China has 150 new reactors planned to go online in the next 15 years. They have a variety of different experimental fission plants such as molten salt and high temperature gas (helium) in addition to reliable "traditional" third gen reactors. It's a question of political will more than engineering at this point.
I am far more interested in white papers on how nuclear fission can be made significantly cheaper. Otherwise advancements in design are great but ultimately pointless.
Solar and wind continues to get cheaper and battery technologies continue to get better e.g. solid state in the medium term. And that makes once crazy ideas like shipping power from Australia to Singapore or from Africa to Europe economically feasible.
Gen-IV forum is a worldwide consortium with a long-standing effort in making nuclear fission viable for the second half of this century. It is less glamorous than fusion but much steadier, and at least three of the six Gen IV reactor designs under development are now approaching the technical readiness level for demonstrators. China and Russia are a bit ahead, to be fair, but the European Union is also doing well with their sodium and lead fast reactor concepts. This white paper just shows they care about more issues than blunt commercial viability. Also consider that they cannot rely upon Russian gas for much longer, and that the C02 zero emission target for 2050 is going to be tough, without nuclear (if they put it into the incoming tassonomy). United States are more pragmatic, and with a dual path: big national laboratories + commercial startups.
Lex Fridman had an interview with his dad on his podcast a while back. His dad is a plasma physicist and had some interesting thoughts on your point here:
In summary, he thinks hot fusion is obviously possible but the system is too complex to ever scale. He has an interesting idea about doping atoms with muons to increase the overall density of the material. I don't know that there's any substance to it, but I like his perspective that the ultimate solution will likely come out of nowhere by an individual or small team that's allowed to pursue wild ideas.
Note: haven't watched the video (I might). Using muons as catalyst is "old news" [0], it didn't seem to work well enough in experiments. It was used by Arthur C. Clarke as the energy source for the ships in 2061: Odyssey III
Oh wow! Thank you for this! I clearly didn't understand the mechanism that Alexander was discussing, definitely not that the change the atom was sufficient for it to effectively self-fuse and release the muon afterwards.
So all we need for a Mr Fusion on our Delorean is a portable muon gun. :P
IIRC creating muons is not that hard, but the fusion ratio was too small anyway. I still think combining it with some weak containment might work… I guess it doesn't and this is why it's not worked on anymore
Yeah I don't know anything at all but it does make me wonder if you could use the energy created by the muon-catalyzed fusion to push a lower-temp hot plasma into fusing over the edge. Almost like hierarchical catalysis of fusion.
Again per your observation, if it had any value we would do it so i'm missing one to many major things. I just enjoy thinking about it.
Interesting. If solar/wind and batteries continue their downward price trajectory we may never see commercial hot fusion on Earth, not because it's impossible but because it could never be economically competitive with the giant free gravitational confinement reactor at the center of the solar system. Fission and fossil fuels would eventually be eaten too, though I think you might still see some use of them in regions that are isolated and have very poor solar and wind options.
Fusion would be very important in the far future for the potential of interstellar probes or even interstellar migration. AFAIK it's the only power source that could enable flight at a meaningful (e.g. double digit) fraction of the speed of light.
Solar and wind are only cheaper when you don't count the adaptation the network needs to make to accommodate them (storage or gas plants to deal with intermittency, but also regional interconnections, frequency stabilization mechanism, the “smart grid” thing, etc). And people usually bring that “nuclear doesn't account for the decommissioning cost”, as if renewable did…
It's the same issue as trucks vs trains: trucks are cheaper because they don't pay for the roads at all.
Nuclear isn’t load following either. To get close to 100% you need to have vastly more power plants constructed which then mostly sit idle, but trying to do that is horrifically expensive.
It’s actually cheaper and safer to pair Nuclear with batteries than it is to have that many idle nuclear power plants which makes it directly comparable to wind and solar. Unfortunately, in a head to head competition to fill batteries, Nuclear simply loses. It would even loses when directly compared to solar + batteries in most areas for a grid with steady 24/7/365 power demand.
> To get close to 100% you need to have vastly more power plants constructed which then mostly sit idle
There's a difference between being forced to build enough capacity to provide 100% of your power, and have some idle when usage is at the cycle's low point, and not being able to load-follow i.e. not being able to adjust within that cycle...
> to have that many idle nuclear power plants [which] makes it directly comparable to wind and solar.
...And that's where that difference is important: a renewables production system has to install way more capacity than what's used on average, because they can't control the load factor, not just because they have to adapt to consumption cycles.
You can build a more expensive a nuclear reactor to quickly adjust power output across a wide range just look at nuclear subs. You can’t however do that and meet close to 100% of annual demand by keeping average power output below 50% and sell electricity at under 25c/kWh without truly monumental subsides.
That’s why nuclear isn’t load following in practice. Sure you can build reactors that occasionally dip few a few hours each night without breaking the bank, but something else needs to be covering your daily peak demand let alone extreme demand.
PS: The more expensive bit relates to both design changes to deal with reactor poisoning (https://en.wikipedia.org/wiki/Neutron_poison) etc and thus construction cost increased alongside increased maintenance costs due to increased thermal cycling etc.
France produces 70% of its total electricity from nuclear, and has one of the cheapest electricity in Europe, AND uses nuclear to adjust production to demand (no we don't do that with fossil, because modern fossil plants have a yield that is widely affected by the stability of the production, with a big warm-up time: just look at french production in winter, gas is mostly constant during the day and nuclear is taking half of the variation with hydro doing the rest).
Varying the power output of a nuclear plant (regular 2G plant) between 40 and 100% is straightforward and needs no specific design changes, and you can even go as low as 10% if need be. Xenon poisoning is a real thing, but it is relatively easy to deal with (simply by tweaking the Boron concentration the position of the control rods) even though it requires a bit of practice to master the skill. It is done routinely every day in all French nuclear plants!
You are simply living in a parallel universe which seams to share very few similarities with the real world…
France has low retail electricity rates because the government massively subsidized nuclear power. That doesn’t mean cheap electricity, that means hiding the costs for electricity.
Also, it doesn’t use 70% nuclear power, it currently exports ~20% of the nuclear power it generates and imports almost exclusively non nuclear energy. Net result it’s customers actually use about 55-60% nuclear power meanwhile average nationwide nuclear capacity factors actually fell below 60% and only recently reached 70% when they started to phase it out.
So you're not responding on the Xenon poisoning and power variation, I guess you stand corrected.
> France has low retail electricity rates because the government massively subsidized nuclear power. That doesn’t mean cheap electricity, that means hiding the costs for electricity.
No. The impact of nuclear electricity on public finances is a clear net positive over the lifetime of the park. And for the past 10 years, nuclear has actually been anti-subsidized by the political will to boost RE and competition in the french electricity sector (see ARENH if you're interested).
> Also, it doesn’t use 70% nuclear power, it currently exports ~20% of the nuclear power it generates and imports almost exclusively non nuclear energy. Net result it’s customers actually use about 55-60% nuclear power
This makes no sense. You can't just discount export from nuclear and count import as non-nuclear…
First of all, most of our biggest neighbours also have nuclear electricity (UK, Belgium, Germany) so we are definitely also importing nuclear (well, when it makes sense to actually export nuclear, see the next point).
Second, most of the time when we export we're actually producing electricity from other means than nuclear (gas, wind, hydro) why would you count only nuclear as being exported? If you want to draw an accurate line, you'd need to count the electricity at the marginal price as the one being exported and most of the time it's not nuclear.
Third, what about the past few days? We've artificially reduced nuclear production to import wind, because the weather leads to a massive overproduction of wind power in Europe. It's not that nuclear couldn't have produced this energy right, just that we had to turn it down to protect the grid from a collapse. Yet in your calculations you are somehow counting it as a liability for nuclear.
I agree with the statement that “between 40 and 100% is straightforward and needs no specific design changes” it’s not wrong. But to be clear a minimum of 40% is nowhere near enough if someone wanted to go close to 100% nuclear. He then mentioned one of several such changes needed in France which does go outside that range and again I agree with what was said. He even mentioned that France regularly goes below 40% output, but below 10% has much more serious issues. Again nothing I disagree with.
I am not sure how you think it’s a refutation of my point?
> You can’t just discount export from nuclear and count import as non-nuclear
I didn’t. They do export some non nuclear generation, but the majority is nuclear which you can verify by looking at the generation during times of export. ~20% nuclear is just what the actual number is. Similarly, they import almost zero nuclear power simply based on when generation occurs and when new power is brought online by the countries involved.
> No
Feel free to dig into hard numbers such construction costs, liability for accidents, and funds set aside to pay for decommissioning and where they came from etc. You obviously feel strongly about this so I can only assume you have exact numbers and a nice economic breakdown.
It takes some doing including digging into Areva‘s finances before and after it’s restructuring (bailout) etc.
> It takes some doing including digging into Areva‘s finances before and after it’s restructuring (bailout) etc.
Why the hell would you dig in Areva's finance on that topic? Look at EDF's finance instead for the past 40 years instead, because that's where the money is. Area's situation is a project management and strategic failure, but they aren't relevant to the discussion because they aren't electricity producers.
> I am not sure how you think it’s a refutation of my point?
Well, you argued that:
> They do export some non nuclear generation, but the majority is nuclear which you can verify by looking at the generation during times of export.
RTE's[1] disagrees with you … I don't know where you get your ideas from but they simply don't align with facts.
> That’s why nuclear isn’t load following in practice.
and
> The more expensive bit relates to both design changes to deal with reactor poisoning
Which I refuted: nuclear is indeed load following, and you don't need to make design changes to make it so.
This is a complex point so I want to address it separately:
> Yet in your calculations you are somehow counting it as a liability for nuclear.
It’s a liability to people operating nuclear power plants as they generally get paid by the market for generating power. When a grid adds solar or wind the wholesale price of electricity drops while those power plants costs stay the same. If it hypothetically costs you 10c/kWh of capacity averaged over the lifetime of the power plant your paying that money independently of how much you’re operating it. Workers showed up, equipment aged, and interest on your loan was due etc. Now actually generating electricity adds some small extra cost for the fuel etc so shutting down may actually save you money when the wholesale price falls low enough, but your losing money either way. The specifics do get complicated as a grid operator may have signed various contracts, but as the marginal costs for wind and solar are much lower than nuclear oversupply fucks nuclear in a pure market.
Of course in France the government subsidizes the industry, so it’s not at the mercy of pure market forces. However, someone is still stuck with the bill even if it’s future taxpayers as seen with recent bailouts. The obvious question then shows up of what exactly the mix of Nuclear vs Wind and Solar should on a carbon free grid. As a pure economic question the cheapest option is to have some large scale energy storage. Unfortunately, once that’s on the table the capacity for steady state operation suddenly becomes vastly less useful. Which is the long way around to what I have been saying where nuclear can’t follow the grid and be economically viable close to 100%.
> It’s a liability to people operating nuclear power plants as they generally get paid by the market for generating power. When a grid adds solar or wind the wholesale price of electricity drops while those power plants costs stay the same. If it hypothetically costs you 10c/kWh of capacity averaged over the lifetime of the power plant your paying that money independently of how much you’re operating it. Workers showed up, equipment aged, and interest on your loan was due etc. Now actually generating electricity adds some small extra cost for the fuel etc so shutting down may actually save you money when the wholesale price falls low enough, but your losing money either way. The specifics do get complicated as a grid operator may have signed various contracts, but as the marginal costs for wind and solar are much lower than nuclear oversupply fucks nuclear in a pure market.
This is true, but here we're going back to square one [1]: wind and solar are only cheap because we don't count the externalities. Each time you build a wind farm, you hurt the viability of other electricity sources but at the same time, so you raise the price of all electricity on the market because you need these other sources to deal with intermittency anyway. But it is not counted as part of the cost of wind, even though it is directly linked to the amount of wind poer you have in your mix.
> Of course in France the government subsidizes the industry, so it’s not at the mercy of pure market forces.
No, no no and again no. You really need to stop believing in this bullshit and face the facts: France doesn't subsidize nuclear at all, it's completely the opposite: it has been a milking cow for decades. And for the past 10 years, because the EU wants to promote competition, EDF actually has to give away some of its nuclear electricity to its competitors because having a monopoly on such a money-making energy was considered unfair. (you definitely didn't look for ARENH as I advised you to do so yesterday, right?)
> As a pure economic question the cheapest option is to have some large scale energy storage. Unfortunately, once that’s on the table the capacity for steady state operation suddenly becomes vastly less useful. Which is the long way around to what I have been saying where nuclear can’t follow the grid and be economically viable close to 100%.
You don't really know what you're talking about. First of all, I don't think why you're obsessed by the idea of getting 100% nuclear, France shows that 70% is a good mix and I think everybody agrees that going higher would make little sense. But you can't run on 100% solar or wind either (you'd need at least to 2 weeks of storage if you wanted to to do, which is far out of reach) and in fact there isn't a single demonstrated example of a working grid with even “just” 70% solar/wind, and just going above 50% is going to be quadratically more expensive for each percent (the load factor decreases, and the need for storage quickly ramps up).
This rolls back to here: “Each time you build a wind farm you hurt the viability of other electricity sources.”
If you say nuclear is a fine 70% solution, what’s the other 30%? If it’s wind then 70% nuclear is no longer viable because that was already the limit and it just became less viable. Thus the core problem, wind and solar cost less per kWh, but as you add them they push out nuclear. Essentially what mix actually works? If say 70% wind and solar 25% nuclear 5% hydro is economically cheapest nuclear needs a very special power purchase agreement designed to support it.
Which gets to the hart of the matter, much of the problem with the current grid is power purchase agreements. No utility wants to sign a power purchase agreement for the lifetime of a power plant which pays nuclear even when it’s off. Even 35 years after construction means the power plant needs to compete on the open market in ~40 years. Which means subsides.
PS: “France doesn’t subsidize” among other things it’s agreed to limit liability in the event of a major disaster. That’s incredibly valuable, suppose you wanted insurance for a 1 in X per year per 1GW reactor risk of a 500 billion dollar failure what does that cost? Well X becomes really important, we have had what 2 major disasters in what 441 current reactors * ~70 years of operation = 1 in 16,000ish. Now I think a reasonable argument can be made that French reactors are safer than that, but napkin math says even if it’s 1 in 50,000 per year you’re still looking at ~10 million per year per reactor * 56 reactors or 1/2 billion per year for insurance. Even if their 100x as safe as historic examples it’s still significant money.
The biggest recipient for energy subsidize in France is fossil fuel, and the lowest is nuclear. This is despite the fact that fossil fuels account for the least amount of energy, and nuclear the highest. Renewables sits in between. Fossil fuel subsidizes is on the rise, especially for oil and gas.
What France (and other countries in EU) should do is to cut those subsidizes for everything, and an additional carbon tax on top that directly targets the energy sector. That money can then go into a tax reduction/social support for individuals most effected by the increase energy prices, while at the same time letting fossil fuel industry die a horrible death.
Interesting can you list all the actual nuclear subsidies and the money set aside for plant decommissioning or are you just repeating something you read somewhere? Including those related to nuclear weapons production via state owned nuclear reactors.
Page 6, figure 2, Subsidies for different energy sources, as percent of GDP and in billion euros in 2019
The yellow slice is nuclear, green renewables and grey fossil fuels. The image text describe the additional colors that refers to "other" and "shared by all".
In term of actually money collected by tax agencies in EU countries, fossil fuels and renewables is the primary recipient of citizens money. Year in, year out.
We are talking about France not the EU. Look at your own numbers: “ Figure 2 - Subsidies for different energy sources, as percent of GDP and in billion euros in 2019”
France is a rather large anomaly even though their nuclear reactors have been built and paid for. This is exactly the time period they should be most profitable, yet still a significant subsidy. More, that’s in comparison to all energy consumption not simply the electric grid. And further, their subsidies for wind and solar are actually paying for new generation to just continued production.
I am trying very heard right now to not invoke the statement "learn to read", but fair enough, graphs are hard for some people.
In page 6, figure 2, bottom line, there are country codes. FR is the country code for France. From left is the 11th, which only represent the numbers from France. France and only France.
Amazing. Not only you don't even bother getting sources, but when someone finds a source that directly contradicts you, you end up distorting the figures you read to fit your twisted view of the world.
A misunderstanding isn't when somebody don't believe in the same things as you. The GP's source was clearly not going in your direction, yet you acted as if it was by just refusing to understand what it was about. The level of self-delusion is truly incredible.
Suppose someone said the US had a budget surplus and linked to an article showing the opposite. Should I not correct them?
When we are talking about electricity grid subsidies in France and they bring up EU wide gasoline subsidies that’s a rather clear misunderstanding. I could have gone further and talked about subsidies on a per kWh over the lifetime of a power plant rather than the relatively tivial amounts mentioned to keep already built nuclear reactors operating but that’s likely to go above their head.
As I have told you I don’t actually care about correcting the ignorant. At best I get some dark joy at making some people look foolish.
> Nuclear isn’t load following either. To get close to 100% you need to have vastly more power plants constructed which then mostly sit idle, but trying to do that is horrifically expensive.
France meets its peak electricity demands with fossil fuel generation, hydropower, and imports. It doesn't have enough nuclear capacity to service annual peak demands without fossil power because that would leave expensive reactors idle much of the year, as the grandparent post said. Throttling down nuclear is not technically difficult but the high cost would sink a commercial reactor that operates at only 50% of its technically achievable capacity factor.
That, however, is an economic decision unrelated to load following. Being able to load follow but deciding not to build the last plant that would be used 5% of the time is not the same as not being able to load follow, and being forced to have batteries or alternative production means.
> France meets its peak electricity demands with fossil fuel generation, hydropower, and imports.
Hydropower is used for peaks a lot, because France has a lot of reversible hydro, but fossil plants and imports aren't routinely used for peak use.
Take a second look at those imports, they only did that when nuclear was unable to meat demand which was every single weekday.
Exact numbers get tricky as they would export and import at the same time due to grid balancing still they needed to export ~1/5 of all nuclear power generated at a loss and import ~30+tWh just to meet demand. Even then utilization tells the story it was in the 60% range in 1988 while the US saw utilization in the 90% range, and again that’s with exporting below costs. Low utilization increased prices by about 30-50% depending on year and that’s with heavy dependence on fossil fuels and a relatively low percentage of nuclear consumption in the country. Nuclear simply doesn’t scale across an electric grid unless you want to more than double the price of electricity or provide subsidizes on that scale.
Just look at what happens to the EU grid if Germany and every every other country outside off France was 1/3 nuclear.
I have linked to several sources in this thread, but as am not going to bother when someone says “your wrong” and nothing else. It’s the same reason I don’t talk with flat earth people, trolling is amusing for some people show a little investment and I may educate you.
Did you notice people saying French nuclear power had zero subsidies? Notice how I didn’t mention the links from other people in this thread showing French nuclear subsidies.
Perhaps I am sadistic, then again I am not the one defaulting to personal attacks rather than cold hard facts.
Zero “net” subsidies, which is what matters. The French govt earn much more money from nuclear than they spend in nuclear, so saying that the govt subsidies nuclear is a lie.
Permanent goalpost moving, pulling imaginary figures from sources you never refer to (and pretending you did) isn't “cold hard fact”: it's bullshit.
Nuclear Power has the same cost problem as public transit infrastructure (and healthcare & education for that matter too). It's not mass produced, there's no competition, capital productivity factor is very low.
But! Uniquely from that group nuclear power can be very easily industrialized. Basically if there would be a group that orders 100-200 (and of course more) plants ... the unit costs would go down dramatically.
As long as each one is a unique little bespoke snowflake, each site needs special plans, every pipe and valve and weld and button needs loving care and precious human attention ... it's going to be expensive.
My impression is that dis-economies of scale in fission reactors (safety systems, low-volume manufacturing & construction, local grid management, etc.) are severe enough that neigh-all more-recent proposed designs are for far smaller reactors than the older generation (~1GW per reactor).
Small Modular Reactors (SMR) are currently the darlings of the nuclear world. They promise cost reduction through mass production in factories rather than on site, smaller size enabling simpler passive emergency cooling, hypothetically simpler licensing through something like type certificates used for, say, airliners rather than starting from scratch for each reactor, etc etc.
This all sounds promising, and it's certainly worth investigating further, however for actual reactors that have been built and operated, cost reduction through increasing the size has been one of the very few approaches that has been empirically demonstrated to work.
Not necessarily - one of the proposed use cases for SVBR-100, soviet/russian SMR design, back in 1995 was to provide small power plant distributed to cities or industrial areas, providing 100MWe power or combined heating and electricity. They would compete directly with still common metropolitan power plants, especially CHP in cities with municipal heating.
One of the main issues of the design was that we simply do not produce enough bismuth for the lead-bismuth coolant, and lead-only variant was iirc more problematic due to higher operating temperature. The lead-bismuth design was tested for mnay years as submarine reactor.
Russian research actually found polonium to be a net benefit!
Namely, the short half-life and the fact that it undergoes alpha decay into lead meant that a) it was very easy to detect leaks b) cleanup was also easier. Also, the rate of polonium production turned out to be pretty low.
The bigger issue is that Bismuth production would be utterly, completely drained by mass production of SVBR-style reactors.
Sure, even if all the previously mentioned SMR advantages would turn out to be true, it would still make sense to place multiple SMR's at the same site in order to take advantage of common grid connections, security, etc.
IIRC Nuscale is designing for up to 12 of their 60 MW reactors in the same plant.
I would say proliferation resistance is nearly completely irrelevant.
There are hundred of weaponizable reactors in not really good places of the world. In case of the next big war, nobody will ask IAEA for permission to build the nuke.
As some the saying goes, the genie is long out of the bottle.
This is short sighted. This isn't about getting the genie back into the bottle but about sharing nuclear technology with developing countries that we otherwise can't help because they'd need a source of enriched uranium and it'd provide them with an industrial source of weapons grade radioisotopes.
This is one of the few ways we can shortcut the developing world's reliance on fossil fuels while still supplying plentiful base load power that can grow with them.
"Developing countries" with an axe to grind on USA already have dozens of research, or power reactors. This is my point.
If somebody will bomb your cities with nukes, killing millions in the process, you will not need Mariano Grossi, he will be useless to USA. You will need capable military allies.
So far, the US does everything to turn away any potential ally, including by denying them access to powerful weapons, with thermonuclear weapons being the biggest one.
> "Developing countries" with an axe to grind on USA already have dozens of research, or power reactors reactors.
Right. This isn't about them. This is about all the other developing countries that don't. This is about making them allies by giving them nuclear technology, something that the developed world is terrified of doing because of nuclear proliferation. That's why proliferation resistance is relevant.
If any of them really wanted so, they would've build a small clandestine reactor somewhere long time ago.
Nuclear bombs are forties technology. Industrial output of forties USA is met by dozens of countries today.
People continuing to advocate for non-proliferation cannot believe the that the most basic point of theirs is false: from technical viewpoint, from logical, political, and military one.
Of course they can, but they don't. It's extremely expensive and the vast majority of countries can't afford the massive upfront cost without nasty consequences. Most of the 32 countries with nuclear reactors built them using technology from Russia, France, the US, UK, or China. China still builds nuclear reactors using imported Russian technology because its own industry can't keep up with the demand - the demand for exporting Chinese nuclear reactors to countries that US/EU won't sell nuclear technology to.
The article we're commenting on is part of the EU and US response - they want to sell their reactor technology to a bunch of countries to prevent China from getting that business and gaining influence, but they want to do it in a way that reduces proliferation because the constituents that vote for them care about nuclear proliferation. It's entry level geopolitics.
But then, they would have to get fissile material, which isn't exactly an open market. It's easy to justify buying the material if you have a commercial reactor to feed it to, not so much if you don't.
In fact, international inspection mechanisms worked pretty well so far.
> they would've build a small clandestine reactor somewhere long time ago.
And have some advanced nation genocide all the relevant brains on your country? I would try to lead them in some other enterprise, even better if it doesn't have to be clandestine.
The only reason one needs enriched uranium is because of political pressure from nuclear-armed countries to both force everyone to use reactors that require enrichment and take control of availability of enriched fuel.
NPT with annexes is already plenty enough to prevent hidden repurposing of waste into weapons, it just doesn't fit American immediate econopolitical wants.
It's not an all-or-nothing issue. Every single reactor that produces weapons-grade material increases the risk.
Because it's not just state actors you need to be worried about. A gun-type fission bomb is simple enough for a well-funded terrorist group to construct from scratch - if they can obtain enough enriched Uranium.
There are currently no commercial power reactors that use highly enriched uranium. A terrorist group that stole a trainload of fresh fuel for one of today's power reactors would be unable to make a bomb from it.
Fast spectrum reactors require a larger core inventory of fuel and higher concentrations of fissile material within the fuel. That makes fuel diversion more dangerous with fast reactors. Presumably that's why people interested in fast reactors are trying to improve their proliferation resistance, to bring it in line with what's considered acceptable today.
As a side note, purely technical measures cannot render reactor designs proliferation-proof against rogue governments that might repurpose reactors to make weapons. At best you get some warning from breach of anti-proliferation measures so other countries know when legal/diplomatic countermeasures become warranted.
> There are hundred of weaponizable reactors in not really good places of the world.
Usually, a reactor is considered a proliferation problem if it allows you to produce plutonium easily. Getting radioactive material isn't good enough to be a significant security risk, because dirty bombs are much more complicated than conventional explosives, for an impact that isn't significantly bigger.
So a reactor can be a risk if it lets you get plutonium easily (and, specifically, the Pu239 isotope). The other plutonium isotopes are not good weapon material, and so you want to be able to use a reactor that lets you produce material with as much pu239 as possible.
Depending on design, if you can get easy access to the nuclear fuel, you can filter it frequently to retrieve the pu239, before it has a chance to react with another neutron and become one of the bad Pu isotopes.
BWR and PWR reactors are bad for enrichment, because in order to access the fuel, you need to stop the reactor and depressurize the primary cooling circuit, which takes time. On the other hand, if you let the fuel spend a long time in the reactor, you end up with spent fuel that has lots of Pu240 and Pu241, which sucks.
Other types of reactor, like CANDU, RMBK or UNGG reactors let you access the fuel more easily, without shutting down the reactor. Therefore, it is easier and quieter to retrieve the fuel frequently and get the Pu239 before it has time to react again.
That is why some types of reactors are nicer to have if you want to build a bomb: they take less time and would raise less suspicion from other countries to collect the same amount of Pu239.
Currently, most commercial reactors in the world use BWR/PWR technology.
I think that sums up the gist of what it means to make a reactor less proliferation-friendly, and why it's important in order to make that technology more friendly to export.
