Hacker News Reader

My comment is not misleading, you're just using outdated data from 2022.

Sure, happy to quibble over units.

The most recent mid-2025 data is from lazard here, it echos exactly what I'm saying.

Website: https://www.lazard.com/research-insights/levelized-cost-of-e...

PDF of report: https://www.lazard.com/media/5tlbhyla/lazards-lcoeplus-june-...

Go to page 8 of that PDF and you will see these ranges for LCOE:

* Solar $38-$78/MWh

* Solar + battery $50-131/MWh

* Gas combined cycle (cheapest fossil fuel) $48-107/MWh

Yes, we are finally at price parity for the technologies.

I didn't disagree that there is price parity for the levelized cost. There is still not price parity for levelized full system cost. If we used wind and solar for 95-100% of generation, the price would be much higher.

My point is not that we can or should replace wind and solar with nuclear. It's that it is far cheaper to use a combination of nuclear, wind, and solar than it is to use 100% wind and solar.

I think it’s quite conceivable that nuclear would be cheaper for a 100% carbon free grid.

But I don’t understand how the combination of nuclear, wind and solar would be low cost. Wouldn’t you effectively have to build out enough nuclear to cover still cloudy days at which point your wind and solar is not very useful? That sounds expensive.

I suspect we won’t end up building much nuclear because we will already have built out so much wind and solar. Nuclear is a poor fit for filling gaps in generation by intermittent renewables because fuel costs are negligible so it costs the same whether you run at 50% or 100% of rated output.

To eliminate carbon emissions entirely we will need some green hydrogen for turning into aviation fuel and as chemical feedstocks. Perhaps the gas backup will eventually burn that.

Green hydrogen is prohibitively expensive and are still way more expensive than using fossil fuels to create hydrogen (called black hydrogen). Burning green hydrogen for electricity when we have yet to make green steel economical viable is not a good idea. Nuclear is still a magnitude cheaper than that.

Green hydrogen has to first prove itself that it can become economical viable. One of the biggest test trials for that is the Swedish initiative, and that one is mostly paid through subsidies and grants. Sadly it isn't looking very great even if the government did decide to continue sending more billions into the project.

I completely agree that green hydrogen is prohibitively expensive at the moment and it currently makes no sense to burn it for electricity generation. But it will likely be necessary in the future if we are to decarbonise aviation fuel, steel making, fertiliser production, etc. What matters at the end of the day is reducing total carbon emissions for the whole economy.

Intermittent renewables and batteries will get us to 80% carbon free electricity generation for more quickly and cheaply than nuclear. While nuclear might make sense in the very narrow use case of 100% carbon free electricity generation, given we also need to decarbonise non-electrical emitters, it will probably reduce more carbon emissions per dollar spent to instead spend that money on even more cheap intermittent renewable generation capacity and use the excess to generate hydrogen. At the point hydrogen based fuels may make sense to use as a buffer for intermittent electricity generation.

Agreed. I misunderstood your comment and got too hot-headed. Sorry about that.

Yes, the 95% renewables is the number we should be shooting for not 100% as that causes battery backup price to explode.

I have been pro-nuclear for a long time, to disappointing results naturally. So, with how well renewables are doing I've really just jumped on this train and seen nuclear as more of a distraction from the critical next 10-20 years given how long it takes to come online.

At the end of the day the grid is only about 30% of the emissions problem (depending where you look).

I may have misinterpreted your original post as saying we should be going full renewables. I think we're basically in agreement about prices. We might just disagree about the percent of energy that should come from nuclear.

I don't see nuclear as a distraction, I see it as a piece of the puzzle. We will always need a source of reliable, uninterrupted power. Whether that comes from natural gas, nuclear, geothermal, hydro, etc. depends on geographical considerations and what tradeoffs we are willing to make in terms of cost and carbon emissions. I'm still optimistic that small modular reactors are going to see success in the coming decades.

Yeah, my opinion on how much should come from nuclear is that current levels (~20%) are enough to fill the rest in with renewables.

I'd love to be France (~50%) but there is so much pushback against the technology due to accidents that happened decades ago with generation II plants (chernobyl + three mile island). We're now building tech for gen III+ plants and there is just almost no appetite to build them, we finished the vogles and now are completely pivoting to SMRs, which is fine.

SMR is probably what makes the most sense even if they're less efficient because until now the nuclear plants have not been very standardized which increases costs.

Why do I think nuclear is a distraction? Because I don't think it's a like-for-like replacement of fossil fuels and this admin knows that. They're willing to invest because it won't disrupt their biggest donors. The time horizon on nuclear is long, and there is a future (I hope) where we have nuclear plants hooked up to carbon capture technology and we pull these gasses out of the atmosphere. But until then what is the cheapest and most efficient path between current emissions and a massive cut in them? Renewables and battery tech (that's currently undergoing very dramatic cost reductions!).

> We will always need a source of reliable, uninterrupted power.

Which can be for example gas turbines running on carbon neutral fuels. Optimizing for lowest possible CAPEX and acceptable OPEX.

The nuclear power lock in are engineer brained imaginary perfect solutions rather than accepting good enough.

> I'm still optimistic that small modular reactors are going to see success in the coming decades.

We’ve been trying to build ”SMR” since the 50s. It has never worked out. The industry likes producing fancy PowerPoint reactors in hopes for handouts and stupid money investment.

When they get far enough and have to present real costs and timelines the projects are shunned and forgotten. Like NuScale and mPower. And the boosters online move to the next juicy SMR project.

When they calculate that Solar + battery would cost $50-131/MWh, how is that number reached? What is the number of charge cycles and over what time span? It seems obvious that the cost of producing, installing and operating a 1MWh system of solar and batteries will cost more than a one time payment of $50-131.

Most of the time when I try to find any data there is the underlying assumption that the charge cycle is a day and night cycle, where the day produce the energy needed during the night, and not a seasonal storage that basically has a single charge cycle per year.

Those are American prices with tariff insanity.

It is much cheaper in the rest of the world. Recent Chinese storage prices are down to ~$50/kWh.

First. $120/MWh for new built nuclear power is cheaper than any modern western reactors. Real costs are ~180-220/MWh when running at 100% 24/7 all year around. As based on Vogtle, FV3, HPC, proposed EPR2s, proposed Polish reactors etc.

The problem with these ”system costs” analyses is that they don’t capture the direct physical incentive structure of our grids.

Why should someone with rooftop solar and a home battery buy $180-220/MWh when they have their own electricity available?

Why should they not sell their excess to the grid cheaper than said nuclear power? It is zero marginal cost after all.

You can call it tragedy of the commons but new built nuclear power simply is unfit for our modern grids.

We need firming for near emergency reserves coming from production with the cheapest possible CAPEX without an outrageous OPEX.

Likely gas turbines running on carbon neutral fuels. But only if we determine that they are needed in the 2030s.

New built nuclear power simply doesn’t even enter the picture in late 2025.

Thanks for that.

A cost model has a lot of independent variables. It can be a weird function of the quantity you want of each technology. Not everything gets cheaper at scale. And you need to be able to manage time-varying demand.

For easy example: a few solar or wind farms cost $X to bring up, but to go large scale you need to also store or transmit the energy, plus keep fallback options. That makes 95% or 100% reliance prohibitive.

There is also the speed of powering on/off. Gas combined cycle turbines are fastest to come online/go offline, followed by hydroelectric (if you have it). Coal and nuclear are at the slow end. You need to have the ability to match total sources and loads at any time.

Just some intuition why total cost is a complex function.