China Reaches Energy Milestone by "breeding" Uranium From Thorium
Key topics
Regulars are buzzing about China's groundbreaking achievement in "breeding" uranium from thorium, a nuclear breakthrough that could revolutionize energy production. As commenters riff on the implications, some highlight the potential for thorium reactors to offer a safer, more abundant alternative to traditional uranium-based reactors. Others dive into the technical details, discussing the complexities of the process and China's progress in mastering it. The thread crackles with excitement as experts and enthusiasts alike explore the possibilities and challenges of this energy milestone.
Snapshot generated from the HN discussion
Discussion Activity
Very active discussionFirst comment
1h
Peak period
81
0-6h
Avg / period
17.8
Based on 160 loaded comments
Key moments
- 01Story posted
Nov 22, 2025 at 12:49 PM EST
about 2 months ago
Step 01 - 02First comment
Nov 22, 2025 at 2:15 PM EST
1h after posting
Step 02 - 03Peak activity
81 comments in 0-6h
Hottest window of the conversation
Step 03 - 04Latest activity
Nov 24, 2025 at 5:07 PM EST
about 2 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
https://www.stdaily.com/web/English/2025-11/17/content_43298...
For comparison: A commercial nuclear power plant is 1 gigawatt, a 10x difference. I assume this would be the next step.
That to say, a typical commercial reactor might be 30x the power of a 100 MW research device.
But regulation, while it has its purposes, stifles many things. At the same time time it’s not even doing what they were meant for.
There are a number of countries being run far better than the US or the EU
It will be funny if China is what convinces the US to be more open to free industry. Opposite day vs the 1970s
If it's just your company or some trifling consortium trying to develop nuclear energy advances in a "free industry" environment, the guy who is just slapping up windmills, [T Boone Pickens RIP], is just gonna mop the floor with you. There's just no way to compete on moonshots like that.
TerraPower is not secret.
If you have easy access to uranium, you just use it directly instead.
The fuel costs of a NPP are a tiny rounding error. If you want electricity and want to build it today, Uranium not Thorium. You are using arguments from 50 years ago when many incorrect assumptions about cost structure and fuel availability were used to make decisions.
The pros you mention are theoretical - because the cons came out in force when actually tried, and they’ve been tried many times by many different countries.
0 - https://en.wikipedia.org/wiki/Uranium-233
The French interest in breeder reactors and nuclear reprocessing also originates from a similar concern about lack of domestic access to raw uranium. Though Super-phoenix [0] was a more traditional uranium -> plutonium approach and not thorium. They gave up because just using uranium is way, way cheaper than synthesizing your own fissile materials.
[0] https://www.world-nuclear-news.org/articles/indias-prototype...
[1] https://en.wikipedia.org/wiki/Superph%C3%A9nix
It had uranium-233 from breeding from thorium in other reactors.
The main problem with these things is they seem very unprofitable. The US reactor ran from 1964 to 1969 and produced a small amount of power but is still running about $10m a year in decommissioning costs. You thing you can run these things a while and think it's over but:
>Sampling in 1994 revealed concentrations of uranium that created a potential for a nuclear criticality accident, as well as a potentially dangerous build-up of fluorine gas: the environment above the solidified salt was approximately one atmosphere of fluorine. The ensuing decontamination and decommissioning project was called "the most technically challenging"...
I guess soon the west has to copy chinas tech.
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.
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.
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...
(Edit: cycomanic explained it much better and more patiently than me)
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
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
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...
https://mastodon.energy/@Sustainable2050/113695635521714572
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.
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...
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.
Most thorium: India, Brazil, Australia, US, Turkey
Most uranium: Australia, Kazakhstan, Canada, Russia, Namibia
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.
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.
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.
This advantage is conserved by all non-water moderated reactor designs.
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.
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.
Is anyone doing that? Everything I've seen is Rankine.
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
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.
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.
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.
Plenty has been learned by the US/West from copying their approach to agriculture, robotics in factories, mining, drones, etc. Have you seen their electromagnetic catapult technology?? That stuff seems like its from the space-age! There's even plenty of tech that we can't really explain like the all-moving wingtips on the new J-50s. (and yes, I'm avoiding talking about their supersonic cruise missiles)
Thorium MSRs don't make sense for the Americas, Europe or Australia. We have plenty of uranium.
Nuclear is receiving solid research backing in both America and China. (India is playing too. Austrlia is having an identity crisis.) Our different geologies mean there will probably be one solution for China, India and North Africa, on one hand, and the rest of the world, on the other hand.
That was tongue in cheek. It's being indecisive. I guess that's conserved across the Anglosphere.
Who said this?
> considering a liquid fueled reactor makes heat in the 900C range and a AP1400 makes heat in the 300C range, they aren't really substitutes for each other
Nobody said this either.
There are more reactor designs in the world than LFTR, PWR and BWR, particularly if we're talking at the demonstration scale like this reactor.
Is the article about a production power plant?
Came online ~10 years ago. One could quibble about design and construction timelines; the reactor is still half-experimental, and the Russians are conducting that breeder program very slowly. But it's not a 1980s design frozen in time.
That covers the input side of th equation. Thorium can help transform the outputs of our existing reactors into waste with orders of magnitude better in terms of dangerous lifespan
I'm glad people are finding more research and hopefully this will unlock other tech but this has limited impact on the current trajectory of commercial nuclear and the designs currently in the labs.
Though the commentary in here does remind me how much hype has infused the nuclear space - good thing on the whole as long as an eventual AI shakeout doesn't knee cap all the good work being done.
Try it someday. You _will_ be surprised by some of the technologies there.
why spend millions and a decade doing R&D when you can just hack American companies and steal it all for free!
For nuclear the playbook goes - design of technology is in the west. China copycats the reactor and puts it through their deployment engine (see current nuclear deployment). Maybe that changes -- but this doesn't prove that.
China pushing the development is fantastic though for the world to give their head a shake and finally get back in the game.
MSRs are riding the Oklo hype train and have a long way to go.
This unlocks a lot of options for the fuel cycle, including the use of thorium.
This work builds on a previous molten salt reactor experiment at Oak Ridge, decades ago. There's a whole lore about MSRs.
Notable, but not unique. The unique bit is it burns thorium.
What people need to understand about the cycle efficiency is that when you mine uranium, the fissionable part of uranium (U-235) is only 1% of that uranium, the rest is nonfissionable U-238.
Thorium is about twice as abundant as Uranium (all isotopes). The MSR uses Thorium to create U-233, a fissionable but not naturally occurring Uranium isotope.
So the "unlimited energy aspect" is that about 200-300x more breedable Thorium exists than fissionable U-235.
A MSR nation could also try to breed U-238 into plutonium, which would provide another 100x more breeding stock, although LFTR never talked about U-238 breeding. IIRC the plutonium may be difficult to handle because of gamma rays, but I don't recall exactly.
While I don't have confidence that even LFTR/MSR reactors can get economical enough to challenge gas peakers, it may be possible to make truly price-competitive MSR electricity with the right modular design. I wish the Chinese the best of luck, because if they do it will spur the rest of the world to adopt this about-as-clean-and-safe-as-it-gets nuclear design.
China has thorium, and while less than others [1], it’s better than they do with uranium [2].
> it may be possible to make truly price-competitive MSR electricity with the right modular design
Yes. But probably not in the near term with thorium. This isn’t designed to be cheaper. It’s designed to be more available to China than being dependent on Russian deposits.
[1] https://www.nature.com/articles/492031a
[2] https://www-pub.iaea.org/MTCD/Publications/PDF/Pub1800.pdf
https://en.wikipedia.org/wiki/Geoneutrino
Economic recoverable reserves are another matter, but there's plenty there.
https://en.wikipedia.org/wiki/Thorium-based_nuclear_power
Source for the fuel cycle?
Thorium 232 -> 233 is neutron negative. But after that you get all kinds of nonsense.
Even the daughter uranium 233 only produces on average 2.48 neutrons per fission, so it’s very difficult even in a combined lifecycle process to have enough - thorium doesn’t produce uranium 233 immediately (takes almost 30 days), neutron capture with that low a ratio requires a LOT of thorium, which is going to mostly just suck up all neutrons and you won’t have any extra for addition uranium 233 fissions, etc.
It’s quite difficult (impossible?)to have actually work without a source of a large amount of additional neutrons.
Unless 100% of those neutrons is being absorbed by the thorium, this means you'll have neutron flux at the boundary. Which, for a liquid moderator, means all the pipes and tanks and pumps.
And considering that people made these things work 60 years ago without modern computers, the idea that its impossible or needs 40 years of research seems pretty far fetched. What is left of the nuclear industry wants to build current designs like the AP1400. That is a great idea, but there are things you can do with a LFTR that you can't do with an AP1400. The biggest of them is making synthetic fuel. The other advantages are the amount of waste produced and the fact that you can make a LFTR into a waste burner consuming the spent fuel rods from a AP1400. The downside is you actually have to fix nuclear regulations to do this and getting politicians to do that has proved impossible.
There are no technological barriers, this is entirely political.
That you’re even discussing graphite moderated (?!!) makes this pretty clear.
And why would this be? Is graphite expensive? No it isn't. Also, we created a working one of these designed in the 1960's without computers. You seriously think this is hard compared to other types of engineering we do today?
A LFTR can also do things that a PWR or BWR can't and has several major advantages. But since it uses pencil lead apparently we can't even try it.
That's just Westinghouse. There is a lot of research happening in small and medium-sized reactors.
> There are no technological barriers, this is entirely political
To thorium MSRs? The main barrier is economic.
The most pressing is that fissionable material is spread throughout the fluid, so fission and decay of fission products is occurring right up to the edge of the fluid. The walls and pipes containing the molten salt, and anything dipped into the salt, are exposed to unmoderated neutrons. One can shield using (say) graphite, but then damage to that (and soaking up of radioactive materials) become issues.
The Molten Salt Reactor Experiment at Oak Ridge was near the end of its radiation exposure lifetime when the program ended.
Contrast this to light water reactors. These are designed so that no lifetime component sees unmoderated neutrons. There's a thick barrier of water between the fissioning fuel and the reactor vessel wall and the support structures for the fuel bundles. The bundles themselves are exposed, but they are replaced for refueling and are not lifetime components.
The metric to look for is called "DPA" (displacements per atom), the number of neutron collisions that a material can tolerate before losing enough structural integrity to fall below the acceptable limits. The best modern reactor steels are at 150-180 DPA.
And a lot of potentially cool reactors like TWR (travelling wave reactor) end up being logistically impossible because lifetime-limited components will be exposed to multiple hundreds of DPAs.
Not sure whether it would be possible to do something similar to a liquid fueled reactor, including all the hot pipework. Maybe, but yet another cost. Notably some of the recent MSR projects propose replacing the entire reactor every now and then (Terrestrial or whatever they were called, not sure if they are still around).
You can make the vessel thicker to compensate, but then you can just make it thicker in the first place and skip annealing.
It helps the atoms displaced by neutron collisions to "snap back" into the correct places in the crystalline structure. But it can never restore the material completely, and over time the annealing breaks will have to be more and more frequent.
It also can't be used for everything. Some pipes will experience large thermal stresses if annealed, and some components can't be heated properly due to complex geometry.
As with everything in engineering, all problems can be solved with additional complexity. It's possible to design LFTR reactors to be more annealable, but it will likely make them impractically complex.
There are also other issues with LFTRs. A significant part of the energy production will happen _inside_ the pipework carrying the molten salt, as delayed fission happens and daughter products decay. This will cause inevitable problems with the reactor power control.
Modern light water reactors are engineering marvels. They are incredibly compact for the amount of power that they generate, and they are now designed with the anticipated 70-100 year operating lifetime. Getting LFTRs to the same level of maturity might be possible, but it'll require literally hundreds of billions (if not trillions) invested, just like with the classic nuclear.
A somewhat lower temperature "wet" annealing process has been applied to two test reactors. I don't think it's been applied to any full scale power reactors.
https://www.nrc.gov/docs/ML2428/ML24282A716.pdf
Not the whole reactor vessel though, just one critical circumferential weld seam.
Oh dear god, no. Graphite is a very good moderator, it is in no way a shield. Those two properties are (sort of) opposites of each other. Lead makes the cheapest and best shield. Also, those parts that are exposed to neutron flux stay radioactive for about 10 years. So it shortens their lifetime in the reactor but the waste isn't a big issue.
There is a rabbithole for almost all of these material choices, especially in nuclear. Not going down that rabbithole in a discussion targeted at folks who don't spend their lives working in nuclear doesn't make that person wrong. It makes them an effective communicator.
PS Lead is a very very common shielding material in nuclear.
Oh my, definitely no :-) Do not use lead for neutron shielding. You're thinking gamma radiation but then we're talking apples vs oranges then. You want atoms comparable in size to neutrons, so something with plenty of hydrogen. Think water or PET (plastic) when you don't want water to "leak" when transporting a source. For thermal neutrons maybe PET impregnated with boron. Now neutrons may generate gamma when captured by hydrogen, then you may want some lead for secondary effects like that but I am not sure how strong those are.
The oakridge experiment ended and not a lot of R&D has been done on salt reactors. It makes sense that China is still basically in research and testing phases for molten salts.
We also have the ability to electrify most transport except maybe long haul trucking and long haul aviation. Aviation (ALL aviation) accounts for less than 5% of global CO2 emissions, which means we could leave that alone and cut elsewhere until we have batteries and other infrastructure good enough for that.
Build all this out and it'll be cheaper and more scalable than what we currently have.
We in the USA choose to stick with ancient technology because we have a sunk cost and an existing political power structure built around it. Meanwhile China is eating our lunch. Make America Great Again! By... pretending it's 1945 and trying to LARP the previous century.
Classic innovators' dilemma at the national level.
Some of the highest temperature reactor concepts use solid fuel (see e.g. various VHTR gen4 concepts).
As an aside, some nuclear proponents claiming synthetic fuel production as some unique selling point of advanced nuclear sounds more like wishful thinking combined with admitting being unable to produce electricity at competitive price. With the 'electrotech revolution', most things will switch to being powered by electricity, leaving a relatively modest market for synthetic fuels (long range aviation and shipping, mainly, and some chemicals production), assuming regulation prevents usage of fossil fuels.
> And you need Thorium for a liquid fueled reactor.
No, why would you? You can use U235 in a non-breeding thermal reactor (Terrestrial being an example design), or you can run the U-Pu breeding cycle in a liquid fueled fast reactor (such designs use chloride salts as the fuel carrier rather than FLiBe).
> That's why this design is so popular.
So popular that despite being invented in the 1960'ies, it hasn't yet progressed beyond the prototype stage?
Yes, and also vast oil and gas reserves China doesn't have.
Also there is strong public fear and dislike of nuclear power.
In countries where there are no or little fossil fuels it is mainly this public opinion which has crippled the nuclear industry. Germany is a prime example.
Public opinion is obviously much less important in China.
That really isn’t true. The reason Shanghai didn’t expand their maglev to Hangzhou is because residents were worried about electrical magnetic radiation, which I don’t think is really a thing. Nuclear took a long time to get started in China because people thought the government to be inept and corrupt, an image that has only recently faded away in the last decade. Without free elections, public opinion is actually much more important if you want to avoid economically destructive riots.
But this all happens in back rooms, the legal system isn’t very relevant, so if you have an issue but it isn’t a very popular one, you don’t really have any recourse. For example, niche environmental issues, or ones that aren’t widely recognized yet as dangerous…
In the US public opinion doesn't really matter either. It's the oligarchs' opinions that matter
On the aviation note, sadly, aviation bats higher than its C02 accounting. Contrails add another 1-2% on top of contribution from it's C02 emissions. It's entirely avoidable and could be resolved at relatively low cost.
https://contrails.org/faq/#how-are-contrails-contributing-to...
You don't want to discount the cultural attachment people have to what their parents did and their childhood.
Marginal versus bulk. It can make sense, economically, to keep building coal plants even if solar is cheaper if you’re building solar as fast as you can and still need more power.
1.While China scaled up the EV production, the development of Hydrogen based technology is still going on. There are some progress but lost in the bigger noise of EV.
2.China became the largest automobile exporter, leading by EV. But most people thought that's because EV took over ICE. That's partially true because EV dominate the export. What the most people missing is a quite portion of export are ICE cars. Because the ICE engine from China achieved higher energy transformation efficiency than Japanese and German cars. Again the information was lost in the EV noise.
I don't see the U.S rushing to adopt either renewables or nuclear. We're just increasing our fossil fuel burning (natural gas).
This is wrong. Natural gas is falling from 42% of U.S. electricity generation in '23 and '24 to 40% in '25E and '26E [1]. Renewables, meanwhile, keep marching from 23% ('24) to 24% ('25E) and 26% ('26E). (Nuclear falls from 19% ('24) to 18% ('25E and '26E).
[1] https://www.eia.gov/outlooks/steo/
No that’s generation. It’s on page 49 of the report. Table 7d Part 1 “US Regional Electricity Generation” it’s measured in billions of kilowatt hours.
https://www.eia.gov/outlooks/steo/pdf/steo_full.pdf
And if anyone is interested I have some of my own graphs on top of the EIA data to make it easier to read - https://eia.languagelatte.com/
Irrelevant. The question is what we're investing in. "The U.S" is "rushing to adopt...renewables."
> FF plants run most of the time
"CCGT capacity factor rose from 40% in 2008 to 57% in 2022" [1]. "In the western United States," meanwhile "the capacity value of PV plants can be in the range of 50% to 80%" [2].
> That's how they skew the numbers to make renewables seem viable when they produce a shockingly low amount of actual power
This is a report from Trump's EIA.
[1] https://www.publicpower.org/periodical/article/average-utili...
[2] https://docs.nrel.gov/docs/fy13osti/57582.pdf
152 more comments available on Hacker News