

Dam!

No more dam jokes.

This made me chuckle. Well done!

Please can we just get back to building nuclear

$80B is buying 40MW of hydro. What would that get you for nuclear? 4MW?

It’s 40GW of hydro

If we use Vogtle as a cost benchmark you'd get roughly 5 GW (note you typo'd units to MW).

Given these projects will be overseas we shouldn’t use the extreme outlier of Vogtle in the US as the benchmark.

Looks like at least 30GW

https://www.reuters.com/sustainability/boards-policy-regulat...

Without flooding enormous ecosystems and disrupting river flows, and on average half the CO2 emissions per unit of generation than hydro has, and a staggeringly lower land use per energy footprint (hydro is 100x larger, wind is 10x larger). Nuclear seems like one of the only sane choices from an environmental point of view.

Wind's land-use footprint is almost completely non-exclusive.

Except when deployed in wilderness areas that require access roads, staging areas, and electricity networks to be cleared. Which is often the case for wind -- hills and mountainous areas that are inherently less suitable for building and farming.

Much smaller nuclear footprint inside existing industrial sprawl is usually preferable in terms of land use.

There is a finite amount of hydro in the world. They will run out of viable dam locations pretty quickly at this rate.

All the good sites were used by 1940.

The ideal big dam is Hoover Dam. Large, deep canyon in a desert. Narrow, deep canyon dam site. Hard rock geology. No major towns or agricultural areas in the area to be flooded above the dam. That's the best case.

Most later dams are at worse sites.

In which continent?

For the curious:

Cadillac Desert is a great history of American dam building and the Bureau of Reclamation

Except now thanks to climate change the water levels might drop below the intake in lake Mead. Lake Powell is even worse, Glenn canyon dam was always a mistake, and it might stop producing power next year.

Tidal, on the other hand, still has plenty of opportunities.

Make nuclear cost-competitive and people will start building it.

I thought it was cost-competitive with something on the scale of mega-dams and takes about as long to finish too. Except unlike dams, you don't disrupt river flow and cause water-rights disputes (and potentially wars).

> Except unlike dams, you don't disrupt river flow and cause water-rights disputes (and potentially wars).

The water requirements of nuclear power stations cooling systems can cause significant issues. The discharge of heated water back into rivers and the sea is also a major problem.

How major? Would not a temporary lake be sufficient to dissipate heat?

The largest nuclear power plant in the US (Palo Verde) is cooled by evaporating treated wastewater. Water is water and you get about 0.3kWh of electricity for each evaporated kg of water.

The unit economics of dams get more and more competitive when their size increase.

So, hell no, nuclear is not competitive with mega-dams. It's not even competitive with small dams.

You mean the ROI? I'm talking about the cost-to-build.

Nuclear is cost competitive if you have a reliable cadence of building plants and if folks get out of the way of location permits and waste storage and people can actually make decisions about them without endless debates and lawsuits. The problem is one-off designs and the decades long gap between project inception and when investment returns start coming in. As opposed to solar where I could order a few panels and accessories online and start producing energy within a week. Obviously larger solar projects take more planning but if you've got a roof, a credit card, and an electrician on hand you can start producing electricity or expand your production in a very short time achieving break even in a few years.

It was cost-competitive before it faced ridiculously unfair regulations due to being so easy to fear-monger about.

Nuclear just isn't economically viable. Maybe fusion someday.

Tell that to China

China has been building them at a relatively lukewarm pace compared to its solar and wind expansion

> China installed 1GW of nuclear last year, compared to 300GW of solar and wind, Mr Buckley said.

https://www.abc.net.au/news/science/2024-07-16/chinas-renewa...

> isn't economically viable

That is not an intrinsic truth. We have chosen to make it economically unviable.

Most things get cheaper to build with time. Nuclear is an outlier where it used to be affordable and now it isn’t. That’s insane.

why is it different like that?

Mainly because regulators got scared and started adding line items for bogus safety reasons (like expecting to see background radiation levels far below those seen at coal plants due to the fly ash).

Coal plants are dead, partly due to the fly ash. I'm not sure you want to put nuclear in the same boat?

I'm not doing so. I'm pointing out that coal plants were allowed to produce more radiation than nuclear plants were allowed to out of fear of that radiation. Because the public is ill-informed about what "radiation" actually is.

I'm broadly supportive of nuclear fission, but I completely understand more nervousness of it than coal, particularly in the days before we cared much about climate change or industrial pollution. the worry about nuclear isn't low-level day to day pollution, it's disasters. coal disasters are essentially local incidents, and while they can have a big impact politically (f.e., Aberfan), they don't strike fear into people's hearts on a national or international level because the only people at risk to a disaster live very near a coal plant. when Chernobyl surged, half of Europe was affected. yes coal plants can have fallout too, but it's nowhere near as dangerous, even if the standard pollution is worse over time

coal plants were allowed to produce more radiation than nuclear plants because nuclear plants are nuclear plants and radiation management is an intrinsic part of the process. although they undoubtedly should have been, no one was paying attention to coal plants' radiation output, because they were invented before radiation was even discovered, and it's just not the star of the show there, nor is it the first consideration people have when considering coal pollution

Nuclear has never been particularly affordable. It's always been more expensive than coal or gas, even in France in the 70s-80s and China in the 2010s.

Here's how much it cost to build nuclear in France during its golden age: https://ifp.org/nuclear-power-plant-construction-costs. Adjust for inflation and draw your own conclusion.

China's nuclear costs are more opaque, but are estimated at $3B per GW. Again, not competitive.

Ontario produces a majority of its electric power from nuclear (https://en.wikipedia.org/wiki/Nuclear_power_in_Canada) and manages to sell it at competitive rates (https://www.oeb.ca/consumer-information-and-protection/elect...) similar to those in Alberta (https://ucahelps.alberta.ca/your-utilities/rates/historic-ra...) where electricity comes mainly from natural gas after a recent transition from mostly coal. A 2021 report cited at https://www.cns-snc.ca/learn-nuclear/basics-of-nuclear/how-m... finds that nuclear power in Ontario was more cost effective than everything else except hydroelectric power (granted, solar has become much cheaper in just the last few years). And this is in spite of multiple reactor shutdowns.

Yes, running a nuclear power plant is pretty cheap, if you ignore the cost of building and decommissioning it.

The solar panels installed in China during the past 2 years produce as much electricity as all of their nuclear plants combined.

Finland spent 18 years and 11 bn euros to get 1.6 GW of nuclear, the US spent 7bn in subsidies and got some 20 GW of solar in 2022 alone.

Countries going for nuclear will wait decades to get the same power that solar can add in weeks.

Nuclear basically makes no sense at all in 2025.

(For nighttime use dirt-cheap batteries and natural gas now, even cheaper batteries and generated hydrogen gas later).

But what to do in Finland during the winter months? I'm massively pro solar, and I'm sceptical of nuclear, but this seems like a problem to me. Batteries work well on shorter time scales but not over the entire year.

Finland's electricity production is already 95% non fossil fuel, so you don't need to worry about them.

Yes but ~37% of that is nuclear. My question was: If they're not supposed to use nuclear what should they use.

The good thing about northern regions is that they tend to be wet and have low population density. This is pretty good for pumped hydro even if batteries aren't cheap enough at a particular time. But it's not clear when the manufacturing costs of batteries will hit a minimum. So far they continue to decrease.

Maximum elevation in Finland is is 1 324 m and that is in extreme north with less water. In general there just isn't elevation differences that are useful enough.

Short term for self-reliance Finland can use natural gas (won't need much), wind and hydro. Since they are a nation with many friends they can also buy electricity from neighbors.

To solve the variable production from solar and wind, most nations should probably have a safety valve in the form of synthesized fuels. Meaning that during summer when energy is abundant and has to be dumped at negative prices, we use the surplus to synthesize fuels instead.

Synthesizing fuel is inefficient, but since you use surplus energy that doesn't matter.

These are options that are viable right now, but there are also promising developments in batteries that could make them viable for season storage too.

Do you know what the main obstacles are for producing synthetic fuels from surplus renewable electricity production today? It doesn't seem like a lot of companies are doing it at large scale, even while the electricity price difference between summer and winter is large.

I think the main obstacles are the existing players in the energy space. They are huge organisations for the most part and take a long time to change, even if they were motivated to do so. Things have changed rapidly the past few years.

It can also be argued that many of them are not motivated, because they make money from selling electricity from highly valued assets (power plants). If electricity gets permanently cheaper, they stand to lose a lot of money.

There are probably other obstacles as well, but I don't think any of them are insurmountable.

Synthetic fuel production has a very low efficiency, but that does not matter if electricity cost is cheap. So the idea is to overbuild electricity consumption and let them sit idle when electricity is expensive.

But, AFAICT, to overbuild consumption the capital cost is so high that it does not make much economic sense.

> But what to do in Finland during the winter months?

Is this a serious question or "raising concerns".

A quick search brings up wind: https://en.wikipedia.org/wiki/Wind_power_in_Finland

and hydro: https://www.andritz.com/hydro-en/hydronews/hn-europe/finland

And connections with friendly neighbours: https://cinea.ec.europa.eu/news-events/news/cef-energy-finla...

I'm sure that there is a role for Nuclear or gas to cover the last few % that renewables find hard to reach. For now.

It's a serious question. Someone was suggesting that instead of building nuclear, like they did I Finland, they should have built solar. But in the winter in Finland there is not a lot of sun, so I was curious what the commenter, who seemed knowledgeable, had to say about that.

I think the statement is that "Nuclear basically makes no sense at all in 2025."

That's slightly different from "they should have built solar" in 2005 when the last nuclear plant was started, according to https://world-nuclear.org/information-library/country-profil...

It's 20 years later, and "not nuclear" isn't always "go 100% solar".

I guess you mean me, and I can't give a definite answer of course, it's complicated. I can mention some things I think should be considered.

- Solar is bad in Finland during the winter, but good during the summer. Whatever other fuel is used during the summer now can instead be offset to winter. Even natural gas, but preferably imported wind, hydro and such.

- Solar was not an option when Finland decided to build o3. They couldn't have chosen solar instead, back then.

- Finland is a net exporter of electricity, and is in fact exporting a similar amount as o3 produces manually. So basically Finland didn't really need o3 for themselves, they built it to sell electricity.

- When other countries add lots of cheap solar, it will be difficult to sell nuclear since it's much more expensive.

- If the price forces o3 to close ahead of time, Finland still has to pay for it, and its waste, for decades if not centuries.

And at night, nuclear power plants produce infinitely more power than solar. Same during winter months.

Solar simply can't work alone for northern countries without insane amount of batteries. We're talking about having a MONTH of supply in reserve for Germany. It's probably even worse for Finland.

I would add that also a lot of solar power is funded by small capital (homeowners). There is no real way for small capital holders to fund and gain profits from nuclear energy, and installing wind or water turbines are not realistic at 99% of homes.

Great comparison to use the most delayed neuclear power plant constrution in human history to extrapolate an argument. Really fairly done.

Japan builds them in 3 years. USA took about the same during the heights of its use.

> Japan builds them in 3 years

Japan hasn't built a new nuclear plant in 20 years.

>Great comparison to use the most delayed neuclear power plant constrution in human history to extrapolate an argument. Really fairly done.

Not as unfair as disregarding my points entirely just because I used the latest reactor built in europe as an example. ;-)

We can use Hinkley point C instead?

When Hinkley point C is completed, estimated 2029 at the earliest, it will have taken 12 years and will produce as much electricity per year as the solar panels installed in Germany last year. And those panels are already producing, now.

Hinkley point C is estimated to cost 44 bn euros by 2029, and it will still cost money after that of course, for maintenance and operations.

If that money had been spent on battery storage for solar instead, the UK might have gotten way more energy at a way lower cost per GWh, and they would have had it _already now_.

When they are in place, battery banks and solar panels have a basically negligible maintenance cost compared to a nuclear plant. Its an investment where you pay upfront and get benefits down the line.

I don't see how nations that wait 10 years for 2-3 GW of nuclear will be competitive in any energy-consuming endeavor against countries that add 40-50GW of solar every year.

> When Hinkley point C is completed, estimated 2029 at the earliest, it will have taken 12 years and will produce as much electricity per year as the solar panels installed in Germany last year.

I agree that engineering in the UK is a mess, which explains a lot of Hinkley Point C, but just a technical point: people in these conversations often do what you've done and pick a "levelling" stat that elides all of the non-nuclear's disadvantages.

E.g. in this case maybe solar in Germany does produce all that power, but does it do it consistently? And does that cost include the batteries needed to level it out?

It's like saying you're more likely to get rained on in Johannesburg than in London because the yearly rainfall's higher there. It's an accurate stat, but in Johannesburg all that rainfall is concentrated between 3pm and 5pm in summer months, so you're much less likely to get rained on all year round.

>I agree that engineering in the UK is a mess, which explains a lot of Hinkley Point C, but just a technical point: people in these conversations often do what you've done and pick a "levelling" stat that elides all of the non-nuclear's disadvantages.

This is a fair point, I only mentioned the total power produced per year, but it wasn't intended to be deceptive. I wanted to emphasize that solar is not a "toy amount of energy" any longer, which it was as lately as say 2020 in many countries.

It used to be that solar doubled from tiny to slightly less tiny, but it's basically doubled semi-annually for many years by now, and the rice on the chessboard is starting to get heavy, so to speak. China went from "same as all nuclear" to "twice all nuclear" in two years. I think that's remarkable and might have gone unnoticed.

That you can't predict very reliably when you get peak production is for sure a problem with solar, and wind, but it's not without solutions. Solutions exist and are also being deployed already.

And even before we have built enough of those, solar still saves the use of hydro reserves, reduces the burning of oil, coal and gas during the day, and so on.

The leveling argument is often employed the other way too btw - all downsides to nuclear tends to be erased, mainly by pretending there is no cost or risk related to the long-term waste management, or for financial risks related to accidents.

Nuclear is generally operated under a government umbrella that covers all such costs, so they are forgotten/ignored, but they shouldn't be.

> all downsides to nuclear tends to be erased, mainly by pretending there is no cost or risk related to the long-term waste management, or for financial risks related to accidents

I think this is all priced into the TCO of a nuclear plant, as they are very well known.

>I think this is all priced into the TCO of a nuclear plant, as they are very well known.

Fukushima was not well known and is not at all covered by profits from selling nuclear power in Japan. Same for Chernobyl. See also Asse II in Germany - not known. There are many many more examples.

But isn't Asse II in Germany a fallout research centre and waste repository? How can it not have waste priced into it? And Chernobyl was a spectacular failure, and done a long time ago - I'm not claiming that its cleanup was priced into that specific project.

> The solar panels installed in China during the past 2 years produce as much electricity as all of their nuclear plants combined.

Because it doesn't have very many nuclear power plants relative to its size? France has the same number of nuclear reactors as China despite being a much smaller country.

I'd argue 50-60 nuclear power plants having the same energy output as millions (billions?) of solar panels is a win for nuclear - it's much higher energy density, much smaller environmental footprint, much smaller infrastructure investment, etc.

Worth mentioning that while France has 57 operating reactors and China has 58 operating reactors, it's also the case that China has an additional 38 under construction (with plans for more again) compared to France having zero additional reactors under construction (despite announcing plans for new reactors three years past).

* https://world-nuclear.org/information-library/country-profil...

>I'd argue 50-60 nuclear power plants having the same energy output as millions (billions?) of solar panels is a win for nuclear - it's much higher energy density, much smaller environmental footprint, much smaller infrastructure investment, etc.

I don't think that's correct. The infrastructure investment is clearly much much smaller for solar, in practice. IEA and other organizations have observed that solar is the cheapest source of electricity that humankind has ever developed, and this was already a few years ago, when it was more expensive than now.

Consider that several countries are adding the equivalent of several nuclear plants of energy generation yearly by now. Germany, Japan, Canada to name a few. Adding the same capacity with nuclear would be a budget-defining decision for years.

Solar is just so much cheaper and faster to make that nuclear becomes "too little too late" by comparison.

If any nation could decide to make 100s of nuclear plants to match the output from solar it's China, but it just doesn't make sense. It makes way more sense to invest in energy storage to stabilize the massive amounts of energy from solar. China does that too.

Nations that have nuclear weapons will of course keep nuclear plants around anyway, but it is really really hard to make a case for nuclear just for energy supply in 2025.

Sorry if I'm mistaken but I reckon solar panels (and especially batteries) produce much more waste. Also they require vast areas for the same energy. You should keep all the variables in the equation and not just say how quickly you can dish out some panels.

Probably not reliable but this is what ChatGPT outputs over 100 years, assuming equal output (100 TWh total):

    Type                 Nuclear_Waste(tons)    Solar_Waste(tons)
    -------------------------------------------------------------
    Solid_Total          12_000–23_000          6_000_000–20_000_000
    Hazardous_Toxic      2_000–3_000 (HLW)      600_000–1_000_000
    Longevity            >100_000 years         ~500 years (some metals)
    Land_Required        ~0.1–0.2 km²           ~5–10 km²                 For 1 TWh/year of continuous output

I still dream of a future where nuclear batteries can be fitted in every item that needs it, but we can't get there without development. There's only so much energy a square meter of panels can output.

"Waste" is a nebulous term; it usually means "is not currently recycled". But you can't build a waste-recycling industry until after there's lots of waste to recycle, for simple economics of scale. Using waste as a justification to not build new stuff is just stupid.

Batteries especially are just absurd - they're ~10% lithium (and it's mostly in the electrolyte, which realistically means the electrolyte is 100% electrolyte, excuse the tautology), whereas 'lithium ore' is mostly 1-3% (there's some higher, even as high as 8%, but it's mostly 1-3% IIRC). With sufficient scale, that stuff will disappear like scrap copper left on the curb for an hour.

I want to highlight two things as counterarguments - 1) nuclear waste is not comparable with solar panel waste and 2) Land use for solar is not comparable with land use for nuclear.

Solar panel waste does not require army supervision to prevent it being used for terrorist acts. The US army has personell permanently stationed at plants that have been closed for several decades by now. They keep costing money for decades after they stopped producing any power.

As of 2025 there isn't a single nuclear site that has ever been in operation that has stopped costing money for the population of the country it is in, simply because of the waste. And there is no end in sight.

As for land, solar panels are usually deployed on land that can still be used for other things. (Rooftops of homes and office buildings, grazing grounds for sheep and farmland for crops that need shade).

The last point is of course why many countries have been able to deploy solar that matches the output of their nuclear generation in just a few years. You have hundreds of thousands of carpenters and electricians that can work simultaneously on building solar panel installations and they get approved by homeowners without any bureaucracy.

Looking at China, the US, the EU, Japan and even a nuclear pioneer like Canada, you see that Solar adds the equivalent of several new nuclear power plants per year, and the power is available immediately.

The only argument with merit is that nuclear works at night and during winter - but so do many other things, much cheaper things, things that don't take a decade+ to build and don't require eternal expensive vigilance.

Nuclear fuel is cheap, if you can somehow reduce the capital costs associated with building them, they are a good base load solution. Couple them with hydro storage and you have the ability to deal with demand spikes. I don’t think China is going to stop investing in nuclear or solar anytime soon.

Question to you. Why do you think Olkiluoto 1&2 took 5 and 7 years to build instead of 18?

Nobody is arguing anyone should build reactors the way Olkiluoto 3 was built.

>Question to you. Why do you think Olkiluoto 1&2 took 5 and 7 years to build instead of 18?

I can only speculate, but I suppose two major reasons could be that Europe had several companies that could and did build reactors back then, and that o1 and o2 were much smaller than o3.

O1 and o2 were also built before we knew about things like Chernobyl and 9/11 of course.

>Countries going for nuclear will wait decades to get the same power that solar can add in weeks.

And it will last 80 years, day and night. Solar can give you what, 20 or 25 years and 12 hours each day? And China can't hamstring your country either by just refusing to sell you more solar panels. There is almost certainly a place for photovoltaic in a nation's power grid as a sidekick to some other more serious technology.

>For nighttime use dirt-cheap batteries and

Also made in China.

This comparison makes no sense at all since both factors: the huge delays for nuclear and small delays for solar are direct results of policy. So countries going for nuclear will do the going by removing those hurdles

>This comparison makes no sense at all since both factors: the huge delays for nuclear and small delays for solar are direct results of policy

It's not policy it's physical logistics. Just building the roads to the site where you want the nuclear plant takes months and it's hard to transport anything else to the site before you have those.

You also need to sign thousands of workers who need to be physically on location, meaning you need housing, parking, plumbing, salary administrators, doctors, catering, janitors etc.

Solar is entirely trivial by comparison. Any small town has enough carpenters and electricians to assemble megawatts of capacity in weeks. Which is why it grows so much faster than anything else.

The nuclear industry was destroyed in the 2000s. You can try to rebuild an industry with 100s of billions of dollars…

… or you invest that money into renewables and battery technology.

We have cheap fusion. A giant reactor in the sky. Solar should be massively scaled, along with battery capacity and pumped storage. China is crushing in this.

Hard agree on pumped hydro and other forms of grid-scale storage, but can you understand concerns around batteries? There are environmental ones on the mining/metals side, but producing and disposing of them in a clean manner is often hard. Getting them from anywhere save mostly china is hard (if you want large, dense, affordable, and grid-scale options) and depending on somebody who's nobody's geopolitical friend is probably a bad idea.

Ditto for the panels themselves.

> but can you understand concerns around batteries?

No I can't. Just recycle the batteries, and you've solved both concerns in one go. Lead acid batteries have a >99% recycling rate, the economics for recycling EV & grid storage batteries are even better.

"Just Recycle the batteries" is a massive understatement of the effort involved. The economics of Recycling ev batteries is autrocious. (They're glued into the frame of the car (differently for each model), needing manual dissasembly to not risk fire)

It took 20 years of standardisation and effort to raise lead acid batteries to 99%, and they're as simple as batteries get. Large scale recycle of litium batteries (including the cobalt and nickel) requires changes in how batteries are made to be (either or both) less energy dense and more expensive.

Pumped hydro is the best bet for gridscale. And i'm hoping sodium batteries roll out for EVs within the near future.

Sand batteries provide a very nice scalable solution IMO. https://polarnightenergy.com/sand-battery/

Sand batteries aren't batteries - they output heat instead of electricity. You could use them to feed a steam turbine, but then you're paying for the steam turbine.

That said, sand batteries are amazing for the heat users who can use them - just plug em into some off-grid solar/wind (and/or on-grid, and make money by buying electricity when the prices go negative) and you get 24/7 heat for the price of intermittent renewables.

Tesla (for one example) has massively scaled their battery recycling tech, they are largely recyclable at this point. I haven't read as much on the recyling of the panels themselves. But they have long lives at this point, and there will always be SOME environmental cost to large-scale energy production. Maybe ASI will arrive and give us practical fusion reactors sooner than we think.

Hydro is a great complement for solar, you can decrease flow when the sun is shining, and increase it when it isn't.

Yep, and if you can pump water back into the lake with excess solar power when the sun in shining, you now have a giant storage battery as well.

Very situational on where it can be used, and requires some very careful cost calculations.

Ignoring the local effects of their construction, a damb breach is one of the worst man-made disasters possible. Mantinence and error margin must be very very carefully accounted for. There is a reason the world bank stopped funding them, and it wasnt purely enviromental. (Some badly managed projects led to expensive and dangerous situations)

So when relevant it's most powerful energy source avalible. But the list of preconditions and caveats is massive.

Consistent rainfall in certain areas seems to be an issue with climate change , does your mega damn rely on snowmelt for topping up, that may start to be a problem?

The sun doesn’t have this issue. It's ubiquitous.

During the day. We still need storage.

Batteries and nuclear my man.

Reply on news.ycombinator.com