Google's Revived Nuclear Power Plant Gets $1.9 Billion Loan From US Government (techcrunch.com)
- Reference: 0185546240
- News link: https://hardware.slashdot.org/story/26/09/09/0553205/googles-revived-nuclear-power-plant-gets-19-billion-loan-from-us-government
- Source link: https://techcrunch.com/2026/09/08/googles-revived-nuclear-power-plant-gets-1-9b-loan-from-us-government/
> James Danly, Deputy Secretary of Energy, said that the Iowa power plant's restart in 2029 will "drive down electricity costs," though he did not explain how. Just 50 megawatts will be set aside for the local power cooperative, NextEra CEO John Ketchum said during an [3]earnings call (PDF) last year. That capacity would cover 18% of Iowa's demand growth since 2021, the year before ChatGPT was released.
>
> Google is reportedly looking to build up to six data centers near the Duane Arnold Energy Center, which hasn't operated since 2020 when an intense rainstorm damaged the power plant. Rather than repair it, NextEra decided to mothball it. At the time, cheap natural gas was flooding the market, making nuclear power economically unappealing. A lot has changed in the last six years, though. After decades of little growth in demand, the sudden rise of AI coupled with broader electrification of the economy meant that utilities and power providers were suddenly scrambling to find new generating sources of electricity. New data centers are expected to nearly triple the sector's electricity demand by 2035.
[1] https://techcrunch.com/2026/09/08/googles-revived-nuclear-power-plant-gets-1-9b-loan-from-us-government/
[2] https://hardware.slashdot.org/story/25/10/27/2312225/nextera-energy-partners-with-google-to-restart-iowa-nuclear-plant
[3] https://www.investor.nexteraenergy.com/~/media/Files/N/NEE-IR/reports-and-fillings/quarterly-earnings/2025/Q3%202025/Final_Q3%202025%20Script%20vF.pdf
Its very puzzling, isn't it? (Score:4, Insightful)
It will strike everyone as very puzzling. After all, everyone on slashdot and Ars Technica knows that wind and solar are far cheaper than nuclear and just as reliable. So why on earth is Google looking into nukes? Why do they not just build a huge wind farm, throw in a few batteries, and get unlimited cheap power?
Because the wind and the sun are free, after all. But instead they seem to be resorting to a ridiculously expensive nuke, and then on top of that, where you have to pay for the fuel, I mean, someone has to dig the stuff up and refine it. Whereas the wind just blows and the sun just shines, so they obviously are cheaper.
They have accountants and lots of MBAa, so they must see this, so the question is why they are going against what their analyses must show them. If it was coal or gas we would know, they were in thrall to the fossil fuel industry. But is there a nuke lobby, and are they not bigger than it anyway.
Help me understand this! Its a great mystery. Unless I have missed something. But what could that be? I mean, it cannot be that wind and solar plus a few batteries are actually more expensive, could it?
Re: Its very puzzling, isn't it? (Score:5, Insightful)
TFS says "US Government", so part of the mystery might be plain old corruption?
Yep (Score:2, Troll)
So you can't let the public at large know that we could transition to wind and solar and power everything we've got for basically free (before all the dipshits pile in free in the sense that the value generated by the electricity far out strips the cost, if I spend $1 to make $10 that $1 is effectively free)
We can't have that because it'll start the discussion of publicly available electricity instead of private ownership of electricity. And we can't have that because we have to have the Epstein class r
Re: Its very puzzling, isn't it? (Score:1)
Well a new solar plant with batteries costs $1.2 billion for the same power generation. Given restoring an existing nuclear power plant costs slightly more at $1.9 billion, you are in a losing argument here.
Re: (Score:3)
> Well a new solar plant with batteries costs $1.2 billion for the same power generation. Given restoring an existing nuclear power plant costs slightly more at $1.9 billion, you are in a losing argument here.
Time is likely the premium being paid for here. Even being forced to adhere to every safety standard, I'm betting it's a lot faster to knock the dust off an existing nuclear plant than it is to build out a solar solution. Assuming you own enough land to build a solar solution relative to the data center site, which is a critical factor.
Positions are justified from the organizations perspective. Project is sunk as a long-term capital expense spread out over years of depreciation. AI revenue will dispose
Re: (Score:3)
Bring up a decommissioned plant will take years. Don't forget about the thousands of smaller systems that have to be tested and re-certified. Pumps that haven't spun in years. Seals and gaskets, wiring, the control computers. You can't just plop in a raspberry pi and write a pid loop. Everything inside that plant has to be audited and signed off. As the for sale listings like to say ran when parked .
Re: (Score:2)
> Everything inside that plant has to be audited and signed off.
I admire your optimism and wish I could share it. Sadly, to me it borders on naivete.
And even if you're right, corruption often finds a way around safety. The name "Boeing" comes to mind...
Re: Its very puzzling, isn't it? (Score:4, Interesting)
> Well a new solar plant with batteries costs $1.2 billion for the same power generation. Given restoring an existing nuclear power plant costs slightly more at $1.9 billion, you are in a losing argument here.
Um yeah, no. The only way that 1.9 billion is etched in stone is if they just stop refurbishing at 1.9 billion, and switch it on in whatever state it is in.
Nuc plants seldom end up costing what the first projection says. And anyone feeling warm and fuzzy about 1960's technology reactors is fooling themselves.
Re: (Score:1)
Not really. A few rainy days and the economics go sideways. Nuclear power is ideal for a steady state workload like a couple of datacenters. I'm quite sure Google did the math.
Re: (Score:2)
It's partly optics. Concerns about AI pushing up power prices by competing with small business and households.
Here Google (and Microsoft, in the article) are increasing capacity by bringing back mothballed generators with a generous loan.
Re: (Score:3)
> A few rainy days and the economics go sideways.
It does not.
As Google did, you can also do the math. [1]Duane Arnold [wikipedia.org] had a net power output of 581 MW or 14 GWh output per day. To store 14 GWh of electricity right now, you have to pay about US$560 Mio (or $40 per kWh). With the 1.9 billion dollar grant, you could build enough capacity for four days of continuous rain. And differently than a nuclear power plant, this battery capacity could do power arbitrage, charging when electricity is cheap and provide it when prices are high, generating revenue and at t
[1] https://en.wikipedia.org/wiki/Duane_Arnold_Energy_Center
Re: Its very puzzling, isn't it? (Score:2)
Could be that they are treating it more akin to a base load power plant. For solar or wind, the costs go way up, due to having over provisioning and battery storage. This is a 615 MW plant. Even at a $6 billion investment, that's $10/MW for baseload power, so pretty cheap.
nice try at sarcasm ... (Score:2)
... but it doesn't go well with ignorance. gigawatts worth of solar plants are working/commisioned just for ai in the us, google has a good share of them. a simple google search would have saved you the embarrassment.
just in case your's was a serious question, i guess they're all scraping everything they can, money doesn't seem to be an issue. nor does profitability apparently. long and mighty be the bubble! or maybe they're guessing now that a nuclear winter is a possibility?
Re: (Score:3)
> I mean, it cannot be that wind and solar plus a few batteries are actually more expensive, could it?
Nope. The TL;DR version: It would actually be cheaper to build wind/solar and batteries, if not for the fact that the cost of building the plant was covered fifty years ago and they're getting another $1.9B in free money for it.
But of course this was never about you "missing something" - this is about you having an agenda. At least I hope it's about having an agenda, because the economics of the situatio
Re: (Score:2)
Aww, there you go again, doing your research and trotting out the facts. Shame on you! :->
Re: (Score:2)
Also NIMBYs. The plan already exists do they can't object to it, but solar and wind farms, and battery storage arrays would all be new. In fact it being an industrial site already probably makes it easier to get the datacentre built there.
If the Trump admin wasn't so hostile to renewables they would probably be buying European and Chinese offshore wind and batteries, and deploying it already.
Re: (Score:2)
There was a plan to put a solar farm + batteries on that site but they were scrapped for reasons I'm sure you could guess.
=Smidge=
Re: (Score:2, Insightful)
> ... everyone on slashdot and Ars Technica knows that wind and solar are far cheaper than nuclear and just as reliable. So why on earth is Google looking into nukes?
I dunno - might it be the $1.9 billion government loan referred to in the first sentence of TFS?
The current regime is against wind and solar power, largely because Trump is an illiterate idiot who makes a not-terribly-bright third-grader look like a Nobel laureate. Also, welfare recipients - among whom I include the vast majority of tech companies - seldom turn down money. Hence, Google Nukes.
If Google was offered money to revive a dead coal-fired power plant with sufficient capacity to power a major data c
Re: (Score:3)
> Help me understand this! Its a great mystery. Unless I have missed something. But what could that be?
You are missing the fact that they are not building a new power plant, they are restarting an existing power plant. The capital cost is already paid for.
In essence, it's like my question "should I buy a new car, or just get the one I have fixed?" New cars may get much better gas mileage, but the car I have is already paid for.
Re: (Score:2)
Maybe that isn't what their analysis shows? I mean, they actually did one and based it on far more information than either of us have.
So Google paid a fraction of the cost (Score:3)
Risk costs money, this is a subsidy for Google.
As expected.
Re: (Score:3)
I'm not sure about the US, but in the rest of the world, every form of power get subsidies, one way or another.
Better to be subsiding nuclear (and batteries for solar) than coal and gas, I suppose.
Whats this I hear about Trump paying out billions to NOT build wind farms though?
nah - never mind, I'm not sure I can take any more of that stuff.
what are the odds (Score:3)
What are the odds of this loan actually being paid back? NextEra (and Google) are certainly large enough to have the means to repay, but will they actually, and will the Administration force them to? What is the penalty if, after burning through all this money, no electrical power is actually produced?
This, in my opinion, is more like a grant - a giveaway. What is the public benefit for the risk being assumed by the taxpayer? Given the current government's [1]love of ownership stakes [cato.org] (despite decrying "socialism" with a pearl-clutching gasp), why not provide the money in return for stock? Or make it a corporate bond such that, if NextEra defaults on the terms, the rest of the corporate bond market can extract their own form of punishment?
And furthermore, if this turns out to be a flop, will the politicians in favor of it get raked over the coals the way Obama was over Solyndra? Heck, that was only $500 million of lost taxpayer money!
[1] https://www.cato.org/blog/government-ownership-stakes-companies-becoming-routine-under-trump
Re:If you are made about (Score:5, Insightful)
100% bullshit.
My goto example is the Barakah plant over in the UAE. [1]https://en.wikipedia.org/wiki/... [wikipedia.org]
You think there were leftists and environmental protesters slowing construction over there? Plus it was joint effort with South Korea.
As with every nuclear plant it was years late and ended up costing $32 billion instead of the projected $20 billion.
[1] https://en.wikipedia.org/wiki/Barakah_nuclear_power_plant
Re: (Score:2)
Because when it comes to Nuclear, everyone has their hand out wanting a piece of the pie. Many hands are self-inserted middlemen, but many more are hands out asking to be paid to go away.
The Story Doesn't Say (Score:2)
The story doesn't say how old this reactor is, why it was decommissioned, how much the government or consumers paid for the stupidly expensive shutdown process and likely ongoing maintenance, how much it will cost the government or consumers to shut it down again in the not too distant future... Decommissioning takes 10 - 15 years. Expensive, union labor.
I'm all in favor of nuclear power. But there is a considerable cost at the end of the line that always seems to get forgotten about until suddenly the util
Can't Afford Plant--then Can't Afford Cleanup (Score:2)
If there is an accident, Google could not afford to clean it up. It would seem that google could not afford the decommission.
Wikipedia: "A banana republic is a country with an economy of state capitalism, where the country is operated as a private commercial enterprise for the exclusive profit of the ruling class. Typically, a banana republic has a society of extremely stratified social classes, usually a large impoverished working class and a ruling class plutocracy, composed of the business, political,
Re: (Score:2)
Obviously, connect the dots. : )
What could possibly go right? (Score:3)
I realize they're not just going to walk in and start throwing switches, but reopening this plant that's been shuttered all this time and doing it correctly has got to be seriously expensive, so what are the odds on whether it will be done correctly?
Re: (Score:3)
> I realize they're not just going to walk in and start throwing switches, but reopening this plant that's been shuttered all this time and doing it correctly has got to be seriously expensive, so what are the odds on whether it will be done correctly?
Pretty good, unless Google thinks it can withstand that kind of explosive scrutiny?
Chernobyl didn't exactly blow over for the USSR. And that Too Big To Fail gimmick doesn't really work out when your bailout revenue glows in the dark with a half-life of 10 business days or less.
Re:What could possibly go right? (Score:4, Informative)
They won't be doing it themselves. They'll have to contract out. Most nuclear plants use Bechtel, Sargent and Lundy and like a dozen other NRC certified contractors. People work their asses off to get into those programs, but you can easily get stuck in a situation where you're moving every 2~4 years to a new nuclear plant. People who want families often try to get FTE positions.
The big danger is draining people from TVA or other power providers. Data center electricians are pulling in $200k/year right now! A loan like this could draw out people from other plants for that big payout. But in reality, unless they plant to build new plants and use this to train up new nuclear workers, they're just going to hurt the existing energy industry.
But that's the big issue: no one can build a new reactor today. And no, not because of NRC regulations. America literally has no steel plants large enough to create a nuclear pressure vessel. These things are massive, and historically had to be transported by river or have specialized roads put in to move them. The last steel plant that could make these closed decades ago. The smaller reactors people are talking about are totally non-viable. Nuclear works because of the scale. Each pressure vessel and reactor core can provide 1Gw ~ 1.5Gw of power. All three TVA nuclear plants (max of 7 reactors between SQN, WBN and BFN plants .. I think only 6 are operational) produce between 6Gw ~ 7Gw of power. Keep that in mind when you hear about data centers needing like 20Gw. That's a fucking insane amount of energy!
Re: (Score:3)
> But that's the big issue: no one can build a new reactor today. And no, not because of NRC regulations. America literally has no steel plants large enough to create a nuclear pressure vessel. These things are massive, and historically had to be transported by river or have specialized roads put in to move them. The last steel plant that could make these closed decades ago.
"For very large 'Generation III+' reactors, production of the pressure vessel requires, or is best undertaken by, forging presses of about 140-150 MN (14-15,000 tonnes) capacity which accept hot steel ingots of 500-600 tonnes. These are not common, and individual large presses do not have high throughput – about four pressure vessels per year appears to be common at present, fitted in with other work, though the potential is greater than this.
Reactor vendors prefer large forgings to be integral, a
Re: (Score:3)
> I realize they're not just going to walk in and start throwing switches, but reopening this plant that's been shuttered all this time and doing it correctly has got to be seriously expensive, so what are the odds on whether it will be done correctly?
A 1960's technology plant that was decommissioned in 2020, a little earlier than planned since it was damaged by a derecho, and was considered too expensive to repair.
Plus how will this be operated? By a company that abandons projects after a short time? Will it be operated with the standard profit motive? Tech bros needing to make more profit each quarter. Bean counters working hard to squeeze out profit. Gonna be interesting, apparently a privately owned old reactor working with corporate oversight.
Re: (Score:2)
> reopening this plant that's been shuttered all this time and doing it correctly has got to be seriously expensive
The article says "a $1.9 billion loan to NextEra Energy to restart Iowa's shuttered Duane Arnold nuclear plant", so, yes, I think you could call 1.9 billion dollars "expensive".