Congratulations, one of your very best columns. I appreciate your wish to cite the 1969 hit record, I’m The Urban Spaceman. But subsequent generations may be more familiar with a story written by Hans Christian Andersen from Denmark, concerning the small boy who alone was not taken in by the King’s much vaunted new clothes, declaring that the King was in fact naked. That seems thoroughly pertinent regarding SMR hypes.
So only a VERY small part of cost (and cost overruns) is driven by material cost (concrete, steel etc).
There are 2 main reasons for the astronomic cost of NPP:
1. An appalling lack of project control, especially over engineering. What idiot starts construction, before even the basic design is approved (Olkiluoto 3). Endless rework, redesign and other delays drive up finance cost (remember the 67%).
2. Strategic misrepresentation. Deliberately under-budgeting the project, because both the principal and the contractor want the project to be approved. Knowing they will not be on the hook for the inevitable cost overruns. The tax payer will be.
The only way to reduce the cost of nuclear, is to change the project delivery.
In such a way, that it does not take 15 years from FID till commissioning.
When you succeed in doing this, you target those 67%, and not the 11%.
But the conclusion remains the same.
SMR will not solve the problem, the propagandists claim it will solve.
It's just like with transformers. No matter whether it's a 3 KVA or a 75 KVA.
The working principle and the complexity are identical. Just the size differs.
A SMR (no matter what definition) is no less complex than a large PWR.
The material cost per output unit are higher. But that is the 11%.
Only when you would build assembly-ready modules off-site, would it be possible to exclude engineering creativity, and thus reduce cost.
But this all depends on unit cost reductions, for which you need sufficient total output.
Excellent and true article. I worked in the 60s at the uk atomic energy site which designed and developed nuclear reactors. At least one reactor had progressed to the final stage of supplying a fair amount of electricity to the National Grid for quite a few years with no problems. However it was decided to go ahead with the large size reactors and stop development of the various different types being researched even though some of them again had gone to quite a sophisticated level. What most people seem to have missed is that in the future we are going to need the best load power generation capability that is climate change invariable and also available all the time. Also able to produce huge amounts of gigawatts with a pretty reasonable long life. With these characteristics it's is likely to be expensive but that's not the point is it. there is not an alternative to nuclear power and we should be building as many as we can.
I find your take on this article surprising. Maybe I am missing something, but isn't the author saying the opposite? Many different designs have been tried, and most dropped. Building any nuclear reactor is very expensive and takes a long time. There simply isn't enough time and expertise (and funding) available to build many of them, regardless of relative size. It's a dead end. As for climate invariable - where are most nuclear power stations located? On the coast. Is the coast going to be impacted by climate change? Yes. How? It's going to move inland, sometimes by many miles. Also, more storms and storm surges massively increase the risk to existing NPS sites, requiring much more expensive mitigation (see Hinkley C). So, costs rise even further. It's a spiral of despair.
Firstly the uk built quickly two different successful reactors classified as small relative to the huge ones currently being buit. Neither expensive nor took a long time. Plus a successful fast breeder. And easy to replicate these. I know because I worked at the uk reactor develooment site. Neither needs to be near water. So your answer is totally incorrect. Research winfrith, sghwr and dragon
Sghwr 100mw provided power to the grid for 23 years 650mw variant also designed as scale up. We could be building lots now.
A successful fast breeder? Dounreay was closed in 1998 after multiple problems. You're missing the key arguments in my post about the relative diseconomies of scale of smaller nuclear power plant.
Rubbish it successfully supplied the grid for 15 years. It was shut down for political reasons. We should have several now greatly helping with our power problems.
I’d go in a slightly different direction. SMRs are interesting for grid operators that want to retire coal plants but want clean energy. SMRs are a potential answer by providing dispatchable power from extant centralized hubs while operating carbon free.
So the comparison shouldn’t be on the basis of a nuclear power plant vs. gas peakers or renewables, rather it should be the total cost of the plant and the marginal benefit of being able to site dispatchable power in extant brownfields without the need for infrastructure upgrades and being able to make use of extant power infrastructure.
As the experience in Iberia last month illustrated, having inertia in the net is quite beneficial as generators that just follow can make the grid very brittle. Especially if you can only import 3% of your grid needs like in Iberia.
Now how to make the economics work is a different matter. So far, few markets seem to have cracked the nut re: how to make use of solar/wind while pricing in the benefit of reliable power.
I’d also add that economies of scale are more likely to occur in an industry where hundreds of SMRs are made every year, not one large plant every decade.
As for the need for ultra specialized gear, regulations, etc. I have relatively little sympathy for that. Operators have time and time again shown themselves to be extremely lax re: common sense (ie siting nukes and gensets in a known tsunami area of impact) and most of these regulations were written by the industry itself.
You want cheap nuclear power? I suggest the French EdF has perhaps a pretty convincing answer that supplies over 70% of the French power needs: standardization. Makes construction and later maintenance much easier than the unicorns the rest of the west keeps putting up.
As I argue in my post, SMRs are a throwback to an age when smaller nuclear power were abandoned because they cost more than large ones. So inasmuch as we'll get any SMRs over the next few years (very few I'd guess) they'll turn out to be so expensive that the idea will eventually be abandoned. There's lots of ways of getting more inertia in systems that are a lot cheaper and much quicker to deploy.
Agreed. SMRs are not expected to be deployed at scale until the 2030s or beyond, which is too late to meet 2030 climate targets. Renewables, storage, and demand-side measures are already deployable and cost-effective today.
While SMRs are being heavily promoted, particularly by governments and nuclear industry players, their high costs, slow timelines, and unproven scalability make them a high-risk and potentially distracting solution compared to rapidly advancing clean energy technologies.
I don’t disagree re: costs. The overruns in more recent years have been astronomical. The only reason we have LIPA in its present form is a nuke and Cuomo.
…but that’s where SMRs may offer an out via standardization and avoiding the costly change orders that dogged many of the large plants built in more recent years. Ie creating a standardized format may allow the collection of learning benefits much like EdF did with its fleet.
As for inertia, no doubt. Whether is flywheels like the one they are building in Ireland, or fat interconnects to willing adjoining iso grids (ie looking at you, Germany) there are ways to do it.
I still think SMRs have a place, especially in context of helping secure grid stability on a local basis vs. relying so much on imports like my location does. How to value said reliability is another matter, however.
Arguably, they are learning. EPR2 is a simpler design, has a smaller scope of components, is easier to build. But it’s precisely the regularity of nukes to run grossly over budget that makes them so uneconomical in most markets.
You don’t have the same scale problem with gas peakers. They are well understood and generally have few installation issues. That’s the benefit of scale and learning - both at the manufacturing stage as well as on site preparation and integration.
The fever dream of SMR proponents is to wring similar efficiencies both in construction and siting if scales increase enough. Time will tell.
Wow. Somehow you conveniently skipped over GE-Hitachi's BWRX-300 currently under construction in Ontario for Ontario Power - a proper commercial model. It is a First of its Kind and 3 more are in line to be constructed - with the first one scheduled to come online by 2030 or so. Alberta and Saskatchewan are also in line to construct this.
So feel free to speak for yourself when you practically dismiss SMRs as a gimmick - but reality says otherwise.
There’s nothing under construction! Just press releases. If one does get completed then regret will set in, and nothing more than press releases will follow
Like I said, feel free to speak for yourself. Construction on that site has been going on for about a year now under a site preparation permit. The full construction will be starting this month under a full construction permit. All of this is publicly available information that can be found with minimal effort.
And you want to categorically declare that the whole thing will be a regret when you have literally no skin in the game, sure go ahead - just know that you are speaking for yourself, and not for the ground reality.
Also, see my commentary at the bottom of the post which I have just added - Darlngton is heading to be the most expensive nuclear power station per GW ever built
Considering you do not seek to be properly informed, let me inform you that actual foundation construction is already well underway at the site - under a construction permit no less. But again, it is not like facts really matter for those who have already made up their minds.
A quite brilliant dissection of the SMR myth. Here at STAND (Severnside Together Against Nuclear Development https://www.nuclearsevernside.co.uk) we are fighting the development of 4 or 6 "SMR"s on the River Severn. Our research and experience over the last 45 years has led us to exactly the same conclusions about SMRs as you, but you have encapsulated the arguments so cogently and succinctly. A recent poll showed that in spite of the fact that the official Green Party line is anti-nuclear power, 40% of Green party members are pro-nuclear. We will be attending a local meeting of the Green Party on Thursday, where pro Nuclear speakers have been invited, and your article gives us some excellence ammunition to counteract the disinformation that I know we will be hearing.
The Chinese have an experimental high temperature molten salt Thorium Reactor producing energy. They are now constructing a larger version. You do not mention TerraPower, the Bill Gates/ Hitachi collaboration molten salt reactor. This is a genuine SMR in that they have plans to mass produce them after the first one is built and working. With these modern designs the whole idea is that by the laws of physics it is impossible to melt down, so you do not need the concrete protection of a traditional reactor. They should therefore be much cheaper to build. The TerraPower reactor is to be 325 MW which is big by the standards of the 1950s. This is a good not a bad thing.
there have been experimental molten salt reactors before. There's nothing new about this and no reason to think it is at all 'cheap'. It will still have to follow the same safety rules of other nuclear reactors
Yes molten salt reactors were developed the 1980s, but there is a lot new. In the same way wind energy has been used for hundreds of years, but a modern wind turbine is much more efficient than a windmill In a tulip field. The whole purpose of the TerraPower design is to use passive safety (i.e. using gravity and the laws of physics) rather than a lot of concrete and humans turning dials. This design is far safer than the gen 3 reactors such as Vogtle and Hinkley Point 3. I agree with you just building a small version of a traditional reactor is pointless and likely more expensive than a large one, but this is not what TerraPower and the Chinese are doing.
Wind power was developed in 5000 BC. Solar electricity was generated around 1900. What matters is how it is developed now. The Chinese recently tested one of their next gen reactors with a total power shut down (that is no cooling as happened by accident at Fukushima). As the science tells us it should, the reactor just stopped. The billions Gates and the other investors have put into TerraPower is on the assumption they can build reactors cheaply.
There was an article in the Eastern Daily Press today talking about IMSR (Integral Molten Salt Reactor) at the Bacton site. Maybe in 2030s. It was a confusing article.
Great article David! Sounds like they are looking for a marriage proposal by putting lipstick on that old piglet.
Congratulations, one of your very best columns. I appreciate your wish to cite the 1969 hit record, I’m The Urban Spaceman. But subsequent generations may be more familiar with a story written by Hans Christian Andersen from Denmark, concerning the small boy who alone was not taken in by the King’s much vaunted new clothes, declaring that the King was in fact naked. That seems thoroughly pertinent regarding SMR hypes.
Excellent write-out.
Unfortunately this is about 1.234.567 times too long for the attention span of the nuclear propagandist. But very useful to refer to.
On important point though, on the cost.
78% of the total lifecycle cost of a NPP is fixed.
11% is overnight capital cost (mostly material cost).
67% of cost of financing.
See this figure (nuclear energy agency)
https://blog.policy.manchester.ac.uk/wp-content/uploads/2022/10/Figure-1-PNG-1-768x502.png
So only a VERY small part of cost (and cost overruns) is driven by material cost (concrete, steel etc).
There are 2 main reasons for the astronomic cost of NPP:
1. An appalling lack of project control, especially over engineering. What idiot starts construction, before even the basic design is approved (Olkiluoto 3). Endless rework, redesign and other delays drive up finance cost (remember the 67%).
2. Strategic misrepresentation. Deliberately under-budgeting the project, because both the principal and the contractor want the project to be approved. Knowing they will not be on the hook for the inevitable cost overruns. The tax payer will be.
The only way to reduce the cost of nuclear, is to change the project delivery.
In such a way, that it does not take 15 years from FID till commissioning.
When you succeed in doing this, you target those 67%, and not the 11%.
But the conclusion remains the same.
SMR will not solve the problem, the propagandists claim it will solve.
It's just like with transformers. No matter whether it's a 3 KVA or a 75 KVA.
The working principle and the complexity are identical. Just the size differs.
A SMR (no matter what definition) is no less complex than a large PWR.
The material cost per output unit are higher. But that is the 11%.
Only when you would build assembly-ready modules off-site, would it be possible to exclude engineering creativity, and thus reduce cost.
But this all depends on unit cost reductions, for which you need sufficient total output.
Wright's law, aka experience curve.
https://en.wikipedia.org/wiki/Experience_curve_effects
Paul Martin wrote an excellent article on LinkedIn, where he explained that the required number of cumulative output, will NEVER be attained.
https://www.linkedin.com/pulse/scaling-examples-pt-1-small-modular-nuclear-reactors-smnrs-martin
To conclude
1. Material cost is not the fundamental cause of the demise of nuclear
2. SMR will not save nuclear; because it is the umpteenth version of the emperor's clothes
the figures quoted at the bottom of my post comparing Darlington and Flamanville 3 are pure overnight costs, nothing else
Excellent and true article. I worked in the 60s at the uk atomic energy site which designed and developed nuclear reactors. At least one reactor had progressed to the final stage of supplying a fair amount of electricity to the National Grid for quite a few years with no problems. However it was decided to go ahead with the large size reactors and stop development of the various different types being researched even though some of them again had gone to quite a sophisticated level. What most people seem to have missed is that in the future we are going to need the best load power generation capability that is climate change invariable and also available all the time. Also able to produce huge amounts of gigawatts with a pretty reasonable long life. With these characteristics it's is likely to be expensive but that's not the point is it. there is not an alternative to nuclear power and we should be building as many as we can.
I find your take on this article surprising. Maybe I am missing something, but isn't the author saying the opposite? Many different designs have been tried, and most dropped. Building any nuclear reactor is very expensive and takes a long time. There simply isn't enough time and expertise (and funding) available to build many of them, regardless of relative size. It's a dead end. As for climate invariable - where are most nuclear power stations located? On the coast. Is the coast going to be impacted by climate change? Yes. How? It's going to move inland, sometimes by many miles. Also, more storms and storm surges massively increase the risk to existing NPS sites, requiring much more expensive mitigation (see Hinkley C). So, costs rise even further. It's a spiral of despair.
Firstly the uk built quickly two different successful reactors classified as small relative to the huge ones currently being buit. Neither expensive nor took a long time. Plus a successful fast breeder. And easy to replicate these. I know because I worked at the uk reactor develooment site. Neither needs to be near water. So your answer is totally incorrect. Research winfrith, sghwr and dragon
Sghwr 100mw provided power to the grid for 23 years 650mw variant also designed as scale up. We could be building lots now.
A successful fast breeder? Dounreay was closed in 1998 after multiple problems. You're missing the key arguments in my post about the relative diseconomies of scale of smaller nuclear power plant.
Rubbish it successfully supplied the grid for 15 years. It was shut down for political reasons. We should have several now greatly helping with our power problems.
I worked for the ukaea on reactor development.
Before more are built, solve the waste problem. We can't protect the human genome from increasing loads of radioactivity.
I’d go in a slightly different direction. SMRs are interesting for grid operators that want to retire coal plants but want clean energy. SMRs are a potential answer by providing dispatchable power from extant centralized hubs while operating carbon free.
So the comparison shouldn’t be on the basis of a nuclear power plant vs. gas peakers or renewables, rather it should be the total cost of the plant and the marginal benefit of being able to site dispatchable power in extant brownfields without the need for infrastructure upgrades and being able to make use of extant power infrastructure.
As the experience in Iberia last month illustrated, having inertia in the net is quite beneficial as generators that just follow can make the grid very brittle. Especially if you can only import 3% of your grid needs like in Iberia.
Now how to make the economics work is a different matter. So far, few markets seem to have cracked the nut re: how to make use of solar/wind while pricing in the benefit of reliable power.
I’d also add that economies of scale are more likely to occur in an industry where hundreds of SMRs are made every year, not one large plant every decade.
As for the need for ultra specialized gear, regulations, etc. I have relatively little sympathy for that. Operators have time and time again shown themselves to be extremely lax re: common sense (ie siting nukes and gensets in a known tsunami area of impact) and most of these regulations were written by the industry itself.
You want cheap nuclear power? I suggest the French EdF has perhaps a pretty convincing answer that supplies over 70% of the French power needs: standardization. Makes construction and later maintenance much easier than the unicorns the rest of the west keeps putting up.
As I argue in my post, SMRs are a throwback to an age when smaller nuclear power were abandoned because they cost more than large ones. So inasmuch as we'll get any SMRs over the next few years (very few I'd guess) they'll turn out to be so expensive that the idea will eventually be abandoned. There's lots of ways of getting more inertia in systems that are a lot cheaper and much quicker to deploy.
Agreed. SMRs are not expected to be deployed at scale until the 2030s or beyond, which is too late to meet 2030 climate targets. Renewables, storage, and demand-side measures are already deployable and cost-effective today.
While SMRs are being heavily promoted, particularly by governments and nuclear industry players, their high costs, slow timelines, and unproven scalability make them a high-risk and potentially distracting solution compared to rapidly advancing clean energy technologies.
I don’t disagree re: costs. The overruns in more recent years have been astronomical. The only reason we have LIPA in its present form is a nuke and Cuomo.
…but that’s where SMRs may offer an out via standardization and avoiding the costly change orders that dogged many of the large plants built in more recent years. Ie creating a standardized format may allow the collection of learning benefits much like EdF did with its fleet.
As for inertia, no doubt. Whether is flywheels like the one they are building in Ireland, or fat interconnects to willing adjoining iso grids (ie looking at you, Germany) there are ways to do it.
I still think SMRs have a place, especially in context of helping secure grid stability on a local basis vs. relying so much on imports like my location does. How to value said reliability is another matter, however.
Well the EPRs being constructed by EDF are supposed to be standardised as well. That’s not gong well.
Arguably, they are learning. EPR2 is a simpler design, has a smaller scope of components, is easier to build. But it’s precisely the regularity of nukes to run grossly over budget that makes them so uneconomical in most markets.
You don’t have the same scale problem with gas peakers. They are well understood and generally have few installation issues. That’s the benefit of scale and learning - both at the manufacturing stage as well as on site preparation and integration.
The fever dream of SMR proponents is to wring similar efficiencies both in construction and siting if scales increase enough. Time will tell.
Wow. Somehow you conveniently skipped over GE-Hitachi's BWRX-300 currently under construction in Ontario for Ontario Power - a proper commercial model. It is a First of its Kind and 3 more are in line to be constructed - with the first one scheduled to come online by 2030 or so. Alberta and Saskatchewan are also in line to construct this.
So feel free to speak for yourself when you practically dismiss SMRs as a gimmick - but reality says otherwise.
There’s nothing under construction! Just press releases. If one does get completed then regret will set in, and nothing more than press releases will follow
Like I said, feel free to speak for yourself. Construction on that site has been going on for about a year now under a site preparation permit. The full construction will be starting this month under a full construction permit. All of this is publicly available information that can be found with minimal effort.
And you want to categorically declare that the whole thing will be a regret when you have literally no skin in the game, sure go ahead - just know that you are speaking for yourself, and not for the ground reality.
Site preparation should not be mistaken for construction.
Also, see my commentary at the bottom of the post which I have just added - Darlngton is heading to be the most expensive nuclear power station per GW ever built
Considering you do not seek to be properly informed, let me inform you that actual foundation construction is already well underway at the site - under a construction permit no less. But again, it is not like facts really matter for those who have already made up their minds.
A quite brilliant dissection of the SMR myth. Here at STAND (Severnside Together Against Nuclear Development https://www.nuclearsevernside.co.uk) we are fighting the development of 4 or 6 "SMR"s on the River Severn. Our research and experience over the last 45 years has led us to exactly the same conclusions about SMRs as you, but you have encapsulated the arguments so cogently and succinctly. A recent poll showed that in spite of the fact that the official Green Party line is anti-nuclear power, 40% of Green party members are pro-nuclear. We will be attending a local meeting of the Green Party on Thursday, where pro Nuclear speakers have been invited, and your article gives us some excellence ammunition to counteract the disinformation that I know we will be hearing.
thanks!
The Chinese have an experimental high temperature molten salt Thorium Reactor producing energy. They are now constructing a larger version. You do not mention TerraPower, the Bill Gates/ Hitachi collaboration molten salt reactor. This is a genuine SMR in that they have plans to mass produce them after the first one is built and working. With these modern designs the whole idea is that by the laws of physics it is impossible to melt down, so you do not need the concrete protection of a traditional reactor. They should therefore be much cheaper to build. The TerraPower reactor is to be 325 MW which is big by the standards of the 1950s. This is a good not a bad thing.
there have been experimental molten salt reactors before. There's nothing new about this and no reason to think it is at all 'cheap'. It will still have to follow the same safety rules of other nuclear reactors
Yes molten salt reactors were developed the 1980s, but there is a lot new. In the same way wind energy has been used for hundreds of years, but a modern wind turbine is much more efficient than a windmill In a tulip field. The whole purpose of the TerraPower design is to use passive safety (i.e. using gravity and the laws of physics) rather than a lot of concrete and humans turning dials. This design is far safer than the gen 3 reactors such as Vogtle and Hinkley Point 3. I agree with you just building a small version of a traditional reactor is pointless and likely more expensive than a large one, but this is not what TerraPower and the Chinese are doing.
There have been molten salt reactors since the 1950s eg https://www.osti.gov/biblio/4237975
Wind power was developed in 5000 BC. Solar electricity was generated around 1900. What matters is how it is developed now. The Chinese recently tested one of their next gen reactors with a total power shut down (that is no cooling as happened by accident at Fukushima). As the science tells us it should, the reactor just stopped. The billions Gates and the other investors have put into TerraPower is on the assumption they can build reactors cheaply.
I cannot see how molten salt can be made "safe" or efficient in the long run. Pipes corrode - will they do so less in high temperature liquid salt?
Apparently this has been solved by Terra Power using a different compound for the piping utilizing Austenitic stainless steels.
Nuscale Power Corp has a market cap of $5bn. Are you saying it’s all hot air?
I didn't bother mentioning NuScale as it has failed in its planned development, see https://www.utilitydive.com/news/nuscale-uamps-project-small-modular-reactor-ramanasmr-/705717/. I'm not going to comment on how and why people decide to invest their money in ventures like NuScale - but I wouldn't!
Yes they have given up...
Last Energy. I looked it up.
What it does: 35110 - Production of electricity; 35300 - Steam and air conditioning supply
Who it’s owned by: Last Energy Inc, described as a start-up October 2024 by Reuters
Mini company, so audited accounts not needed. Has a deficit of £1.2m
Has been going since November 2021
DESNZ would give it £40m?
Probably not!
There was an article in the Eastern Daily Press today talking about IMSR (Integral Molten Salt Reactor) at the Bacton site. Maybe in 2030s. It was a confusing article.