Hacker News Reader

Comment by pfdietz | Hacker News Reader

Breeding is a technology looking for a business case.

It's more expensive than just using fresh uranium in current market conditions. It's a way from keeping future uranium shortages from making nuclear power more expensive; it's not a way to make nuclear cheaper than it currently is.

View on HN

It also apparently provides a way to make reactors that don’t depend as much on water so they don’t all have to be near the coast.

This would allow Western China to also develop reactors to help underpin their renewable and coal energy.

> The interest in MSR technology and Thorium breeding did not disappear however. China's nuclear power production relies heavily on imported uranium,[10] a strategic vulnerability in the event of i.e. economic sanctions. Additionally, the relative lack of water available for cooling PWRs west of the Hu line is a limiting factor for siting them there.

https://en.wikipedia.org/wiki/TMSR-LF1?wprov=sfti1#History

> also apparently provides a way to make reactors that don’t depend as much on water so they don’t all have to be near the coast

Non-water microreactors broadly fall into two categories: ones using a different moderator, most commonly sodium, a sodium salt or helium; and those using heat pipes. Most microreactor designs don’t use water.

Nuclear plants don't need more water than a coal plant of the same power, they both use the same steam turbine with water as cold source.

Emphasis on current market conditions. Relations with uranium mining countries and environmental opposition to uranium mining could shift conditions.

The truth is that nuclear power is not that financially attractive at the present and would the price of uranium rise enough that breeders would become economically viable most countries would just stop bothering with nuclear power altogether.

> The truth is that nuclear power is not that financially attractive

Let me fix that for you: "The truth is that nuclear power is not that financially attractive in the bureaucratic high cost litigious Anglo-sphere". And that's pretty much all infrastructure these days, unfortunately.

They’re not financially attractive in other parts of the world either. China, a zero litigation single party state, is building some but a tiny % compared to their renewable buildout

They need a lot of energy from a variety of sources. China has 30 or so reactors under construction (half or so of all active projects).

"China currently has 58 operable reactors with a total capacity of 56.9 GW. A further 30 reactors, with a total capacity of 34.4 GW are under construction" [1]

So, yes, but...

China installed 256GW of solar in the first 6 months of 2025 [2]. A full year estimate of ~350gw. So, the total of all nuclear under construction is 1/10th of the solar they installed in one year.

Don't get me wrong, its cool to see diversity of non fossil sources, glad they are building some, but its a niche in their overall energy buildout. And they can only build that small niche because they dont have to be market priced, its state subsidized.

[1] https://www.world-nuclear-news.org/articles/ten-new-reactors... [2] https://ember-energy.org/latest-updates/global-solar-install...

Comparing nuclear reactor capacity to solar capacity is misleading because renewable capacity dramatically overstates actual generation. IIRC The capacity factor for solar ranges between %5-%25 of total capacity generated.

Yes yes, one of the usual reflexive context free points repeated every time solar comes up. Whatever the actual capacity factor is(5% is not a serious number ), I’m sure chinas energy planners know that, it’s hardly a gotcha. And still they’ve choose to build solar at a volume massively dwarfing nuclear

(Edit: cycomanic explained it much better and more patiently than me)

I don't think it's reflexive to point out that evaluating the headline numbers like that is misleading. People who are new to the topic will misinformed if they think you're describing an apples to apples comparison.

The issue is that even with your correction the point stands. You weren't refuting, you were being irrelevant.

That doesn't significantly change the argument. Most solar plants have capacity factors of around 20% (5% might apply to home systems, but not commerical), compared to nuclear which has around 80%. So a factor of 4. So numbers change a bit on the previous poster, China just installed 3x more solar in 1 year than all the nuclear under construction, or they essentially installed the same amout of solar in one year as all existing and under construction nuclear combined. And if we look at projections, next year they likely will install twice as much solar...

While China is often put up as the poster child for nuclear power, they are actually a great example of how nuclear is being overtaken by renewables. China's 2019 plan was that by 2035 nuclear would account for ~8% of generated electricity (up from ~5%). Since then percentage dropped to 4.5% (and the drop seems to be accelerating). Unless something dramatically changes nuclear will account for less than 4% (not the planned 8%) of generated electricity by 2035. All that is due to the raise of renewables (largely solar). I suspect we will not see China build close to those projected 200 GW and the percentage to be even lower, just due to the exponential growth in solar.

source: https://en.wikipedia.org/wiki/Nuclear_power_in_China

As if their massive solar production (and innovation) is market based...

It's not the litigiousness that makes it expensive. France was producing nuclear power plants at a cost per watt that nearly matches modern China. In fact, the mind-numbing cost overruns seem unique to the US.

Here's a Nature article about it:

https://archive.ph/Tpe0j

Seems to me like it's more of a story of corruption than of over-regulation

france cant do it any more either. Flamanville was 12 years late and [1] 400% over budget. EPR2 is already delayed and over budget and they havent even started building yet!

UK cant do it either, see hinkley point c [2]

[1] https://www.nucnet.org/news/long-delayed-nuclear-plant-conne... [2] https://www.world-nuclear-news.org/articles/edf-announces-hi...

That might be somewhat true but Flamanville was still about $4/watt while Vogtle 3 and 4 (which were built around the same time) were about $15/watt. It's still hard to place France and the US in the same bucket. The US really is uniquely inept at nuclear costs

Flammanville 3 was about €14/W, or about $16/W.

https://mastodon.energy/@Sustainable2050/113695635521714572

The UK does the same thing. In fact, its across the entire west. Its almost as if absurd over-regulation is expensive. The Vogtle plant construction for example had to deal with 3 different tranches of changes to the design caused by regulators. Its not corruption, its over-regulation. If it is corruption, it is corrupt politicians intentionally over-regulating because their backers make lots of money extracting FFs.

The cost of nuclear power is almost entirely capex and financing, not opex. Uranium input cost for nuclear power plants is 0.5c/kWh. With breeders you can divide that by about 100.

At least as of a couple years ago nuclear costs just a little more than solar plus storage and that’s not stopping anyone heh.

Capex and financing is still an issue for many countries, and the opex is a non-zero commitment beyond just the fiscal portion. Most countries that pass-over nuclear energy are fairly justified in their decision. The status-quo is still not super psyched about nuclear proliferation.

There is room to change that, but the cards are very heavily stacked in China's favor. America's bad at the financing part, fickle when it comes to enforcement & supply chains, and ostensibly 2 days away from bailing on the IAEA itself. The proliferation-resistance of Thorium reactors gives China an export trump card that America will struggle to match.

With recent price drops of solar and storage the difference is now multiples.

This is just plain false. Learn the difference between capacity cost and utilization cost.

It is LCOE outside of the US tariff bubble.

Go ahead a calculate what co-locating ~$52/kWh BESS systems alongside an utility scale solar install costs per kWh.

https://www.ess-news.com/2025/06/26/china-energy-engineering...

China has more uranium reserves and less thorium reserves than the US though

Most thorium: India, Brazil, Australia, US, Turkey

Most uranium: Australia, Kazakhstan, Canada, Russia, Namibia

They highlight less the advantages from breeding, than other advantages of the molten salt design, like not needing a lot of cooling water, which allows this reactor to operate in the Gobi desert, the possibility of replacing the fuel without halting the reactor and various safety features.

Nuclear reactors don't need a particularly big amount of cooling water.

The thermodynamic cycle needs a cold source though, and it's most commonly water. This doesn't depend on the reactor design and this is equally as true of coal plants.

As long as you are making electricity out of a thermodynamic cycle, you need a heat source (be it a flame or a nuclear reaction) and a cold source.

As the reactor is operating in the Gobi desert and China claims that its main advantage for them is exactly this possibility of operating in the inland arid areas of the country, unlike their current reactors that must be installed only close to the sea, in the part of the country with abundant water, they must have a solution for the cold source that does not involve water.

Perhaps they use as a cold source the underground soil, though the soil thermal conductivity will limit the amount of power of the reactor. This reactor has a modest power, which could be explained by this constraint.

If the reactor is as safe as they claim, the moderate output power per reactor could be compensated by installing many such reactors.

> As the reactor is operating in the Gobi desert and China claims that its main advantage for them is exactly this possibility of operating in the inland arid areas of the country

This is mainly a feature of the reactor being small. If you don't have much heat to dissipate, even air cooling becomes feasible.

> unlike their current reactors that must be installed only close to the sea, in the part of the country with abundant water

In reality even current water-cooled reactors can be pretty efficient in terms of water use if you design the cooling system with that in mind. See the Palo Verde Nuclear Generating Station in Arizona.

> Perhaps they use as a cold source the underground soil

I'm not sure this would work, as you'd be storing heat in the soil without a real heat drain so the yield of the plant would decrease until it reaches zero.

For small reactors air or radiative cooling are an option though.

> other advantages of the molten salt design, like not needing a lot of cooling water

This advantage is conserved by all non-water moderated reactor designs.

The use of water for moderation is one thing, the use of water for cooling is another thing, even if in many reactors water is used for both purposes.

A reactor can be moderated with something else than water, e.g. graphite, but it may still need water for cooling.

The amount of water needed for cooling is much more than needed for moderation.

So there is no doubt that many "non-water moderated reactor designs" still need copious amounts of cooling water.

Any "non-water moderated reactor design" that does not have liquid fuel, i.e. it is not a molten-salt design, must have a cooling fluid, though the fluid in the primary cooling circuit may be not water, but something else, e.g. molten metal (e.g. molten sodium) or supercritical carbon dioxide.

I believe the point was that non-water moderated designs typically operate at higher core temperature than LWRs, so they can reject waste heat at higher temperature (or reject less waste heat per unit of electrical energy produced), and that makes rejection to air more practical.

A very high temperature reactor might even be able to work with an open air Brayton cycle system, which would allow heat to be directly exhausted in that air stream. It would probably still need an in intermediate heat exchanger so the air wasn't being irradiated with neutrons.

> very high temperature reactor might even be able to work with an open air Brayton cycle system

Is anyone doing that? Everything I've seen is Rankine.

No, because it would require very high temperature. The air coming out of the compressor of a gas turbine would already be hotter than the water/steam coming out of a LWR. It would likely involve a core temperature of around 1000 C. The turbine inlet temperature of a modern jet engine is at least 1500 C. There would likely be thermal NOx production, so post treatment of the gas might be necessary in an open cycle system.

Westinghouse plans to with their eVinci reactor. It will use an open-air Brayton cycle: https://www.nrc.gov/docs/ML2409/ML24092A122.pdf

There is no business case for basic research, but if you stop basic research long enough you will have no business. The United States and its allies seem to have completely forgotten this.

It makes sense for big monopolies like Bell, or the CCP. The investment can be justified if the ones investing are confident they will be able to capture the value and not some competitor.

Bell Labs also served to maintain positive perceptions of the monopoly. Unix was famously developed despite the knowledge that AT&T would not be able to offer it as an independent product.

No business sense for scientific research? You know you're completely destroying the argument for capitalism, right?

I don't see how it follows. Anyway it's debatable if the current system with antitrust laws is true capitalism. One of those poorly-defined words that people argue over.

This isn't basic research, it's applied research. Applied research lives or dies on the plausibility of the ultimate applicability.

This isn't basic research. The US has had this tech for half a century. There's just no reason to do it. Uranium is plentiful and cheap and arguably safer.

The fuel cost of a NPP has almost no impact on the NPP's operational expenses and a LFTR (like all liquid fuel designs) is a far safer design. Nobody in the energy industry has talked about the fuel cost in nuclear in 50 years. It isn't even a consideration when comparing designs. Waste volume, safety, politics, and construction labor costs are the factors which are considered (also temp of the heat maybe).

Reducing the energy sector to pure business would probably work in the 1990s, but not now, when countries are afraid of strategic dependence on potentially hostile suppliers.

Uranium isn't as ubiquitous as, say, natural gas, and stockpiling it comes with a big heap of physical problems. I can definitely see countries spending on more expensive technology if it comes with more energy security.

> in current market conditions.

That is, as long as we don't build more nuclear power plants.

If you want to increase nuclear power adoption, then you're not going to stay in “current market conditions” for long.

Reply on news.ycombinator.com