Comment by reenorap

> Nuclear inherently need a lot more effort refining fuel as you can’t just dig a shovel full of ore and burn it. Even after refining you can’t just dump fuel in, you need fuel assemblies.

It's true that a pound of nuclear fuel costs more than a pound of coal. But it also has a million times more energy content, which is why fuel is only 15-20% of the operating costs compared to >60% for coal. And that's for legacy nuclear plants designed to use moderately high enrichment rates, not newer designs that can do without that.

> Nuclear must have a more complicated boiler setup with an extra coolant loop.

You're describing a heat exchanger and some pipes. If this is the thing that costs a billion dollars, you're making the argument that this is a regulatory cost problem.

> You need shielding and equipment to move spent fuel and a spent fuel cooling pond.

Shielding is concrete and lead and water. None of those are particularly expensive.

Equipment to move things is something you need at refueling intervals, i.e. more than a year apart. If this is both expensive and rarely used then why does each plant need its own instead of being something that comes on the truck with the new fuel and then goes back to be used at the next plant?

> Insurance isn’t cheap when mistakes can cost hundreds of billions.

This is the regulatory asymmetry again. When a hydroelectric dam messes up bad enough, the dam breaks and it can wipe out an entire city. When oil companies mess up, Deep Water Horizon and Exxon Valdez. When coal companies just operate in their ordinary manner as if this is fine, they leave behind a sea of environmental disaster sites that the government spends many billions of dollars in superfund money to clean up. That stuff costs as much in real life as nuclear disasters do in theory. And that's before we even consider climate change.

But then one of them is required to carry that amount of insurance when the others aren't. It should either be both or neither, right?

It really is. Nuclear is 100-1000x safer than coal. By insisting on such an aggressive safety target, we force prices up and actually incur much higher levels of mortality - just delivered in the boring old ways of pollution and climate-driven harms.

See https://ourworldindata.org/safest-sources-of-energy for detailed stats.

I think we should target “risk parity with Gas” until climate change is under control.

This statistic is very relevant here, and surprising to many! Deaths per kWh produced for all energy sources.

Solar and nuclear both really stand out immensely as the safer alternatives.

People tend to think of nuclear as dangerous, but that's just propaganda. There has been a lot of anti-nuclear propaganda over the years. But the numbers speak truth:

https://ourworldindata.org/grapher/death-rates-from-energy-p...

https://ourworldindata.org/grapher/death-rates-from-energy-p...

> "It’s really not, nuclear inherently requires extreme costs to operate."

Not just to operate, but to clean up and decommission at their end of life. In the UK, for example, early reactors were built cheaply without much consideration/provision for eventual decommissioning. This has left an enormous burden on future taxpayers, estimated to exceed £260 billion, much of it related to the handling and cleanup of vast quantities of nuclear waste [1].

Thankfully new reactors are being financed with eventual decommissioning costs "priced in", but this is another reason why they've become so expensive.

[1] https://www.theguardian.com/environment/2022/sep/23/uk-nucle...

>> Most of it are fake costs due to regulation.

>It’s really not, nuclear inherently requires extreme costs to operate. Compare costs vs coal which isn’t cost competitive these days

Maybe it can't be as cheap as coal, but at the very least it shouldn't be absurdly expensive compared to what South Korea and China can do.

https://www.economist.com/content-assets/images/20250906_WBC...

> It’s really not, nuclear inherently requires extreme costs to operate. Compare costs vs coal which isn’t cost competitive these days. Nuclear inherently need a lot more effort refining fuel as you can’t just dig a shovel full of ore and burn it. Even after refining you can’t just dump fuel in, you need fuel assemblies. Nuclear must have a more complicated boiler setup with an extra coolant loop. You need shielding and equipment to move spent fuel and a spent fuel cooling pond. Insurance isn’t cheap when mistakes can cost hundreds of billions. Etc etc.

Without the fear of dual use, we could just enrich the fuel to higher levels and refuel once per 30 years.

Regulation still plays a role in the final cost. Sure it has to be safe. But we need to draw the line. Nuclear is arguably way too safe currently (zero death for a long time). Some regulations could be relaxed to speed up the construction, and make the operations cheaper.

We should have a discussion and review all the regulations surrounding nuclear.

https://www.businesswire.com/news/home/20250908384960/en/Dee...

I think if you regulated coal on a linear no threshold risk model, you'd find the costs to be somewhat closer.

> Nuclear inherently need a lot more effort refining fuel as you can’t just dig a shovel full of ore and burn it.

You have to take scale into account. This is 20 years of spent fuel.

https://npr.brightspotcdn.com/dims4/default/cca0b8d/21474836...

That's it. 20 years. Just that, for a constant, quiet output of just about a gigawatt. And that's an old, decommissioned reactor.

You're right about nuclear fuel refinement, packaging, and so on being non-trivial, but the amount of it that you need is so miniscule that if you don't talk about volume you paint a misleading picture.

> small modular reactors are only making heat they don’t actually drive costs down meaningfully.

Mass production makes anything cheaper. Ask the French about their efficient reactor program.

  > nuclear inherently requires extreme costs to operate.

True, but you also get large amounts of electricity in return.

You're over simplifying and cherry-picking. Is it a big deal if it costs 10x more if it produces 20x more power? What about 10x the cost, 10x the power (so equal $/MWhr) but 0.1x the land? What about 10x cost, 10x power, 1x land, but 10x more power stability? As in fewer outages. How much will you pay for 99.999 than 99.99?

The problem with the vast majority of these energy conversions is that people act like all these costs are captured in the monetary metric. I'm sorry, the real world is complex and a spreadsheet only takes you so far. There's no one size fits all power source. The best one to use depends on many factors, including location. If you ignore everything and hyper focus on one metric you're not making an informed decision that's "good enough" you're arrogantly making an uninformed conjecture.

I'm surprised how often this needs to be said (even to pro nuclear folks), but nuclear physics is complicated. Can we just stop this bullshit of pretentiousness masquerading as arrogance?

> It’s really not, nuclear inherently requires extreme costs to operate. Compare costs vs coal which isn’t cost competitive these days. Nuclear inherently need a lot more effort refining fuel as you can’t just dig a shovel full of ore and burn it.

This is based on reactors with poor efficiencies that leave a lot of unburned Uranium in their waste. Fast reactors and thorium reactors burn 90% of fissile material, so mining costs are significantly lower for the same power output.

> Insurance isn’t cheap when mistakes can cost hundreds of billions.

Total death count from nuclear is lower than the death count of wind and solar. Falling off roofs happens a lot more frequently than nuclear accidents. This is a nothingburger, particularly given new reactor designs are meltdown proof.

Regarding fuel:

https://xkcd.com/1162/

> A major reason nuclear plants are super expensive is because we do it so rarely

Once you have your supply chain running, and PM/labour experience, things can run fairly quickly. In the 1980s and '90s Japan was starting a new nuclear plant every 1-2 years, and finishing them in 5:

* https://en.wikipedia.org/wiki/List_of_commercial_nuclear_rea...

France built 40 in a decade:

* https://worksinprogress.co/issue/liberte-egalite-radioactivi...

More recently, Vogtle Unit 3 was expensive AF, but Unit 4 cost 30% less (though still not cheap).

Exactly. What is needed is a SpaceX-like enterprise, where the engineering effort is concentrated in building economies of scale. To me it's clear that nuclear energy's pros largely outweigh the cons, and that it is a perfect complement to solar and wind power generation.

It isn't that rare in general - if the U.S. opens the secrets of nuclear submarines - we had had mini reactors for decades.

There are some companies that are trying to get SMRs up and running.

https://www.ans.org/news/2025-02-05/article-6744/new-swedish...

We’ll see how it goes.

We’ve been trying to build ”SMR”s since the 1950s and a bunch has been built throughout the decades.

https://spectrum.ieee.org/amp/the-forgotten-history-of-small...

The problem is: who pays for the hundreds of prototypes before the ”process” has worked?

> If we'd build smaller, modular reactors that are literally exactly the same every single time you would begin to get economies of scale

You can also build standardized, modular LARGE nuclear power reactors. The French and the Japanese did it and managed to builds lots of large reactors with relatively short build times

It can be cheaper to run a nuclear plant than a conventional power plant, due to lower fuel costs. But what kills nuclear is the capital costs of building the plant. It takes a while to reap the reward

It’s always funny to me to see folks with the “HN leans _________” comments every few days with the blank spot filled in with every single political position one can think of.

I don't think anyone wants to get rid of nuclear regs entirely. There is a popular perception (i dont know if actually true) that safety regs were built around first generation reactor designs which were designed in an inherently unsafe way, and for modern designs that are inherently safer, it makes sense to relax some regulations.

Advocating for deregulation in order to achieve innovation is the opposite of conservative.

It’s not a matter of being a for profit or not. It’s an also matter of technological development. Most of the early incidents in nuclear plants happened under the management of public or state controlled companies.

Nuclear safety to provide safety is important but not to stifle any innovation or deployment which is what it has been.

No one is saying there shouldn't be regulations on nuclear.

But our regulations on nuclear are utterly insane -- every time I get someone to read into the reasons nuclear here has been so much more expensive than safe nuclear in other countries with more reasonable regulations around it, they come away shellshocked. It takes a while to understand what's going on, because it's truly death by a thousand cuts, but the unifying principle is the NRC's ALARA ("As Low As Reasonably Achievable") principle (with honorable mention going to the NRC's Linear No-Threshold harm model, which despite the evidence assigns a linear cancer incidence to radiation dosing).

Getting radiation exposure "As Low As Reasonably Achievable" sounds like a nice idea. But there's no lower bound, so the costs scale infinitely, gutting the incentives to innovate and invest. If the prices of other forms of energy go up, regulators intentionally raise the costs of nuclear comparably by increasing what must be spent on reducing radiation exposure. New innovative plant design that increases margins? Guess what -- that's another opportunity to use the money to lower radiation exposure even further.

The lack of a lower bound results in absurd results, because we long ago decreased the exposure from plants to far below background radiation levels, and far below the levels at which we've been able to observe harm.

We need to replace the LNT model with a sigmoid model that aligns with the science on radiation harms, and we need to remove the infinitely-scaling ALARA standard. Doing these will not increase risks, but will decrease costs a large amount in the short run and even more in the longer-term.

I completely agree with you and I'm pro nuclear. But those regulations have to be streamlined and the regulator needs to have enough manpower so licenses aren't stuck in limbo for years.

It's also unacceptable that the regulations can change during builds and then you have to make large parts completely new before you get the license to load fuel into the reactor.

> At peak hours electricity is $0.13/kwh but the delivery charge is $0.50/kwh.

Unfortunately, transmission has a natural monopoly risk, unless the government owns without profit requirements. The price peak is when it is just cheaper to make second set of lines next to old one and you can still pay the investment with fewer customers and lower price.

The goal of making nuclear cheaper isn’t to lower consumer costs. It’s to displace CO2 emitting baseload sources like coal and gas.

At some point the electricity will be near-free, and we'll just pay transmission fees

I'm a bit miffed I can't find the article now, but I recall hearing it was more the reactor design approval process than the operational process regulations that interfered with and drove up costs. Every tiny detail of a site has to be taken into account, forcing modifications to existing designs such that every build ends up being bespoke anyway. On top of that, many of the rules around the design approval process are geared towards older generation reactors and newer generation reactors end up being cost ineffective because they need to account for things that don't apply to them.

If anyone remembers that article, I'd love to cite it here. If not, feel free to ignore what is otherwise unfounded speculation I guess.

> I think industry overreaction to the regs is possibly as large or larger of a problem than the regs themselves.

I see this over and over again in regulated industries like banking and healthcare. No one wants to risk tripping up the regulations so company lawyers write up crazy and often conflicting “requirements” to satisfy legislation. The limitations placed by company council are often far more restrictive than regulations actually require. You have lawyers dictating engineering or software design requirements based off of a shoddy understanding of other lawyers attempts to regulate said industries they also don’t really understand.

And this isn’t to say that engineers are somehow better at this than lawyers. Engineers make just as many of these sorts of mistakes when developing things via a game of telephone. As someone who has played the architect role at many companies, it’s not enough to set a standard. You have to evangelize the standard and demonstrate why it works to get buy in from the various teams. You have to work with those teams to help them through the hurdles. Especially if you’re dealing with new paradigms. I don’t know to what degree this happens for other industry standards. But it seems like mostly folks are left to figure it out themselves and risk getting fined or worse if they misinterpreted something along the way.

I’d like to believe there is a way to balance lenience for companies that are genuinely trying to adhere to regulations but miss the mark at places and severely cracking down on companies that routinely operate in grey areas as a matter of course. But humans suck. And lenience given is just more grey areas for the fuck heads to play in. We cannot have nice things.

> I have yet to find any concrete defense of the idea that costs are coming from regulation, rather than the costs of construction in advanced economies.

One of the main drivers of excessive costs of construction in advanced economies are from excessive regulations, so it's really one in the same. Nuclear is obviously more regulated than other industries, and it routinely faces more frequent, longer delays and higher cost overruns than projects of comparable scale and complexity. This study [1] goes into a lot more detail.

Digging more into the details, it's all linked. The lack of regulatory clarity means that designs have to be changed more after construction starts, requirements for redundancy increase complexity, changing regulations prevents standardization, etc. Prescriptive regulations which were created decades ago limit the cost savings possible with newer technologies, like improved reinforced concrete. This study [1] goes into a lot more detail.

> Our retrospective and prospective analyses together provide insights on the past shortcomings of engineering cost models and possible solutions for the future. Nuclear reactor costs exceeded estimates in engineering models because cost variables related to labor productivity and safety regulations were underestimated. These discrepancies between estimated and realized costs increased with time, with changing regulations and variable construction site-specific characteristics.

[1] https://www.sciencedirect.com/science/article/pii/S254243512...

Its multifold.

1. Regulations are a big asterisk to any project. If you don't think you will get licensed or your project will get axed halfway through or there is a risk (Which has been very high in the past). Investors who would put money up for the project won't do it OR they require a significantly higher cost of capital. 2. There is very little muscle memory in the fabrication of reactors and reactor components in north America because we de facto shut down the industry from 80s until 20s. Therefore the first projects will cost more money as we recover our abilities to fab. 3. The licensing and regulatory costs are also incredibly high - and you cant make any adjustments if you kick off the project or you restart the process. This leads to massive cost over runs.

China and Korea are currently building reactors about 1/6 the costs of the US I believe.

> If regulations are the cost, name them and a solution.

That is a funny ask. Regulation doesnt have to be a single thing. It can very well be cost-overrun by a thousand paper cut. You can drown any project in endless paperwork, environmental and national security reviews. In fact unclear and contradictory requirements are much more conductive to drive costs up than a single Lets-make-nuclear-expensive-Act.

That being said if you need to pick a single thing (which is silly) then the “As Low As Reasonably Achievable” principle of radiation protection is a prime candidate. When you have a safety limit you can design a system to remain under it. When you are designing a sytem for the ALARA principle that in itself will blow your costs up.

All the safety and countermeasure costs here ultimately stem from regulation. If we allowed less safe power plants, they would likely be cheaper to build and operate.

However, I’m not sure I want private for profits actor deciding the level of safety of such projects.

It takes 15 years to build a nuclear power plant. It shouldn't take this long at all and it's strictly because of regulations. If we cut down the time it takes to build a plant the cost plummets.

Shouldn't the burden of proof belong to those that claim that regulation isn't the cost, when it is so extremely obvious to anybody who has ever had to build anything that it is?

Just look at building costs in California vs Texas. Both are nominally constituents of the same "advanced economy".

Yeah it seems like having State control is not a silver bullet

How much of that rate is because China is flooding the market with wind turbine blades and solar panels?

How much would it cost if China turns off that supply?

> Or acknowledge the true cost of $10 billion to build a reactor. Look at recent implementations. Finland was complaining that they had to deal with the mafia. The plant cost €11 billion, original proposal: €3 billion. Yikes.

This particular plant is a terrible example. It was the first of its kind, so it was bound to be more difficult than as part of a series. For example, there were issues with contractors that would not have happened if it had been the 5th reactor with the same specs. There were also issues with project management and changing regulations, which prompted some extensive tweaking of the reactor core almost as it was built. This is not representative of the difficulty of building a reactor that is par tof a fleet with identical designs.

Maybe roll back regulation to when France rolled out the Messmer plan?

They spent 1/4th of what we do today.

More people die every single year from the radiation parts of coal power (excluding accidents), than have died from radiation of nuclear power's entire decades long history, including accidents.

Yes, they should be made safe, but we need some perspective here.

In other words: nuclear is not viable unless the risks are offloaded to the public. Privatize profits, socialize risks.

PG&E is in no way a victim here. Their CEO is being paid $50M a year, and our rates got increased 6 times last year. Nevada the next state over, the prices are 20% of California's.

Unless by "taxes" you mean "delivery charges" this is simply untrue.

The generation is cheap. The delivery, the grid cost, is 3x-5x the cost of the generation.

It's all PG&E and the regulators's fault, for not containing costs more.

> if you agree to have all the nuclear waste dumped in your families backyard

What an unnecessary strawman. Nobody's gonna have nuclear waste in their backyards. It's all gonna get stored safety in glass vials under geologically inactive mountains.

>the urban areas of California are subsidizing the fire prone rural areas of the state

Meanwhile Rural California is where the electricity is actually generated[1]; they're "subsidizing" urban use.

>SVP vs PG&E

This has nothing to do with the ownership model and everything to do with not being obligated to serve rural areas. They get to serve only lower cost dense areas

[1] https://en.wikipedia.org/wiki/List_of_power_stations_in_Cali...

> The same corporate structures that hid data about smoking, PFAS and oxycodone are the ones you want running a nuclear plant?

Thanks for expressing my concerns over nuclear so clearly. It's not the technology I fear, its the people in charge.

Combined with democracy, it means that even if we trusted our governments today to police nuclear companies, they are replaced every few years. Nobody knows who will be in charge in 10 or 20 years time.

We should simply not build this large dangerous technology because rules and regulations will not keep us safe.