U.S. Plans $80B Nuclear Power Expansion | Hacker News Reader

U.S. Plans $80B Nuclear Power Expansion

spectrum.ieee.org

37 pointsrbanffy23 comments

I have been a big fan of nuclear for decades. But why now?

Solar with battery storage is about to be so inexpensive and rapid to deploy that perhaps 100% of new capacity should be added this way.

Start with the solar arrays and then add the batteries. They will add to the max immediately. While you are deploying the solar, batteries will improve.

With batteries, you can use solar power even at night.

Lithium batteries are already cheap enough. Sodium is going to be even cheaper and much safer to boot.

There's not a lot of grift in solar and batteries, it's too easy to acquire and deploy. There is literally limitless ability for grift with nuclear.

Look no further than Trump's media corporation merging with a "fusion reactor" company. What do they have in common? Absolutely nothing, but it's an excellent conduit for bribes and fraud, and a way for Trump to send our tax dollars directly into his own pocket!

https://abcnews.go.com/Business/trump-media-announces-6-bill...

> Solar with battery storage is about to be so inexpensive and rapid to deploy that perhaps 100% of new capacity should be added this way.

Exactly because solar is so inexpensive, it means the private sector does not need government help. Utilities do add a lot of solar power themselves, see for example [1]: 52% (32.5 GW) solar, 29% (18.2 GW) battery storage, 12% (7.5 GW) wind for 2025.

[1]https://www.eia.gov/todayinenergy/detail.php?id=64586

I think because of capacity. This race is mainly driven with the AI power demand estimated to increase 10x in the next 5 years. Currently it's 5GW and by 2030 it is expected to be 50GW.

Is this taking efficiency gains into account? I would expect 10X efficiency increase every 3 years given Moore's Law and the hardware appropriate algorithms tendency.

Yes, of course.

Just for comparison, China is spending $20-$30 billion a year on nuclear with 29 power plants under construction (half of world total under construction). While this $80 billion will fund about 5?

Large construction projects of any kind in the US are really expensive.

It does make one wonder if our gdp is as high as we think. Maybe our PPP estimates are overstated due to the reserve currency status.

Anecdotally have some relatives that worked in nuclear industry in USA and have gone on to consulting for the Chinese due to opportunities being too few and far between here in the USA.

> So just how many reactors will $80 billion buy? Assuming an average of $16 billion per AP1000—slightly less than for Vogtle, and allowing for cost reductions from economies of scale and learning-by-doing—the plan would mean five new reactors. That would represent an increase of about 5.7 percent in total U.S. nuclear energy generation capacity, if all the reactors currently in service remain online.

And those are all optimistic assumptions, and allow no margin for delays and cost overruns.

The people making these decisions will be long dead by the time these costs catch up to us, just like Brexit. Most unfortunate if you have US federal tax liability and can’t avoid paying towards the fiat furnace.

Isn’t that Vogtle benchmark including significant cost overruns, delays, etc?

Yes but presumably there has been some innovation for new types of cost overruns and delays.

Based on recent prices for utility-scale solar, I think $80 billion would buy you about 65 GW of solar nameplate capacity, versus 5 GW for 5 AP1000s. Even after accounting for battery capacity and duty cycle and whatnot, this a terrible bargain.

Isn't the nameplate capacity of solar the peak production when the sun is right overhead and the panels are new? Or is it spec'd differently at utility scale?

Say 20% around-the-clock production versus the nameplate capacity, cell degradation, and battery storage doubling the cost, and those costs are starting to look in the same ballpark. Plus having some diversity of sources isn't bad.

Battery backed solar is cheaper than nuclear today, and any nuclear generator will take at least ten years to build, shovel in ground to first kWh to the grid.

Sure, I don't disagree with that argument, and think the destructionists ending the solar subsidies is a treacherous "mistake". But I'd also say that some diversity is worth it in and of itself, rather than relying on one single thing to solve it all, as technologists tend to want to do.

This argument is easy to make when it’s other people’s tax dollars or capital. I don’t mind this position, as long as it isn’t my capital or tax dollars being incinerated on high risk suboptimal energy system investment, based on the evidence and known trajectories. For those who believe it’s worth it, I fully support them contributing their capital or tax dollars to their belief system.

Given how this regime has squandered money and ballooned the debt for things that are outright harmful to our country, I'd say as long as the reactors don't melt down (whether through sheer incompetence or as part of a deliberate plan) we're coming out ahead. I certainly wish we were at a place where fiscal responsibility arguments had relevance, but we just aren't.

Disclaimer: I used to work for an employee-owned nuclear energy services consultancy c. 90's comprised of mostly ex-GE NE and Mitsubishi engineers.

While I'm fine with very scrupulous megaproject nuclear sites who have many layers of checks and security processes, I'm not fine with "emperor's new clothes" throwing out the proverbial baby with the bathwater slapdash, unsecured SMRs in residential and urban areas managed by a startup lacking the deep bench of technical and institutional knowledge. Safety regs are written in blood.

I just don't see the ROI when an equivalent investment in pumped energy storage, hydrothermal, wind, and solar doesn't come with the same baggage that I'm afraid the current regulatory and political environment isn't interested in respecting and protecting a culture of safety.

SMRs designed, owned, and managed by industry titans never got a chance because of public relations in the day, but I think that train has sailed in the current technology and economic environment. (The AI bubble can't burst soon enough, because billionaires are driving inflation of utilities and imposing undue externalities on datacenter neighbors.)

too little too late

Reply on news.ycombinator.com