Big Tech Has Cozied Up To Nuclear Energy (theverge.com) 77
Tech giants Amazon and Microsoft have inked major deals with U.S. nuclear power plants to fuel their energy-hungry data centers, marking a shift in the industry's power sourcing strategy. The move comes as AI-driven facilities strain companies' climate goals, pushing them towards carbon-free electricity sources.
Microsoft plans to revive the shuttered Three Mile Island plant by 2028, while Amazon secured power from Pennsylvania's Susquehanna Nuclear facility in a $650 million deal. Google is also exploring nuclear options, including small modular reactors still under development. This trend could potentially triple U.S. nuclear capacity by 2050, according to a Department of Energy report.
Microsoft plans to revive the shuttered Three Mile Island plant by 2028, while Amazon secured power from Pennsylvania's Susquehanna Nuclear facility in a $650 million deal. Google is also exploring nuclear options, including small modular reactors still under development. This trend could potentially triple U.S. nuclear capacity by 2050, according to a Department of Energy report.
Issues with nukes (Score:5, Interesting)
Data centers that aren't backing real-time applications could throttle if demand power was needed elsewhere. And nuke plants are slow to adjust.
It seems to me this is a potentially good marriage - let the data centers throttle themselves based on current power grid load, keeping the nuke plant at a steady output.
Re: (Score:3)
Re:Issues with nukes (Score:5, Insightful)
It seems to me this is a potentially good marriage - let the data centers throttle themselves based on current power grid load, keeping the nuke plant at a steady output.
In principle, I agree with you. In practice, Big Tech companies are hardly models of restraint, and they're also empire builders in the worst sense. I predict that eventually the larger grid will take a back seat to data centre "needs".
Of course, refining and rolling out better grid storage solutions could make that point moot. But then, with enough storage to handle fluctuations on both the supply side and the load side, why go nuclear at all? Wind and solar should be enough.
I'd rather see this whole out-of-control data centre gold rush slowed down. It represents a huge drain on resources, in the name of what's essentially "Ponzi scheme 2.0". It funnels more wealth up to the parasite class faster, and it worsens income inequality.
Re: (Score:3)
It seems to me this is a potentially good marriage - let the data centers throttle themselves based on current power grid load, keeping the nuke plant at a steady output.
In principle, I agree with you. In practice, Big Tech companies are hardly models of restraint, and they're also empire builders in the worst sense. I predict that eventually the larger grid will take a back seat to data centre "needs".
I can see the baby AIs hovering around their entertainment center, getting ready to sit down to a nice sitcom about how humans argued with one another during this time period:
Big tech: We need more power.
Government: We need power for the people.
Big tech: WE NEED MORE POWER!
Government: For to why?
Big Tech: For to replace all humans.
Government: Kinda sounds bad-ish. Though that does get rid of voting. Hmm.
Big tech: Also, for to future profits!
Government: Sold. Would we like to burn bodies for fuel?
AI
Re: (Score:2)
Of course, refining and rolling out better grid storage solutions could make that point moot.
Now there's a deal for data centers and power providers. A huge battery system that works two ways - backup power if they get disconnected from the grid, but providing grid stability back to the power company the rest of the time.
Re: (Score:2)
Good call. I hadn't taken my thinking to the point of the batteries being located at the server farms, but you're right - it seems like a natural and mutually advantageous fit.
Re: (Score:2)
Big tech an nuclear are a perfect match. The tech industry doesn't care about the risk or the environmental consequences, and always tries to push the cost onto someone else. Nuclear is desperate for investment.
Re: (Score:2)
But then, with enough storage to handle fluctuations on both the supply side and the load side, why go nuclear at all? Wind and solar should be enough.
They bought expensive hardware to train the models. To get return on the investment they need to run it all the time at full capacity. Nuclear is a better choice in such a case. They do not have any hassle with intermittent renewables and (battery) storage. They can colocate data centers with nuclear power plants and they do not have the cost associated with the distribution grid. This way they can lower the electricity cost which is their biggest recurring cost item.
Re: (Score:3)
No one will be doing load shedding without being paid for it. Many data centres are not backing real-time applications, they are backing profitable applications. There's little profit in not getting your next 2 trillion parameter LLM to market. A datacentre which has the ability to drastically throttle away on demand is a datacentre which was over provisioned.
Re: Issues with nukes (Score:2)
Re: (Score:2)
Re: (Score:1)
Simpsons (Score:1)
As long as they keep them out of Ukraine. At least then we won't run the risk of another Chernobyl, just another mere Three Mile Island.
Re:Simpsons (Score:4, Funny)
Re: (Score:2)
Did nobody get The Simpsons reference? King Sized Homer?
Re: Simpsons (Score:2)
TMI was a wet squib.
Nuclear is our only hope (Score:2, Insightful)
Anyone opposed to nuclear power is a climate change denier, whether they admit it or not.
Re:Nuclear is our only hope (Score:4, Insightful)
Anyone opposed to nuclear power is a climate change denier, whether they admit it or not.
Nice gaslighting. But no, we can very much realise climate change is a problem while at the same time reflecting on the fact that any investment you make now will not bring a single nuclear reactor online in any meaningful timeframe to avert climate change.
You're standing on train tracks looking at an oncoming train knowing it will hit you. Do you take the nuclear approach: Lobby the government to enact a 20 year project to move the train tracks away from where you're standing? Or do you do something else in the 30 seconds you have left to live.
Re: (Score:2)
You do realize that many "environmental" organizations were taking money from oil industry to attack nuclear power [environmen...ogress.org]...
Sierra Club
Has received $136 million from interests that would benefit from the closure of nuclear plants
Natural Resources Defense Council (NRDC)
Has received at least $70 million from oil and gas interests that would benefit from the closure of nuclear plants
Environmental Defense Fund (EDF)
Has received at least $60 million from oil, gas, and renewables investors who would benefit from EDF's a
Re: (Score:2)
Of course I realise that. And fuck those greenies and the fossil industry for not allowing us to invest in nuclear back when it *was* a meaningful and viable solution.
But that time is gone, the time for this discussion was 30 years ago so that we have a steady stream of nuclear baseload coming online now to drop our emissions. Not time to have the discussion now where we'll do nothing but create copious amounts of CO2 emissions running cement factories for the next 20 years while achieving nothing. We need
Re: (Score:3)
Nice gaslighting. But no, we can very much realise climate change is a problem while at the same time reflecting on the fact that any investment you make now will not bring a single nuclear reactor online in any meaningful timeframe to avert climate change.
Climate change doesn't happen overnight and there is no "meaningful timeframe to avert" it. If we stopped emitting all CO2 right now the planet would continue warming for at least a few hundred years while all the excess carbon is absorbed into the already saturated carbon sinks. We could certainly build nuclear plants fast enough that it would matter if we actually do it instead of just debate and study the feasibility of doing it. China is building AP-1000 reactors in 5-7 years, there's no reason other
Re: (Score:2)
Climate change doesn't happen overnight
Nuclear power doesn't happen over 7300 nights. Whatever way you cut it, kicking the can down the road isn't a viable solution.
Re: (Score:2)
But no, we can very much realise climate change is a problem while at the same time reflecting on the fact that any investment you make now will not bring a single nuclear reactor online in any meaningful timeframe to avert climate change.
Only because of bureaucratic procedures that have been purposefully honed to created such a situation. It wasn't always like this, but the 'save the planet' types fought tooth and nail to create this situation.
Nuclear? Maybe. Fission? Likely not. (Score:2)
Fission is steam-age technology. Quite literally. 23% efficiency, cooling problems, flaky setups, error prone, overly complex, outlandishly expensive, questionable runtimes/lifecycles and with a non-trivial 50 000 year waste-problem attached.
There is a reason why renewables are catching on faster than any Nuke-Fission fan (or anybody else) would've suspected are plainly obvious in hindsight: No regulations, no waste at geolithic timescales, no obscene amounts of cash and resources and no epic dangers. Just
Re: (Score:2)
Re: (Score:2)
PS there is no such thing as a gammavolt.
A gammavoltaic cell is a device that converts gamma radiation to electricity. The GP mentioned it because something like that might potentially be an alternative way to get electric power out of a nuclear reaction. Obviously it would take a huge leap forward in technology to use that to replace steam power in nuclear reactors.
In any case, what parallel universe do you live in where you can't make steam with renewables? That's equivalent to saying you can't make steam with electricity. Which is obviously [acmeprod.com] unt
Not just big tech... (Score:4, Interesting)
And of course... (Score:5, Interesting)
Cue social media and algorithm machines suddenly being pro-nuclear.
Most people don't realize how easily they are to influence, and how their opinions are really the ones carried by their rich overlords. Our rich overlords want nuclear? Watch as suddenly everyone will become pro-nuclear. The sponsored research about how nuclear is good for the environment, the economists talking about the potential savings... Our lives will improve so much! The rich billionaires know what's best, always have, always will.
And in case you come after me, I have always been in favor in nuclear energy, I am just tired of how easily the rich dictate what opinion is the proper one to have.
Re: And of course... (Score:1, Troll)
They are, after all, the ones who sold the idea of organic being good to the unwashed masses despite how incredibly wasteful it is. Food was getting too cheap. Food with higher profit margins is better for you. Just slap meaningless words like "superfood" and "natural" on it and they'll literally eat it up.
Don't worry, their stooges like drinkypoo will come over with fresh from the farm propaganda like falsely claiming that regenerative agriculture is exclusive to organic superfoods. He'll then talk about h
Re: (Score:2)
If the outcome is truck-portable, inexpensively produced stable nuclear power production, it's worth the spin. The technology exists to make nuclear power clean, safe, and inexpensive but only at scale. Regulations have been holding back progress for decades.
As two-faced as a quarter. (Score:3)
Tech Bros spent the past 30 years all-in for "green energy."
Suddenly they need much much more, and fast -- to fuel their new money factory, AI -- so now they're cozying up to their prior enemy.
The duplicity is stunning, and revealing of their character.
I hope they energy sector remembers the absolute hatchet jobs tech bros like Gates and others did on the energy industry, and tells Silly-con Valley to go fuck itself.
Bubble pops in 3...2...1...0.5... (Score:1)
Don't worry, there will be an AI bubble poppage soon that will pause AI server farm building for about 20 years until they figure out how to make it profitable. (It won't go away, but overcapacity will take a while to adjust.) So we got about 20 more years until the Apocalypse Machines are cranked back up.
Roughly 3/4 of investment-inducing "breakthroughs" have a bubble hiccup before they are forced to be efficient.
Re:As two-faced as a quarter. (Score:4, Insightful)
Man, my karma is still in the dumpster from when I flamed out on some awful people 10 years ago, but I'm gonna go right back in the trash by posting some sincerity.
Nuclear is green energy, by virtually any measurable standard(besides a vague sense of how "natural" it is). By amount of radiation put into the environment, it's clean. By CO2 generated, it's very clean. By smog and soot, it's incredibly clean. What it isn't is cheap, but as long as that money is coming from easily duped AI hype investors anyways, who cares?
In most respects solar is purely better, but for base load, nuclear is a cleaner answer than batteries.
Re:As two-faced as a quarter. (Score:5, Informative)
Errr no, that's revisionist. Tech Bros are still all in for green energy (not for 30 years though, largely no one gave a fuck about this even 20 years ago), if they weren't they wouldn't be inking deals with nuclear plants, they'd be in for cheaper power.
And Tech Bros have largely been proponents of nuclear energy. The only thing which has changed now is that there's enough money to be made to buy more expensive power from nuclear plants that have already been built and are standing idle because they aren't viable without a customer willing to pay a premium.
I hope they energy sector remembers the absolute hatchet jobs tech bros like Gates and others did on the energy industry, and tells Silly-con Valley to go fuck itself.
What the actual fuck are you talking about. Bill Gates literally funded a nuclear power startup (Terrapower) just shy of 20 years ago, when not only very few people were giving any fucks about CO2 emissions, but when the nuclear industry was still big-bad-number-1 in the energy sector even before people were talking about shutting down coal.
Re: (Score:1)
they wouldn't be inking deals with nuclear plants, they'd be in for cheaper power.
Nuclear power IS cheaper energy, so I'm not sure what you're on about. Perhaps not at the market rate due to the prohibitive regulations and moratorium against modernization, but they're not going to be consuming market rate energy.
Re: (Score:3)
Nuclear power IS cheaper energy, so I'm not sure what you're on about.
Don't ask me. Ask the guys at Three Mile Island, the guys who shut down the reactor in 2019 for purely economic reasons - i.e. they fundamentally couldn't produce power at a low enough cost to sell and were running at a loss. The only reason TMI is restarting is because Microsoft signed a PPA with them well above market rates for electricity.
Perhaps not at the market rate due to the prohibitive regulations and moratorium against modernization
I genuinely don't understand what you're saying. What is "cheaper energy" if not something that is cheap compared to the market rate. The market is the market, regulate
Re: As two-faced as a quarter. (Score:2)
The energy markets in the US are warped to the point of being broken, something that independent monitoring groups have identified for over a decade. The mechanisms for capacity compensation are radically undervalued, leading to an overvaluing of intermittent resources. MISO in particular has had tastes of this, such as when their capacity market prices spiked by a factor of 100 in a single year because they effectively didn't compensate for capacity until they had too little.
What you're seeing with these d
Re: (Score:2)
What you're seeing with these deals isn't altruistic companies cutting deals above market rate out of the goodness of their heart. It's companies making a declaration that they believe these resources are fundamentally undervalued.
No. It's companies engaging in quick knee jerk reactions to resolve their power problems and paying above the market rate in the process. These tech companies are only in it right now as a panic movement. It is faster to startup a stopped base load power source while padding green credentials than to attempt to kick start a whole new power generation project complete with storage.
This is especially difficult as the bottom is falling out of the carbon credits processes and the world is increasingly skeptical
Re: (Score:2)
New reactors don't cost $15bn. New, old reactor designs with existing regulatory requirements in place do.
Been kind of hard to compete against all the other power sources, which have significant subsidization for 'green' initiatives (yes, even coal and NG).
I'm talking about modern designs, like the Zeus 20MW nuclear reactor, which ships nicely in a connex container. It's small enough to be mass produced. Mass production means efficiencies of scale and fabrication binning. Even without mass production, they'
Re: (Score:2)
Sorry but you live in a fantasy land. New reactors absolutely do cost that. They typically get funded at a cost estimate of $6bn, but not even the Chinese can build them for that price without insane overruns. In the west we look to typical overrun costs of 2-3x. But again, don't take my word for it. Look at EDF and what they just built at Flamanville. They may actually go into operation in 2026. But let's face it they've been saying "next year" for over a decade at this point.
Been kind of hard to compete against all the other power sources, which have significant subsidization for 'green' initiatives (yes, even coal and NG).
The nuclear industry is heavil
AI zombie apocalypse almost given (Score:3)
There's nothing that gives me more nightmares than Larry Ellison, Zuckerberg, or Musk owning a nuke plant.
Re: (Score:1)
Eye mate!
What if each one owns one?
Or two even?
Or more?
Re: (Score:1)
Then we get it up the wazoo, down the wazoo, and through the wazoo.
Re: (Score:2)
Good (Score:2, Troll)
We need nuclear energy if we want to mitigate climate change. The tech companies have enough capital and political influence to get new nuclear built.
If we do not pursue nuclear we will have a situation such as Germany. They spent 700 billion euros on renewables and failed to deep decarbonize their grid.
Re: (Score:2, Flamebait)
Nuclear is a good way of spending lots of money while not having any positive impact on carbon emissions for 2 decades. It is literally how you get to your incorrect view of the Germany situation: spend lots of money and achieve nothing.
They spent 700 billion euros on renewables and failed to deep decarbonize their grid.
Only a fraction of the 700 billion euros was spent on the grid. The 700 billion euro figure is the sum total of all green energy programs spent over the past 2 decades, this includes building quite a few gas plants and investing in new gas pipelines among other things.
As for
Re: (Score:2)
Germans oppose nuclear energy because they still believe hitlers propaganda about nuclear science being Jewish science.
And 400 g CO2 per kWh is failing!
Always the same. (Score:1)
Short term gains (well, not that short, but relatively speaking..)
Are these companies going to shell out for long-term nuclear waste storage, decommissioning and clean-up, and if there's any sort of accident?
They should be tied into that.
But nope.
Always on us. More corporate welfare coming down the pipe. They reap the benefits, and we clean up after them, on our own dime.
And for the people saying 'we need nuclear', this isn't for most of us, this is for them.
Re: (Score:1, Troll)
Long-term nuclear waste storage is such a bullshit excuse to oppose nuclear energy. Guess what DF, used fuel (aka nuclear waste from nuclear power plants) has a world wide kill county of zero. Zero! Meanwhile the waste from fossil fuels and biofuels kill 8.7 million a year.
Also decommissioning/cleanup costs are included in the price. As for accidents all NPP's pay into a government ran insurance fund that has never been tapped(since nuclear has been historically safe). There is almost 100 billion in
Re: (Score:2)
You have never heard of the Price-Anderson act I guess. Nuclear companies don't pay for insurance, it is insured by the federal government. Taxpayers are on the hook. The real problem nuclear power faces is that it has been extremely expensive and unreliable with lots of cost overruns and lengthy delays. Construction standards on many of the plants have been awful.
Safety is certainly an issue because we are lousy at assessing low probability catastrophic risk. Witness earthquake preparedness in the Pacific
Re: (Score:2)
Yes, I have heard of it. And yes nuclear power plants pay into a general fund that is nearly 100 billion in size.
You seem to think Chernobyl never happened or, more likely, it can't happen here.
No, It can't happen here. Different reactors, different physics. It couldn't happen if someone intentionally attempted to blow the reactor. The worst possible case would be TMI, but that isn't even likely.
Next-generation plants are even better. They can't meltdown at all. Google Expermental Breeder Reactor II and the IFR.
The builders of nuclear power plants have the same attitude.
No, this is measurable physics. Also known as Science!
Re: (Score:2)
No, It can't happen here. Different reactors, different physics.
I think something unexpected happened that smart people didn't think was likely. You think that nothing unexpected can happen that smart people like you didn't anticipate. Just like the overconfident, smart people at Chernobyl. That much is the same.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: Always the same. (Score:2)
Well, at least in the US, yes, nuke plant operators have been on the hook for containment, clean up, etc. It's up to We the People to hold them to that by ensuring proper regulations and enforcement.
We'd have a nice long-term storage facility if the idiot NIMBYs in Nevada hadn't somehow successfully shit on Yucca Mountain. Maybe we can get that unfucked.
Limited Liability (Score:2)
No, nuclear operators are not generally responsible for those when they are part of the regulated grid. They are guaranteed a return on their investment and their regulated rates include those costs. So ratepayers are also going to be responsible for any additional costs.Which makes it a very expensive power source.
Of course Microsoft is not part of a regulated market. So if there is a three mile island type incident, three mile island's owners will be responsible for the cost. They will likely go bankrupt
Re: Limited Liability (Score:2)
Not sure what country you are talking about, but in the US I am not aware of nuke plant operators that are guaranteed a profit and are not responsible for operational and decommissioning costs. Examples?
Re: (Score:2)
Some absolute figures (Score:4, Insightful)
As of 2022 American nuclear power stations produced 772 TWh [world-nuclear.org], or about 19%, despite a small drop since 1990-ies [statista.com].
Tripling that would be quite impressive indeed...
They think anyone cares ... (Score:2)
This is almost purely about meeting fantasy emission targets which they fool themselves anyone relevant cares about (sure there are some green investment funds but it's not like they are startups which need to care about that).
The restart brings some extra nuclear power back online, but for the most part it only displaces emissions. They are not lining up to take the 30 year risks and prefund Nuclear all by their lonesome. For new nuclear to get build, public money has to take all the risk and private money
Who would pay for a Nuclear New Reactor? (Score:2)
I would like to know is who would be silly enough to pay for the building of A New Nuclear Reactor even SMR 's, which have yet to put build for Commercial use, will be too expensive when compared to Renewables with Battery firming.
Nuclear Power only makes sense if you don't have to pay for the cost building the of the Reactors hence why Bill Gates would be looking at Three mile Island's just retired Nuclear Reactor.
But really, bitcoin bros want cheap power (Score:2)