Due to increased regulation etc you cannot just translate 1985 $, £ or Euro to a 2026 one. There is an actual example in the UK Hinkley Point C current estimate $43b, (£35b) where as sizewell B commissioned in 1987 was $3.2b billion (£2b) or about $7b in todays $. This is probably the worst example but makes the point.

    > $5.6B actually sounds like a good deal. It outputs 2GW+ of power. 

I don't understand. Are you talking about 1985 dollars of 2026 dollars?

After some research, I learned that thermal powerplants (coal/gas/oil) completed in 1985 cost about 0.8B to 1.2B USD per GW. 5.6B USD in 1985 for 2GW sounds like a terrible price -- at least twice the cost.

assuming 300 days/year, 1c/kwh and ignoring opex that's $150m worth of electricity per year.

7B for the first set of batts.

Then 7B in 2046 money which is probably $15 today.

Unfortunately France can no longer build nuclear plants cheaply either. All of the recent nuclear plants built by the French state owned company EDF in France, Finland, and the UK have seen enormous cost and time overruns.

Cumulative emissions matter. We simply don’t have the time to wait the 20 years it takes to build new nuclear plants.

Source please? The numbers I have seen of real opex paint a different picture. In general, nuclear plants close because of cost.

> 100% solar

100% solar is a straw man though, as much as the simplicity of it sounds nice.

> Further if we build more nuclear we'd be better at it and it would be cheaper.

This is far from being clear, nuclear is one technology that tends to have increased costs the more we do of it. Even in France!

The costs of the French nuclear scale-up: A case of negative learning by doing https://www.sciencedirect.com/science/article/abs/pii/S03014...

Human labor is very expensive, and every time we make humans more productive, that makes human labor more expensive, because their time becomes more valuable. Technological growth does that.

The cost of nuclear is primarily in labor and long-term financing, due to the very long lifetime and upfront labor cost. Until somebody has some sort of technological breathrough to decrease the labor cost of nuclear, it's not going to be able to compete. Even decades ago it had trouble, and now it's far worse.

You are talking only about the operations of the nuclear, and ignoring all the high energy process required to mine and process uranium before it can be used as a fuel, and after as waste. But let’s pass this problem to the next generation, they will know what to do :)

Or solar / wind (which mostly anticorrelate) + biomass + storage + interconnectors + smart demand.

The amount of baseload we technically need can be pretty slim.

Take Denmark: fossil powers just 9% of their electricity generation, the majority of it is wind and solar. Wind is strong in evenings/nights, solar during the day.

Then they have biomass (indirect solar) as a form of baseload, more sustainable than coal/gas.

Then there's interconnectors, they're close to Norway which can pump hydro, and Sweden, each day about 25% of the electricity is exchanged between these two countries, and that's a growing figure.

With more east/west interconnectors you could move surplus solar between countries. Import from the east in the morning before your own solar ramps up, export your midday surplus west before theirs peaks, and import from the west in the late afternoon as yours fades.

With interconnectors you can also share rather than independently build peaker capacity. Because a lot of peaker plants only run a small amount of time and therefore much of the cost is in the construction/maintenance, not the fuel.

And of course there's storage, which will take a while to build out but the trendlines are extremely strong. Just a fleet of EVs alone, an average EV has a 60 kWh battery, an average EU household uses 12 kWh per day so an average car holds 5 days worth of power a home uses.

And then finally there's smart demand. An average car is parked for more than 95% of the day, and driven 5% of the time. Further, the average car drives just 40km a day which you can charge in 3 minutes on say a Tesla. Given these numbers (EVs store 5 days of household use, can sit at a charger for 23 hours a day, and can smartly plan the 3 minutes a day of charging it actually needs to do) just programming cars to charge smartly, is a trivial social and technical problem in the coming 10-20 years.

Given this, baseload coal/gas can really be minimised the coming decades. It's not going to go away as a need, but I don't think it requires gas/coal or nuclear long-term going forward.

The issue with "base load" is that people usually omit to mention how much GW they are talking about and for how long. Which makes it a bit of a bull shit argument.

As an insurance against unspecified lack (how much for how long?) of wind and solar (and batteries, cable capacity, hydro, etc.) base load is supposed to swoop in and save the day when those temporarily fail locally. So, it's a valid question to ask how much insurance we need against that. Nobody seems to really know. There are loose estimates of course. And people seem to assume it's months and that renewables are going to 100% be offline throughout that very very long period. In reality in most connected energy markets, we have a short gap of a few weeks or so in winter at higher latitudes of reduced output that we already manage to cover with flexible generation.

It's more constructive to think in terms of dispatchable power rather than base load. When the sun doesn't shine or there is no wind, it's nice if you can quickly bring online additional generation, tap into battery reserves, or bring in power from elsewhere (via cables). That favors flexible power, not inflexible power. Nuclear and older coal plants are a bit inflexible. Shutting down and starting up a nuclear plant is really slow and expensive and requires a lot of planning. And especially older coal plants need quite a bit of time to bring their boilers up to temperature such that they build up enough steam pressure to generate power. Until then, they are just blowing smoke out of the chimney. Modern coal plants are a bit better on that front. Same with gas plants.

The modern ones only need about 10-20 minutes or so. Still quite slow but something you can plan to do. Slow here means expensive as well. Because shutting them down when there is a surplus of renewables (which is a very common thing now) is really inconvenient. Which means consumers have to pay extra for perfectly good electricity from renewables to be curtailed. That happens by the GW in some markets and keeps consumer prices higher than they should be because they have to pay for gas/coal that is technically not actually needed.

Batteries have a much lower LCOE than gas or coal plants (never mind nuclear) and it's being produced by the TWH per year now. A lot of markets are serving much of their peak demand using batteries now. Australia and China are good examples. Even in the US, you see batteries being deployed at a large scale now. That's starting to push gas and coal out of the market. A gas peaker plant that rarely runs is just really expensive.

A requirement for base load is a fallacy promulgated by fossil fuel preservation lobbying

it actually is a choice between nuclear and PV, because base load supply is an obsolete concept. Because actually nuclear is terrible in a grid increasingly full of nearly-free variable sources (solar&wind). The nukes need to stay at 100% all the time selling their power at a high fixed price to have any remote chance of being economical. Cheap variables push nuke's expensive power off the grid during the day, and increasingly into the evenings with batteries. This is unavoidable in an open energy market, and is fatal to the economics of nuclear. You cannot make them work without massive state subsidies.

Gas is far better suited economically to backstop a variable grid. I wish it werent true, because i dont hate nukes, but it is just economics.

I will also point out that california is down to 25% fossil sourced power in 2025, from 45% in 2022. Due to renewables and batteries, and there's far more coming. The amount left to backstop on gas in a few years could plausibly be 10%! which is amazing.

It was not a good idea for Germany (and certain other parts of the EU) to be so dependent on Russian gas. It was also not a good idea to become dependent on LNG from Qatar or the US. Spain uses natural gas from Algeria (via Morocco), also not great. Italy also gets some from Algeria/Tunesia, still not great. Inside of Europe, we are far too dependent on Norway. Not because Norway is likely to turn on us (or we on them), but because the pipelines are relatively easy to disrupt.

The transition from coal to gas gave us cleaner air (and less CO2) but it definitely also had costs, some of them in the form of many thousands of dead Ukrainians, some of them in the form of concessions to the US.

And $ cost is a poor metric to chase when what you really care about includes a lot more-- exposure to the whims of geopolitical forces you can't foresee or control, which have both $ cost and more.

Power distribution is a natural monopoly, power production is commodified/competitive business.

Because if a thing is valued by thing-consumers at x and you set the price to <x, then you are incentivizing people to use more of the thing than they need, even to waste the thing. This thus requires more infra than is actually needed or wanted.

This doesn't go away under socialism/communism/collectivism. If you set the price too low, you either have to build far more production capacity at public expense than needed, or you cope with regular blackouts.

> China shows that it can be built very cheap and fast with good supply chain

I mean, thank you, the USSR already showed this, no more is needed.

25 year replacement for solar is a myth. [1] . They may degrade to ~80% but they keep on working and producing, so far it seems almost indefinitely.

[1] https://www.ecoticias.com/en/goodbye-to-the-idea-that-solar-...

"Ironically, what originally motivated pumped storage installations was the inflexibility of nuclear power. Nuclear plants’ large steam turbines run best at full power. Pumped storage can defer surplus nuclear power generated overnight (when consumption is low) to help meet the next day’s demand peak."

https://spectrum.ieee.org/a-pumped-hydro-energystorage-renai...

To use the strongest plausible interpretation, the writer could be from Belgium, where the meaning of 2.256 depends on whether they wrote the number in French, Dutch or German.

The Belgians apparently typically invert the meaning of . and , in numbers (from how they are used in the US).

  To make large numbers readable, Belgians use either a period (.) or a non-breaking space. Example: Two thousand thirty-six is written as 2.036 or 2 036. In formal Belgian French, the space is increasingly preferred over the period to avoid confusion with the Anglo-American system, but the period remains very common in Belgian Dutch and everyday shorthand.

I would guess Europeans tend to be better at SI units than people from the US. And let's not mention the the cancer of changing the value of G depending on context.

mount signal, the largest PV plant in california makes 1,200GW/hrs per year. it would still take ~15 copies of mount signal for a single diablo canyon https://en.wikipedia.org/wiki/Mount_Signal_Solar

my whole point is solar is great, but the insane scale it requires to get reasonable output is really underestimated. you would need solar fields 100sqmi big. probably many of them. solar alone won’t be the future of humanities energy needs because it’s not efficient enough. we should still keep building solar. but if we aren’t building nuclear too its not enough growth

It doesn't. Watts were a mistake by whatever committee it was that standardized unit names. Power should not have been given a unit; it should have been left as ∆energy/time just as velocity is distance/time.