Comment by epistasis | Hacker News Reader

Comment by epistasis | Hacker News Reader

Parent Full thread

IMHO this podcast is way too biased in favor of nuclear, and exemplifies the entire problem with the nuclear industry: too much magical thinking and not enough hard retrospective self-criticism.

Are we going to be able to construct nuclear in the future? Probably? Maybe? But if folks like this are in charge of the efforts, then my answer is a firm "never."

Take for example when Jigar Shah went on their podcast. Shah is a huuuge nuclear fan, and IMHO is way too optimistic too (for example the DoE's LPO Liftoff report on nuclear gave it a 30%-50% cost advantage out of nowhere in order to make it look the least bit plausible). But for all of Shah's blind spots on nuclear, at least he tried to give the Decouple podcast a small shower of reality:

https://www.youtube.com/watch?v=HwN1MCtBkVk

(BTW, Shah has said on recent podcasts that there will be more announcements of new AP1000s soon, and since he's giving out so much money to get them built, we should probably believe that. IMHO, I'd bet 10:1 that it will be wasted money that would be better spent on batteries and solar, but when it comes to subsidies and government spending, money is not actually fungible, so I support more shots on goal, even if they are very very long shots)

Nuclear power, despite its larger costs, is the only form of non-intermittent green energy production other than geographically limited options like dams and geothermal. Once wind and solar start saturating energy markets during peak hours of production, each new wind and solar plant has diminishing returns. While the plants themselves are expensive, producing wind and solar power at large scale and offsetting intermittency incurs costs in the form of increased transmission capacity and storage requirements.

People really often don't comprehend just how titanic the storage requirements are just to provision half a day's worth of energy to offset daily solar fluctuations. The globe currently uses 60 TWh of electricity per day. That's 30,000 GWh for 12 hours of storage. Keep in mind, electricity is only about half our total energy use, heating and transportation also need to be electrified which will drive this number up. And lastly, diurnal storage isn't the only storage we'll need. We'll also need seasonal storage to offset weather fluctuations.

When we stop asking "what's the most cost-efficient way to reduce CO2 emissions by a percent or two?" and start asking "how will we build a 100% carbon-free grid?" nuclear becomes a much more attractive option.

> While the plants themselves are expensive, producing wind and solar power at large scale and offsetting intermittency incurs costs in the form of increased transmission capacity and storage requirements.

My understanding is that solar and wind are cheaper than nuclear even when accounting for storage.

That's before you get to the externalized costs, such as waste disposal and decomissioning.

> My understanding is that solar and wind are cheaper than nuclear even when accounting for storage.

This understanding is either based on geographically limited storage options (hydroelectric storage), or is incorrect. The amount of batteries required to even out intermittent sources is many times more than the amount of batteries produced. The cost of a 1 GWh facility is very different from a 1 TWh facility. The latter is 2x the amount of batteries produced worldwide each year. But it's only 2 hours of the USA's electricity consumption. That's how big of a mismatch there exists between battery supply and the demands of grid storage.

> That's before you get to the externalized costs, such as waste disposal and decomissioning.

Waste disposal and decommissioning are already factored into nuclear power's cost. They have to pre-pay the cost of disposal and decommissioning.

You need to update your talking points.

Vogtle is $170-$180/MWh. https://www.powermag.com/blog/plant-vogtle-not-a-star-but-a-...

New solar is on average $40/MWh: https://emp.lbl.gov/sites/default/files/utility_scale_solar_...

Back in 2022 when prices were higher, NREL put batteries at $482/kWh:

https://www.nrel.gov/docs/fy23osti/85332.pdf

That's $87/MWh for storage, backing out the battery lifetime and round trip efficiency from the same report.

So depending on the time usage of electricity, the average solar+battery installation will have an averaging of $40/MWh and $127/Mwh electricity.

That's using old prices. It's cheaper today, and will get cheaper in the future.

The cost estimates are for a 4 hour storage system - not enough for diurnal let alone seasonal storage. Furthermore, actually attempting to build storage systems at grid scale would cause battery prices to skyrocket. Even just 4 hours of electricity supply for the world works out to 10,000 GWh of batteries (global electricity demand is ~60 TWh per day). In 2023, 680 GWh of batteries were produced [1]. If you were to try and build 10,000 GWh worth of battery storage you'd buy up all the batteries on the market, causing the price to skyrocket. And even if you literally bought every single battery produced in 2023 you still wouldn't have 1/10th the batteries required to provision just four hours of storage (and not to mention you'd kill 100% of EV production in the process). The scale between the batteries produced and the amount of batteries required to deliver even small amounts of storage are vast.

And we're not even talking about electrifying other fossil fuels uses like heating, transportation, etc. That's going to make electricity demand even larger.

1. https://autovista24.autovistagroup.com/news/which-manufactur...

> My understanding is that solar and wind are cheaper than nuclear even when accounting for storage.

In Ontario nuclear costs 10¢/kWh while wind costs 15¢ and solar 50¢ (Table 2):

* https://www.oeb.ca/sites/default/files/rpp-price-report-2022...

And when wind goes to zero at night, then (natural/methane) gas generators are often spun up.

> That's before you get to the externalized costs, such as waste disposal and decomissioning.

Which wind and solar also have. You may be able to extract some metals from solar panels, but turbine blades are not (AFAICT) recyclable.

> is the only form of non-intermittent green energy production other than geographically limited options like dams and geothermal

That "is" should be a "was".

> The globe currently uses 60 TWh of electricity per day. That's 30,000 GWh for 12 hours of storage.

Bloomberg New Energy Finance tracks battery production capacity announcements, and has a total world capacity of 7.9 TWh/year by the end of 2025 [0]. With a warrantied life of 12-20 years, that 2025 manufacturing capacity would support 95-158 TWh of ongoing storage.

Keep in mind that the talking points you are drawing from are several years of out date and assume an audience that knows little about the world of energy. HN is not the right place to pull out the standard bag of tricks from the electronic information toolkit.

For example, you claim a need for "seasonal" storage out of nowhere, but do not consider the obvious (to the HN audince at least) idea of overcapacity such that there's enough generation at the seasonal minimum to cover needs, and oversupply the rest of the year.

Already, many countries in Europe have seen single-digit /MWh average prices this month [1]. That's going to become far more common during most of the year.

[0] https://twitter.com/colinmckerrache/status/17758808881626562...

[1] https://x.com/LionHirth/status/1777309657804992969

> For example, you claim a need for "seasonal" storage out of nowhere, but do not consider the obvious (to the HN audince at least) idea of overcapacity such that there's enough generation at the seasonal minimum to cover needs, and oversupply the rest of the year.

No amount of overcapacity will make solar panels produce electricity at night. You'll at least need diurnal storage, no amount of overproduction will help solar power offset the day and night cycle. Overproduction is also not a silver bullet: solar production during cloudy days and wind production during still winds are often under 10% of average output. You'd need a massive amount of overproduction in order to avoid building large amounts of storage.

As for your battery production predictions, these seem very far-fetched given the history of lithium ion battery production. Bear in mind that production capacity is a different figure than actual production. Production capacity is what factories theoretically could make if given unlimited input materials. But that's not what matters, what matters is actual production figures - how much battery capacity was actually built.

2023 saw 680 GWh of lithium ion batteries produced worldwide: https://autovista24.autovistagroup.com/news/which-manufactur....

2021 this figure was 476 GWh: https://interactanalysis.com/insight/lithium-ion-battery-mar....

Will we actually see 7.9 TWh of batteries produced annually by the end of 2025? The growth over the last two years was only about 200 GWh of capacity, so seeing over 7,000 GWh of production increase over the next two years seems a bit far fetched.

> IMHO this podcast is way too biased in favor of nuclear, and exemplifies the entire problem with the nuclear industry: too much magical thinking and not enough hard retrospective self-criticism.

Why do things have to be balanced or unbiased? The anti-nuclear folks have their platforms, why can't the pro-nuclear people have theirs?

It is "biased" because it is run by a nuclear advocate? But given they had a three-part series on the problems and what went wrong with Vogtle, it's not like they're trying to hide everything under the rug.

A little while ago they published a video on a debate with an anti-nuclear individual:

* https://www.youtube.com/watch?v=XfL40rSFT00

> Take for example when Jigar Shah went on their podcast.

I have not gotten to that episode (yet?), but have heard Shah on podcasts like Odd Lots discussing his work more generally.

> IMHO, I'd bet 10:1 that it will be wasted money that would be better spent on batteries and solar, but when it comes to subsidies and government spending, money is not actually fungible, so I support more shots on goal, even if they are very very long shots)

As someone who lives in Ontario (same as Decouple host), my experience with renewables haven't exactly been thrilling (click on "Supply"):

* https://www.ieso.ca/power-data

Every time wind goes down (including down to zero), (natural/methane) gas generators tend to have to be powered up (putting junk into the atmosphre). I'd rather we just build more CANDUs for base load and let hydro take up more of the variable load.

Sure, have a pro-nuclear platform, but it would actually be productive it it attempted to improve the field instead of tilting at windmills of political goals.

That's the difference between the success of renewables and the failures of nuclear. The renewable folks went out and built the stuff, made it scale, and fixed the problems. The nuclear folks are star-gazing, hoping and dreaming, without executing at all.

> Sure, have a pro-nuclear platform, but it would actually be productive it it attempted to improve the field instead of tilting at windmills of political goals.

If there are factual issues with what they say, feel free to post a comment correcting them. Just because something is (allegedly) "biased" does not make it wrong.

> That's the difference between the success of renewables and the failures of nuclear.

There's also a difference between the failures (and inadequacies) of renewables and the success of nuclear:

* https://www.ieso.ca/power-data § Supply

When wind goes to zero at night (which it has done in Ontario), what are the options for keeping people warm in January and cool in July?

There's no point in correcting factual errors because what they are attempting isn't even directed towards something that improves the chances of nuclear getting deployed.

The decisions that will determine how many new AP1000s get ordered in the US or BWRX300s get ordered in Canada come down to financial and logistical and contractual issues around their construction. The decisions will be made in utility board rooms, not in the court of public opinion.

To improve those chances, there needs to be serious introspection on how to drive down construction costs, prevent construction delays, etc. etc. etc.

Maybe there are regulatory changes that could help, but I'll be damned if the show ever talks about those in concrete constructive ways, instead regulations are merely an obstacle and an excuse for failure.

In the renewable space, the hurdles are talked about in ways that they can be overcome, addressed, side stepped. In the nuclear podcast promotion space, the hurdles are immovable, somebody else's fault, and an excuse for the failure of current projects.

I never hear a podcast about renewables extolling the nature of solar tech as something that just needs to be done because the tech is so cool, or repeating the failures of nuclear tech as inherent to the technology, because that's a pointless waste of time. Yet that sort of navel gazing is the only type of pro-nuclear podcast I have ever found, of about five that I have stumbled on over the years.

Nuclear does not need any more cheerleaders. It needs more hackers.

> To improve those chances, there needs to be serious introspection on how to drive down construction costs, prevent construction delays, etc. etc. etc.

Build the same designs repeatedly. No other major changes need to be made IMHO.

It worked for France, Ontario, and Japan ((A)BWRs consistently built in 4-5 years)[1] in the past, and it works for every modern widget we produce in factories now.

Ontario's refurbishment is on-time and on-budget because of a repeatable process:

* https://archive.today/xoQDQ / https://www.theglobeandmail.com/business/article-darlington-...

And sometimes not only on-time, but finishing early:

* https://www.ebmag.com/darlingtons-unit-3-back-online-nearly-...

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

Reply on news.ycombinator.com