The term “Five-Year Plan” is a bitter joke, referring to the announcements by the Soviet Union about how they were going to reach some specified level of production of heavy industry, which would notoriously be followed by some combination of inefficient allocation of resources, manufacture of useless or defective products, and flat-out lying about what was being produced in the factories.
On the other hand, sometimes Five-Year Plans, or something like them, do work. Most notably there was WW2, when the U.S. really did build a bunch of factories and produced previously-unimaginably large quantities of airplanes etc. And we really did get to the moon, and it didn’t take much more than five years.
I was thinking about this a few months ago when I was at a conference and some political scientist stood up and told us how, because of competition with China over data centers or AI or something like that, we needed to build a bunch of nuclear power plants. And all the savvy politicians know this, he said. I guess AI isn’t just about making goofy videos or cheating on your homework anymore, or even just about writing more efficient code. You don’t need zillions of gigawatts just to run coding assistants. I think the idea was that this is about building the next-generation AI, whatever it will be.
I hadn’t heard about all this so I did some searching online and I found out that this is the pitch that’s been coming from various software executives. These guys have tons of money, tons of media exposure, and tons of mystique, so it’s no surprise that lots of politicians would fall under their sway. I guess all the politicians except the far right and the far left: at both extremes, their combination of strong ideology and safe seats could make them relatively immune to the siren song of tech zillionaires.
And, I dunno, maybe these guys are right, that it’s absolutely necessary to set national priorities in this way, climate change be damned. (Yeah, I get it that nukes don’t contribute much to global warming, but we’re using tons of power anyway, so if the nukes are going to just be used to power these new data centers, we’ll still need all the other energy sources to keep the heat on and the cars running and all the rest.) I don’t know.
It’s just funny how these guys are going all-in on the whole Five-Year Plan thing. I could imagine someone saying, Hey, if the Chinese want to bankrupt themselves by investing all their resources into gas-guzzling data centers, let ’em! But in this case the centrist consensus seems to be that Planning is the way to go. Or maybe it doesn’t count as Planning if the profits are private. It was just funny to hear a political scientist going on about how this needs to be done without even remarking on the historical parallels, which indeed go in both directions, as noted in the first two paragraphs of this post.
Wouldn’t prioritizing climate change also require something like a five year (more likely a hundred year) plan? It seems like this is something the right and the left might agree on: we need to have long term plans. But there is immense disagreement regarding what those plans should look like. Then there is a third group – the one that denies that plans are necessary or even useful. I admit to have some agreement with that group – most plans go awry, and the information requirements for good plans are often unreasonable. Decentralized markets are often superior to these Soviet-styled plans. For me, the distinguishing characteristic is that modern technologies require more than what decentralized markets can provide (note that many economists disagree with this). I don’t see how long-term energy use, environmental protection, species’ diversity, nuclear weapons, or AI use (to cite a few) can be adequately handled by laissez-faire markets. Multi-year (somewhat) centralized planning may make mistakes, but I don’t see how we can view them as unnecessary.
It’s not clear to me whether you (Andrew) are objecting to a “five year plan” for increased energy production, or a plan for nuclear power specifically. If the latter: This is one area in which the popular view is both severely wrong and environmentally harmful, and I’m very surprised to see you implying otherwise. Nuclear power is extremely safe — even including Chernobyl etc the deaths per Watt are about 4 orders of magnitude less than what we blithely tolerate for fossil fuels (https://www.statista.com/statistics/494425/death-rate-worldwide-by-energy-source/), carbon emissions are of course zero, the specific energy of the fuels are so enormous that the volume of waste generation is tiny, etc. One of the great tragedies of the problem of climate change is that we *already possessed* the technology to avoid much of, and we chose not to, simply because the word “nuclear” conjures images of bombs and fictitious three-eyed fish.
Sorry to be a bit off topic. I have an on topic comment, too, which I’ll try to find time for later!
Raghu:
Unfortunately I don’t think the five-year plan under discussion is to build nuclear power plants to aid the transition from carbon. Rather, it seems that the plan is to build nuclear power plants specifically to run new data centers, new bitcoin computers, etc., thus allowing tons of added energy use without relieving the use of existing resources.
Unless we’re saying: “Let’s prohibit energy to be sold to industry X” , which is unpalatable and hard to imagine, industry X buying energy will increase the demand for energy overall, which is inseparable from the question of where, overall, we get our energy from.
It’s unclear what the polscientist you are referring to actually said about five year plans. But if he suggested that we could build new NPPs in five years, he must be completely delusional. The last few nuclear power projects in the Western hemisphere took 15-20 years to complete at astronomical cost overruns (check Oikiluoto, Hinkley Point, Flamanville). Even in China, NPPs take longer to get built although not as long as in Europe (for reasons you can imagine), and the energy strategy of the PRC is far more, and with astonishing success, geared towards wind and solar than nuclear power.
It is absolutely true that no NPPs will be built without some sort fo longterm government planning. “Market forces” are unanimously saying “no thanks”. And that’s the main reason why nuclear power will be increasingly irrelevant.
Piglet:
Just to be clear, that political scientist did not use the term “five-year plan.” But it sounded like a five-year plan to me, or something like it, in that he was calling for massive government investment. It also had a Cold War flavor to it, as the motivation was a vague feeling we’d be left behind if we didn’t devote massive resources and, indirectly, massive fossil fuel consumption, to this project. Which, instead of tractors or heavy industry or whatever, seems to be computer programs to make video deepfakes and help kids cheat on their exams. But, the message was, if we don’t turn our economic and energy policy inside out, the Chinese will beat us to the finish line.
I don’t know that there was much political science in this message. It just happened to be a political scientist who was saying it.
Yeah I understood that he was “calling for massive government investment”, which I don’t think is unreasonable per se. But I do wonder whether he was straightworward about the scale of the cost (hundreds of billions at least), and perhaps more importantly the time scale (30 years min), of such a project. Maybe he had no clue about it or he was delusional or dishonest, but just calling for huge investment without talking about the scale, and without talking about what we couldn’t have instead (namely, much cheaper and faster to build renewable energy) is pretty disqualifying. Many times the clamor for nuclear power isn’t actually about nuclear power, but about sabotaging the clean energy transition already under way.
Piglet:
I had the impression that he thought it would be a decades-long project but that if we didn’t get moving in the next five years, we’d be too far behind China to ever catch up.
The term, “Five-Year Plan”–with or without the dash–reminds me of my youth because it ignites the Cold War of the bad old days. Back then, as I somewhat hazily recall, the notion of a “Plan” smacked of communist totalitarianism, a denial of the “Free Market” religion of the U.S. Keeping the Government out of the mix was a matter of faith and dogma, albeit convenient violations happened.
It is fascinating to watch a slow-motion train wreck. Capitalism is good for humanity, amirite?
I read the Texas Tribune, which gives fascinating insight into America’s banana republic, Texas. The entire issue is pretty much encapsulated in this statement from the Tribune:
“‘The data center folks have found a sweet spot in the state that has limited regulations, limited enforcement, limited code, and they’re coming faster than we can keep up with,’ said Hill County Commissioner Jim Holcomb. “I think it’s imperative … that we tap the brakes and we get our arms around what we’re faced with and do the research, do the studies.” Holcomb, who voted for the pause [in data center construction], said the move was in ‘no way, shape or form a push to impair anyone’s right to do with their own property what they want to do with it.'”
It’s not just the property. Texas has an unregulated energy industry run by private industry. The data centers are simply going to outbid the communities for power and water, pushing up prices for everyone. The strong eat the weak, that is the way capitalism is supposed to work.
To date, most of the effort by the data center builders in Texas has been old-fashioned political graft, which has always been how you get things done in Texas. It begins when a political entity announces that the issue is being “studied,” meaning that the time is ripe for big donations to the PACs associated with the politicians doing the studying. That gets the politicians on board, but how do you convince the citizens?
Now that even some Texas communities are fighting back, the data center builders have to figure something out, so they tell “just so” tales about new energy and water supplies. Here in Washington State where I live, Microsoft is building a data center near where Alcoa used to make aluminum cans (the plant closed a few decades ago to seek cheaper power overseas, with cheap Columbia River power having been the attraction in the first place). Microsoft says don’t worry, they will be opening a fusion plant next door to provide the power. Fusion!
That banana republic is the top in renewable energy of any state in the union, even though it represents only about 1/4 of the mix at the state level.
And *not* building nuclear power plants for the AI data centers will result in what, exactly? More building of coal, gas, and other sources. Does that sound like a good idea?
One of the actual lessons of climate change so far is that energy requirements are going to go *up* substantially and faster than they were before. Lots of the technologies used to deal with the effects of climate change require a massive amount of power, eg, desalinization, air-conditioning. Nuclear power and solar are probably the two most rapidly scalable for that.
Total:
I don’t know. If there’s going to be a plan, maybe the plan to supply more energy sources should be coupled with a plan to reduce additional energy use. I’m concerned about a plan to supply more energy sources that are dedicated to new energy uses.
The point of the above post is not so much about these details but with the proposal to devote public resources to it (the “five-year plan”). I accept the argument that, in the modern economy, anything that is done will require public resources. I’m questioning the idea to devote these resources to gas-guzzling data centers. This seems like a separate question from expanding nuclear and solar power.
“If there’s going to be a plan, maybe the plan to supply more energy sources should be coupled with a plan to reduce additional energy use. I’m concerned about a plan to supply more energy sources that are dedicated to new energy uses.”
‘Should’ is a lovely thought, if entirely implausible.
Total:
Again, to return to my post, people are suggesting that public resources be devoted to these plans. They’re the ones who are talking “should,” so it seems reasonable for the rest of us to talk “should” too.
Actually, no, because then you’re locked into a larger argument. Stay with just “no”.
Coal production has been steady since 1950s’. Whenever I hear NPR ‘green’ guests whose medications need to be adjusted, I think of an example of how to cheat with statistics,
An uninformed person gets an impression how coal is diminishing with every passing year and the plants are closing down, while unicorns and rainbows are in abundance. What they are playing with is the % of the total mix that is going down. Meanwhile, the absolute amount of coal used in any given year has been basically the same since 1950, and that’s all that matters.
https://www.eia.gov/energyexplained/coal/use-of-coal.php
https://www.statista.com/statistics/184333/coal-energy-consumption-in-the-us/?srsltid=AfmBOooa1LvRJoGf3ZQ9MF6ARJqHE4HfJNQrMjSA57YFHQ3xMxDFcY6z
Heck, worldwide, burning wood for energy (heating, cooking) is larger than all renewables combined.
Any energy source that is labeled ‘dirty’ is the cleanest there is, because of its energy density.
“An uninformed person gets an impression how coal is diminishing”
There is no mystery here. Coal is diminishing in Europe and the US but has been rising in China (seems to have finally peaked in 2025 and hopefully start declining) and India. So overall carbon emissions haven’t yet declined, which is very very bad for the future. But this state of affairs is hardly the fault of “NPR ‘green’ guests” so what’s your point?
“More building of coal, gas, and other sources”, that sounds good! Let’s go with “other sources.”
If you want to provide 1 GW continuously, you can do it with solar + batteries or with nuclear, for surprisingly comparable cost (of the order of $10 B). With solar + batteries you need to be constantly replacing batteries, since they do wear out, but with nuclear you have to deal with nuclear waste and with buying new fuel. The operating cost of the nuclear reactor is about double the cost of solar+batteries. And new battery technologies are driving that cost down much faster than the cost of nuclear waste disposal is coming down.
AI data centers are huge consumers of energy. A single AI data center might use about 100 MW pretty much continuously. The server farm itself, plus associated facilities (cooling, electrical substation, etc.), typically occupies 30 to 50 acres, but a couple of thousand acres are needed to supply enough solar power to run it. That’s a couple of square miles of solar panels, per data center, which sounds like (and is) a lot, but it’s far from prohibitive. Amazon uses about 250 square miles of solar farms (many of them not actually owned by the company), including the Union Solar Farm in Ohio, which is over 20 square miles (you can read about it here https://www.acciona.com/projects/union-solar-farm ).
The big advantage of nuclear is that it provides power 24/7 independent of weather, whereas your solar farm can be unproductive if you have a couple of weeks of heavy clouds (it’s not really feasible to have so much battery technology that you could get through days of discharging). But it’s nearly never cloudy over a really large area for a long time, so if you have enough long-distance transmission you can handle it if one region is cloudy for a week or even more. Adding more transmission is, at least for now, cheaper than building and operating nuclear power plants and dealing with the waste they produce. But the cost of waste disposal could come down a lot if the government ever opens a waste repository like they’ve been planning for thirty years.
Over the past five years, in the US, we’ve added a lot of electricity from solar (200,000 GWh per year) and wind (140,000 GWh per year) and some battery capacity (30 GWh per year), and quite a bit of generation from natural gas (120,000 GWh per year)…and nearly nothing else. A bit of new nuclear came on line, in Georgia, but at great expense: originally budgeted at $9 B, the two reactors ended up costing about $36 B. And even that wasn’t enough money to pay all of the costs: the primary contractor (Westinghouse Electric) went bankrupt in 2017. I expect the future to look like the past, only more so: batteries have gotten cheaper, solar is still very cheap, fossil fuel prices are high and volatile, and nuclear plants are very expensive to build.
source: trust me bro
Phil worked for the department of energy for decades at Lawrence Berkeley Natl Lab and then became a private consultant to industry on electricity production and consumption. Many of us have been on the blog long enough to know that. it would be good if he linked some sources yes, but I actually do trust he has these numbers largely correct.
Good points, but contra this the high costs of nuclear appear not to be because the technology has mysteriously gotten more expensive, but because regulatory issues plus a dwindling of expertise have made it expensive. (We could decide otherwise, in other words.) If I remember correctly, cost/Watt for a nuclear power plant in S Korea, for example, are about 1/3 of what they are in the US.
Right. The deregulation of the 1970s made it riskier to invest money in nuclear plants (if the plant is less safe because there are less regulations on its construction then it is more risky to invest). We need to re-regulate, then nuclear will become cheaper and competitive with S. Korea.
France is also good at building nuclear plants, we should look at what they are doing.
If only S Korea would adopt Japan’s policies, they could cut another 13 months off the median construction time (https://hannahritchie.substack.com/p/nuclear-construction-time). What you are sweeping under the rug here are the reasons for these delays – you clearly seem to believe these are bogus concerns (based on a number of your comments on this post). While I agree that many opponents of nuclear power have not carefully thought through the total impacts of nuclear and other energy sources, I’m not willing to dismiss the public concerns as lightly as you are. Fossil fuel harms are much larger and more immediate than nuclear (as you say), but I’m not aware of needing to safeguard waste from fossil fuel plants for thousands of years. In a world where I have little confidence that the US “democracy” will last another 10 years, who do I trust to keep nuclear waste isolated for thousands? Of course, a failure in this safekeeping is still far from a worldwide disaster, more likely with localized impacts. But if we apply the logic that localized impacts are small in the great scheme of things, then all kinds of policies could be justified which I would find offensive (many airline safety standards, for example).
I’d like to note that nuclear waste is actually fuel its uneconomical to use/reprocess. Ie, its currently cheaper to mine and process new fuel.
Anonymous –
In SoKo, there’s centralized planning, strong regulation, state‑owned utilities, state‑backed financing, and strong political support for a strong state.
Seems to me that you’re avoiding certain aspects of this issue (assuming you’re the same anonymous as elsewhere on the thread).
Once again, nuclear power requires massive up-front capital, long build times (even accounting for excessive regulation), huge downside risk, a long ROI horizon, and a specialized workforce.
The NRC, excessive as it might be, just doesn’t cause all of that.
I’m curious as to what % of the cost of nuclear you’d assign to excessive regulation as opposed to inherent properties of the industry?
And I’m curious as to how you think those issues could be addressed on a large scale without functional state capacity (which exists everywhere that there’s significant nuclear power).
Raghu –
Could you put some meat on the bone where you pin the higher cost for nuclear here vs. South Korea mostly on regulation? From what I’ve seen, that’s too simplistic.
Contra my friend Anonymous, I think higher levels of centralization there (facilitating *uniform* regulation) – that is associated with standardization, state-backed financing and state-owned utilities – is probably more explanatory.
Yeah, I should have given sources, but that would have required me to check everything and find sources for it all. I guess I could let chatGPT do that. I stand by the numbers at a factor-of-two level.
There’s a lot of squishiness when you compare nuclear (24/7 generation, very high up-front cost, history of cost overruns that makes it very risky to lenders) to solar (intermittent and non-dispatchable power generation, long periods of zero or low production even if paired with batteries). So the first paragraph is kind of iffy.
The rest of it… oh man, even documenting the “250 square miles of solar farms” for Amazon is a bit of a pain.
Here’s a list of Amazon solar generation sites
https://sustainability.aboutamazon.com/content/dam/sustainability-marketing-site/05-carbon-free-energy/Amazon%20Carbon-Free%20Energy%20Projects.csv?utm_source=chatgpt.com
You can add up the nameplate capacities, apply a reasonable value for acres per MW of nameplate capacity, and see if you get a number like what I claimed. If I’m off by more than 30% LMK and I will offer a genuine apology.
“The big advantage of nuclear is that it provides power 24/7 independent of weather”
Not quite, since most NPPs are cooled by river water, heat and drought can and has in the past repeatedly had significant effects on nuclear energy production, especially in France.
Orther than that, I agree with your points. Worth noting that China has over 10 years increased its solar electricity production from 45 to 1173 TWh and wind from 186 to 1128 TWh. These are huge numbers (actual generation, not nominal capacity!). Nuclear has increased from 171 to 485 TWh.
https://en.wikipedia.org/wiki/Electricity_sector_in_China
“Nuclear power is extremely safe — even including Chernobyl etc the deaths per Watt are about 4 orders of magnitude less than what we blithely tolerate for fossil fuels…”
Sorry to be tangential, but (and I’m actually asking a question, not advocating a position here): do we really know this. It is my understanding, possibly incorrect, that at least earlier types of nuclear reactors produced radioactive spent fuel that remained hazardous for up to 10,000 years. So may there not be some still-to-occur deaths associated with future mishaps arising from spent fuel? Perhaps a large number of them?
Even if modern reactors do not create such long-lived waste, can we really have confidence that the already existing stuff will be safely managed for an eon?
I had the same reaction/question. Years ago (decades actually), there were concerns about the long life for radioactive waste. I haven’t heard much about this for a while now. Does this mean the concern is no longer valid? I also remember that the concern was not as much about leaks of the material (though that was of some concern), but about the idea that the time scale exceeded the known time scale for most human institutions. Thus, nuclear power was termed a “Faustian bargain.”
I’d like somebody more knowledgeable to clarify the dangers, or lack thereof, of long-lived radioactive wastes.
It’s important to keep in mind the scale of the materials. The US gets about 10% of its energy from nuclear power, and the amount of waste is about 25 ml per person per year (1/10 of a soda can). It’s not some monstrous amount. Also, the alternative isn’t nothing, it’s fossil fuels, whose emissions kill ~30000 people /year just from air pollution. (By the way: a coal-burning electricity plant emits more *radiation* than a nuclear power plant.) All this follows from the specific energy of nuclear fuels being about 20,000,000 times that of fossil fuels; the material needed per Joule is ridiculously small.
Aren’t you mixing emissions with waste material? The emissions may be small but the waste material may be larger and highly concentrated. Does that matter?
I combined both emissions and waste into one comment; sorry, that was perhaps unclear. However, I’ll emphasize that the radioactive waste material is nowhere close to what could be considered large, 2000 metric tons/year in the US, which is less than 10 grams per person. My point was just that overall harms (emission, waste, whatever) are tiny for nuclear power compared to what we seem to be blithely comfortable with from fossil fuels.
“So may there not be some still-to-occur deaths associated with future mishaps arising from spent fuel?”
If we’re going to do that, we have to do it for all fuel sources. How many will the waste pollution from solar manufacturing kill in the future?
The post reminds me that the NSF just announced its new “X-Labs” funding program, https://www.nsf.gov/news/nsf-announces-15b-nsf-x-labs-initiative-pursue-generational . It’s unclear what it actually is, but the implication is that it’s supporting stand-alone labs, “independent teams of researchers, engineers and entrepreneurs”, i.e. not academic research or even national labs. The amount is $1.5B over 10 years, which if it’s actually spread over 10 years is 2% of the NSF budget, so not huge. I don’t have enough information to know what I think about this, but I thought it was somewhat relevant to the post!
Prescribed plans for spending public money on scientific endeavours can be useful stimulants for scientific advance. The human genome project (a 15 year plan completed ahead of schedule) is a good example. It helps if the plan is well defined and has either required technologies available or strong expectations of developing these. It was actually run as consecutive 5 year plans.
One can contrast that with the 5 year plan for returning US astronauts to the moon set out in 2019 – paraphrasing the former vice president Pence (can’t remember the exact quote) “..failure to achieve our goal in the next five years is not an option”. The context then and now was/is also competition with China, who no doubt have their own 5 year plans for going to the moon. As with the recently announced Genesis Mission which seems a rather Soviet style centralized enforcement of how energy science must be done, the aim seems to be to redirect public funds away from more “traditional” research centres to the corporate sector. If the effect of that approach on the US Space program is any guide, the prospects don’t look so positive even if lots of hot air can be predicted.
The nukes the Chinese are building that have people’s shorts in a knit are fusion plants. They have a very substantial R/D program on these with 5 (or 6?) different methods under study. Their tokamak reactor is set to go on line in 2030.
If they whip the fusion problem first it will be a massive shot for their economy and not just for “data centers”. Admittedly, the methods would be very difficult to disguise and most everybody would be able to chime in soon. But … soon is a substantial gap to overcome, especially in a society that is suspicious of public investments in large scale science and, of course, is spooked by the very idea of nuclear power.
Btw, the data that Raghu is using appears right. And, of course, it is only a matter of time before someone decides to see if LLMs can solve the “socialist calculation” problem. 5 Year plans may be closer than we think.
It’s good to push back on statism, but you are barking up the wrong tree here. The problem is that power generation, especially nuclear, is already centrally planned. Facebook cannot start a nuke build without appealing to the existing central planners first.
Yes, central planning is a total failure responsible for massive waste and environmental harm. But its not new; the NRC has been blocking nuke builds for decades.
Were the tech zillionaires radical libertarians, theyd still have to appeal to the state. That’s just how you build energy infra without going to prison.
Anonymous –
Every country that has significantly more nuclear energy than the US, with the possible exception of Finland, got there through state‑centered planning, state financing, and state‑owned or state‑directed utilities.
One of the more substantive obstacles to expanding nuclear in the U.S. has been the popularity of “drown the government in a bathtub” anti‑statism. The result is a technology that’s usually state‑dependent operating inside a hollowed‑out state that’s been systematically undermined and lacks capacity. The sadly ironic part is that the people who attack environmentalists and blame them for the lack of nuclear capacity are largely the same people who have spent decades undermining the state powers that usually create nuclear capacity. That might change with smaller-scale nuclear to some extent but broadly, even that technology is an expensive investment with a long delayed ROI and high liability. Most investors have better options with their money. That’s why attacking “the state” is a pro-nuclear cutting off your nose to spite your face.
We (Finland) have ≈40% of our electricity nuclear, while ≈80–85% is carbon-free (nuclear, hydro, wind, solar). Our nuclear is not exactly private: “the state owns one operator outright, co-owns the other through industrial proxies, decides whether new plants can be built at all (twice — Government and Parliament), regulates safety arm’s-length via STUK, and underwrites the back end.” https://claude.ai/share/6972e7a5-83fc-48ef-a7b6-c8990a4839ae, although there’s an effort to make SMRs easier. But in general the system is very market-driven. Our mean electricity price is about the lowest in Europe, and volatility of the price is the highest.
Data centers will be here too, for example Microsoft will soon have ≈500MW online. Many new data centers take part in district heating, and in general we have no strong opposition against them, although some worry about the growing base load (for a good reason).
In terms of power, new data centers are almost exclusively for AI, not at all for crypto, and building enough for the demand will be a challenge. Yes it looks like a sort of industrial revolution, and I wouldn’t be too cynical about the voices of “billionaire software executives”, as we are already now short of compute, GPUs and memory.
Many new US data centers run on gas turbines which is not optimal, and some sort of centralized push towards renewables (with batteries), grid, and nuclear would be wise. Opinions obviously vary on to what extent nuclear is the way to go, vs. solar and batteries.
Climate vs. industrial growth is a hard one. The theoretically preferred solution there is to not punish a single sector, instead impose a tax on emissions. That hasn’t got too well in the US as far as I know. ;) Don’t blame the data centers though.
>Every country that has significantly more nuclear energy than the US, with the possible exception of Finland, got there through state‑centered planning, state financing, and state‑owned or state‑directed utilities.
What are you trying to say here? This seems to me to be just the trivial observation that nuclear is an industry where everyone involved gets killed, bombed or imprisoned if they don’t have a strong army backing them first. Which is also why I don’t get Andrew’s proposal; if he thinks that it’s hypocritical for the tech industry to talk to govt about nuclear energy, is he expecting them to go build the nukes in Ancapistan and then deal with inevitable CIA visit in some free market way too?
Anonymous –
It seems to me that you’re reconciling a desire for more nuclear with an “anti-statist” world view, and I don’t see how they’re reconcileable. But maybe I misinterpreted your view?
Coupling nuclear development to political decision making and to onerous governmental approval processes has dramatically increased costs, in the rare case where the regulations are not just a fig leaf over a technological ban. I also understand pragmatically that this will not change. The infrastructure is too easy to convert to weapons, too dangerous if unmaintained, whatever, it is how it is.
Where I take issue is when we observe this reality that nuclear is inevitably tied to government power, and conclude that it’s good. No, it’s not, the main effect is to restrict the technology and dramatically increase costs. I can acknowledge that there is no alternative without going all Stockholm Syndrome about it, you know?
In Medieval Europe, there were hundreds of tiny states competing with each other for existence. There was no world police – if you could provide cheap energy for your citizens, you did, or your neighbor did and then took your territory under his management too. We do not live in that world of ruthless competition. If the four or five truly independent governments in the world collectively fall for some stupid idea, whatever, it is how it is, we can carry on with those bad decisions, it’s not existential. Doesn’t mean we have to smile and say “thank you” when they serve us the fruits of that stupidity though.
Anonymous –
I’m not clear on what your point is about nuclear. As for the big picture, figuring out a counterfactual about whether we’d all be better off without large, centralized states is too complicated for me to evaluate, and it’s kind of a moot question anyway. Given that we have large states and will continue to have them absent Armageddon, I think it’s better to have functional large states with the capacity to build nuclear than non‑functional large states (like ours) that can’t.
I don’t think about this in terms of “good” or “bad.” Nuclear is what it is – a technology with high costs, long timelines, and high‑impact risk.
My point is that large‑scale nuclear build‑out pretty much depends on a functional state. Maybe small‑scale nuclear would change that, but we’d need an absolutely enormous amount of it to significantly offset growing energy demand.
In the US, we don’t have the required functional state to do either, and we won’t have much nuclear unless we do. I don’t see a world where massive non‑state production happens even with zero regulation, and I don’t see the public accepting nuclear without regulation.
>My point is that large‑scale nuclear build‑out pretty much depends on a functional state.
Only in the same sense that passage through the Strait of Hormuz depends on a functional Iranian state.
We can all rally around the word “functional”, but we will disagree on the specifics. I think the American problems with nuclear can all be laid at the feet of the NRC, not blamed on Republican saboteurs, and a “functional” government is one where nuclear projects do not have to pass NRC approval.
Anyway, I just wanted to gripe that nuclear has always been captive to central planning, and going “oh gee awful hypocritical of the CEOs to plead with the govt” is just..wow.