Comment by crote | Hacker News Reader

Comment by crote | Hacker News Reader

I am not surprised.

Over the last few years, social media has been absolutely flooded with people claiming that nuclear is the "one weird trick THEY don't want you to know about" to solving climate change. It is supposed to solve every single problem, and probably even cure cancer in the process - often relying on theoretical or experimental concepts which have never been tried in practice.

Meanwhile, the reactors currently / recently constructed are massively over budget and years behind schedule. Nobody is willing to invest in them unless the government is willing to guarantee a massive subsidy on the electricity produced. Additional reactors will at best come online in a decade or two. And during all this solar and wind are rapidly increasing their market share while being weaned off the initial subsidies. Literally everyone wants to build it and in many areas we are now seeing problems because too many people are building it, leading to overcapacity.

Nuclear was a great idea in the 70s, but got a significant reputation hit due to high-profile nuclear incidents, cost overruns, and long construction delays. It was essentially killed by cheap coal becoming available with the rise of surface mining, just like coal is now being killed by cheap gas.

In 2023 the economic reality is that nuclear simply isn't a viable option anymore, and it would be quite foolish to rely on it to solve climate change in any way.

While this is the standard answer that people who have looked into it deeper give, it still surprises me.

The largest solar farm in the US is about 570 MW. Some have calculated that we're going to need to build two 400 MW farms every week for the next 30 years to reach a "middle-road renewable energy scenario." [1]

That's insane.

(Or we could be building one 400 MW solar farm and one 400 MW wind farm every week instead for 30 years. Point is, though, we're far from being able to do that.)

So we're still going to be struggling in 20 years. Wouldn't we be in a better position then if we had also started the process of building new nuclear plants 20 years prior? "The best time to plant a tree was 20 years ago, the second best time is today," etc.

1. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4443474

> (Or we could be building one 400 MW solar farm and one 400 MW wind farm every week instead for 30 years. Point is, though, we're far from being able to do that.)

That's actually lower than what the USA is doing right now. "The US solar industry installed 6.1 gigawatts-direct current (GWdc) of capacity in the first quarter of 2023"[0]

That is approx 469 MW of power being installed every week.

[0]https://www.seia.org/us-solar-market-insight (first link I could find)

Sorry, I realize I was being confusing by measuring in 400 MW farms: the total required was 800 MW/week, which we're not hitting yet.

We are at the right order of magnitude, though, as you show. The question is whether we can hit the actual number required (which is close to double what we have now) and sustain it for another three decades.

> sustain it for another three decades

You keep saying "sustain", but solar has been growing exponentially and there is still no sign of an inflection. "Sustaining" a rate of growth is the wrong concept to model it.

I disagree. Just because there's no sign of inflection yet in is earliest period of exponential growth doesn't mean that the conditions to keep building at this rate (or, rather, faster) will remain for another 30 years.

We're in a period with low-hanging fruit. Lots of available space. Lots of roofs that don't have don't panels yet. Lots of mines churning our rare elements.

What will happen as solar starts to saturate the market, and the price of electricity continues to fall (thus lowering the incentive to build more farms) while the available space shrinks and the number of rare elements decreases (increasing the cost of farms)?

I'm not saying this will happen, but it was mentioned in the Greens Dilemma article posted and, and is certainly worth thinking about.

So now you want to know for how long it can keep accelerating.

That one is a good question to ask. If the rate of change can be sustained is not. And by the way, linearly extrapolating the current rate of change into the future, like you did on your original comment is equality useless.

But yes, this one comment has a useful model. By the way, you can look how much rare earth elements are actually needed, how much mining them can grow, how much space are there for the panels, and how much the panel price can theoretically fall. I don't think anything I say will convince you, so yeah, go look, because your comment shows a lack of awareness of any of those factors.

But of course, it won't keep the exponential growth for 30 years. That would make the PV output orders of magnitude larger than our current energy consumption. If nothing yields by them, it will stop for lack of demand. Demanding it to keep going for that long is completely unreasonable.

I'm not asking how long it can keep accelerating. You keep making things overly-complicated.

We (per the estimates above) need to be installing about 800 MW of new renewable power generation every week every week for the next 30 years. We're currently at about 400 MW installed every week. So we need to grow to somewhere around or above that target, and then sustain that same rate for 30 years. (Or possibly fewer years if we are installing way above that rate.)

My question is simply, given that large rate of installation, can we actually sustain that rate for 30 years?

Presumably we will eventually be hitting that target rate, at least, and hopefully we will continue to do so for some time. But if markets don't continue to support installing solar at that same pace in 10 or 20 years, we won't hit our targets. That's all I'm saying.

So, you are asking if solar can technically replace all of our energy needs? In a time independent fashion?

That one has plenty of stuff written about. Yes, it can.

One other good question to ask is how much of it will be replaced on the next 5 or 10 years. That's where all the discussion is, because if it takes 30 years, we'd better also invest in alternatives, like nuclear.

(And wherever the price of electricity goes, the cost structure doesn't automatically change. Thus if the market stops supporting investing in solar in 10 years, everything except for natural gas and wind will be already bankrupt. Oh, and if the current changes on the cost structure continue for 10 years - there's some reason for them not to - you can cut wind from that list too. In other words, if you concern is about the cheapest form of generation becoming unprofitable, that automatically means we just don't really want more energy.)

> One other good question to ask is [...] if it takes 30 years, we'd better also invest in alternatives, like nuclear.

Yes. That's been my question from the very beginning of the thread. Glad we got there.

The US installed 25.5GW of clean power generation capacity in 2022.[0] That's 490MW a week, so it isn't really all that insane. Meanwhile, the US is on pace to install 3GW of nuclear in 25 years.

Surely increasing the construction speed of renewables from 490MW to 800MW / week is easier than increasing the construction speed of nuclear from 2.3MW to 800MW / week?

[0]: https://cleanpower.org/market-report-2022/

I think the idea is to change that pace.

In 2023 utility-scale solar generation capacity is expected to grow by 29 GW, and wind by 7.5 GW[1] - that's 700 MW per week, so we're almost on track already. (The EIA's "expectations" might be optimistic, but I don't think it's at all out of the question to roll out wind and solar on this scale.)

1. https://www.eia.gov/electricity/monthly/update/archive/febru...

Not sure why you are downvoted, other than you used numbers and real data which contradicts a nuclear narrative.

The EIA can be trusted with their reports on actual numbers. But their projections are absolutely embarrassing and underestimate solar deployments so much that they have completely discredited the entire enterprise of projections. The graphics showing how embarrassing they are all use the IEA numbers, but the EIA projections are just as bad

https://www.bu.edu/eci/2021/11/10/signs-of-hope-reflections-...

It really depends on the cost. Whatever your analysis says, surely it says something different if the cost of solar goes down 90% over the next decade.

> That's insane.

This type of innumeracy from nuclear advocates is rampant. It's not insane, you just put out a number that seems big, to you, and said "wow this must be impossible!"

Meanwhile, we deploy that all the time.

If nuclear advocates were better at numbers, maybe all the nuclear projects wouldn't be 2x-3x over budget and off schedule, and there would be more nuclear built.

> Meanwhile, we deploy that all the time.

Well, no. It's double our current rate of deployment, and that doubled rate will need to be sustained for decades.

It's not "innumeracy" to say that's going to be hard. Not impossible, but hard, and we haven't yet demonstrated that we're going to do it.

If you price in externalities, as we will have to do in order to slow down climate change, then nuclear is way less expensive than coal or gas.

Also, other countries have historically been able to build plants on reasonable timeframes and budgets.

If there was political will, we could have a large fleet of nuclear plants in under a decade.

Oh absolutely, but that's not going to make the comparison with solar and wind any better. It'd probably make carbon capture for gas a really interesting peaker solution, though!

Considering that during the 1970s nuclear boom a single plant still took a decade to construct, there is simply no way we could possibly have a "large fleet" in under a decade. The only country with the capability of doing so is China - and they are not going to be of much use to Western countries.

China is living proof we could do it- they are currently on track to build a _reactor a year_. We let all our nuclear talent age out and die.

China is not building nearly as much nuclear as they are solar, wind, and storage.

A reactor a year is a pathetic pace! 1GW/year! Even I f the US matched that, we will see most of our nuclear fleet age out before it was replaced.

For contrast, China is expected to be able to produce 6TWh/year of batteries in 2030, ~260 "gigafactories" worth:

https://source.benchmarkminerals.com/article/over-400-gigafa...

China's reluctance to invest as heavily in nuclear should be a clear sign that even with great competence at building nuclear, it's just not that attractive of a tech. It's barely functioning as a hedge, and is probably only being used for its side industries rather than as a serious power source.

If you look at large nations that have actually decarbed their grid, they do it with hydro first and nuclear second.

Show me even a medium sized nation that is predominantly powered by solar/wind and isn't utterly reliant on imports.

Nuclear is absolutely economically viable so long as you build one design several times and don't let regulators insist on crazy handycaps. That's how France, Sweden, Korea, China, and Japan successfully built out. Japan's median built time was 3.8 years during its build out. China builds plants today with $3000/kw capital costs which is as cheap as coal.

Nuclear technology is a proven solution despite all the mental gymnastics people do to insist otherwise.

> Meanwhile, the reactors currently / recently constructed are massively over budget and years behind schedule. Nobody is willing to invest in them unless the government is willing to guarantee a massive subsidy on the electricity produced.

The massive costs are entirely due to excessive regulations. In places and times where the rules are reasonable, plants get built on time and under budget.

Reply on news.ycombinator.com