One of the two nuclear reactors at Three Mile Island, the Pennsylvania site of a notorious partial meltdown 45 years ago, could be brought back online in the coming years to provide power to a new Microsoft artificial intelligence data center, officials said Friday.
Constellation Energy, the Baltimore-based provider that spun off Exelon two years ago, has signed a 20-year power purchasing agreement with the tech giant to draw electricity generated at the plant along the Susquehanna River outside Harrisburg and about 85 miles west of Philadelphia.
Pending regulatory approvals, the newly created Crane Clean Energy Center would become the first nuclear plant in the United States to return to service after being shut down.
The $1.6 billion project will restart Three Mile Island Unit 1, which stopped generating power five years ago because it could not compete with cheaper energy being produced by Pennsylvania's natural gas industry. The reactor can be run independently from Unit 2, where the plant's partial meltdown occurred resulting in the worst nuclear accident in U.S. history on March 28, 1979. That reactor is still in the process of being decommissioned by owner Energy Solutions.
"Before it was prematurely shuttered due to poor economics, this plant was among the safest and most reliable nuclear plants on the grid, and we look forward to bringing it back with a new name and a renewed mission to serve as an economic engine for Pennsylvania," Joe Dominguez, president and CEO of Constellation, said in a statement.
[...] In the race to develop artificial intelligence applications, tech companies are scrambling to build data centers, which require enormous amounts of electricity to operate. Such facilities are forecast to make up a growing share of the nation's electricity use in the years to come, prompting companies to look at tapping into existing infrastructure to help meet their needs.
Nuclear power is being touted as a cost-effective solution for these data centers that also limits reliance on carbon-producing power sources. Building and directly connecting data centers to nuclear plants is known as co-location, a strategy that industry leaders favor because it's cheaper and faster to do. Proponents also claim it reduces stress on the transmission grids.
During the years the 837-megawatt unit operated at Three Mile Island, the reactor powered about 830,000 homes and businesses. Constellation officials did not say how much of the reactor's power-producing capacity would be dedicated to powering Microsoft's AI data center, but it's not uncommon for such facilities to have energy demands of 1,000 megawatts – or 1 gigawatt.
An economic impact study commissioned by the Pennsylvania Building & Construction Trades Council estimates the restart of Three Mile Island would create 3,400 jobs directly and indirectly related to the plant and generate about $3 billion in state and federal tax revenue.
[...] When Constellation signaled interest in restarting Three Mile Island in July, doubts surfaced about the technical feasibility of the project. Not only would it be the first of its kind, but it will have to be accomplished next to another reactor whose clean-up and decommissioning is expected to continue through 2078.
The site also remains politically contentious due to the lasting memory of the 1979 accident, which displaced surrounding communities and left a legacy of fear over whether the radiation released contributed to increased cancer rates in the vicinity. About 2 million people were exposed to radioactive fallout as a result of the meltdown.
Public health researchers from Temple, Penn State and the University of Pittsburgh published a report last year finding that long-term studies into the impact of the meltdown were limited by research flaws.
Despite these concerns, Constellation cited a statewide poll showing strong support for restarting Three Mile Island. The poll conducted by Susquehanna Polling & Research found Pennsylvania residents approve of restarting Three Mile Island by more than 2-1 and 70% favor the ongoing use of nuclear energy in the state.
« NHS Scientists Find New Blood Group Solving 50-Year Mystery | Plenty of Ups-and-Downs Are Key to a Great Story »
Related Stories
On Monday, OpenAI CEO Sam Altman outlined his vision for an AI-driven future of tech progress and global prosperity in a new personal blog post titled "The Intelligence Age." The essay paints a picture of human advancement accelerated by AI, with Altman suggesting that superintelligent AI could emerge within the next decade.
"It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I'm confident we'll get there," he wrote.
OpenAI's current goal is to create AGI (artificial general intelligence), which is a term for hypothetical technology that could match human intelligence in performing many tasks without the need for specific training. By contrast, superintelligence surpasses AGI, and it could be seen as a hypothetical level of machine intelligence that can dramatically outperform humans at any intellectual task, perhaps even to an unfathomable degree.
[...]
Despite the criticism, it's notable when the CEO of what is probably the defining AI company of the moment makes a broad prediction about future capabilities—even if that means he's perpetually trying to raise money. Building infrastructure to power AI services is foremost on many tech CEOs' minds these days."If we want to put AI into the hands of as many people as possible," Altman writes in his essay, "we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips). If we don't build enough infrastructure, AI will be a very limited resource that wars get fought over and that becomes mostly a tool for rich people."
[...]
While enthusiastic about AI's potential, Altman urges caution, too, but vaguely. He writes, "We need to act wisely but with conviction. The dawn of the Intelligence Age is a momentous development with very complex and extremely high-stakes challenges. It will not be an entirely positive story, but the upside is so tremendous that we owe it to ourselves, and the future, to figure out how to navigate the risks in front of us."
[...]
"Many of the jobs we do today would have looked like trifling wastes of time to people a few hundred years ago, but nobody is looking back at the past, wishing they were a lamplighter," he wrote. "If a lamplighter could see the world today, he would think the prosperity all around him was unimaginable. And if we could fast-forward a hundred years from today, the prosperity all around us would feel just as unimaginable."
Related Stories on Soylent News:
Plan Would Power New Microsoft AI Data Center From Pa.'s Three Mile Island 'Unit 1' Nuclear Reactor - 20240921
Artificial Intelligence 'Godfather' on AI Possibly Wiping Out Humanity: 'It's Not Inconceivable' - 20230329
Microsoft Research Paper Claims Sparks of Artificial Intelligence in GPT-4 - 20230327
John Carmack's 'Different Path' to Artificial General Intelligence - 20230213
(Score: 5, Insightful) by Rosco P. Coltrane on Monday September 23, @01:41AM (4 children)
An incompetent nuclear operator powering an incompetent software maker. Seems like a perfect match.
(Score: 4, Funny) by KritonK on Monday September 23, @05:17AM (3 children)
Nah, Microsoft will upgrade the reactor's controlling software according to their latest standards. What could possibly go wrong?
(Score: 2) by turgid on Monday September 23, @02:05PM
This sounds like a job for MinWin!
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].
(Score: 2) by Tork on Monday September 23, @04:08PM (1 child)
🏳️🌈 Proud Ally 🏳️🌈
(Score: 3, Informative) by DannyB on Monday September 23, @05:47PM
Outlook Excess indeed!
Need I remind you . . .
* I Love You (email virus that rampantly spread because Outlook would happily execute incoming email attachment scripts for your convenience)
* Windows for Warships [schneier.com]
*
CloudStrikeClownStrikeDon't put a mindless tool of corporations in the white house; vote ChatGPT for 2024!
(Score: 2) by Rosco P. Coltrane on Monday September 23, @01:49AM (3 children)
Yeah, the half that didn't melt down was really reliable.
(Score: 2, Insightful) by Anonymous Coward on Monday September 23, @02:51AM
It's ludicrous that someone could maybe possibly learn from mistakes.
Go study Chernobyl accident. There was much they didn't anticipate.
(Score: 2, Interesting) by VLM on Monday September 23, @11:51AM (1 child)
Yeah, exactly. A liberal arts guy name Perrow ended up writing a book about the (then new) "normal accident" theory which was kind of an attempt to tell non-math people what they should think about engineering statistics and the mathematics of risk.
It's kind of "meh", I remember having to read it as part of the common engineering curriculum and maybe I still have it on some bookshelf.
It got ripped on at publication and after, because the book was written because of TMI and more or less about TMI and inspired by TMI but the marketing people got a hold of it shortly after the Challenger space shuttle exploded, and as marketing people they decided a cover picture of the then recent exploding space shuttle would sell more books than a picture of a nuke plant, so ... and that resulted in, IIRC, more people discussing predatory marketing and profiteering off seven dead astronauts than about actual nuclear risks or any of the topics in the book. Its, um, memorable, just look for the sociology book with the space shuttle explosion cloud on the cover. But the contents of the book are entirely about TMI IIRC LOL.
Even a decade after it was published people were talking about it and it was an assigned book at school and I still remember it so it made at least a minimal impression, I guess.
It's a strange book because it's trying to talk about statistical theories without using numbers or math or logical reasoning so as to sell the book and its concepts to the liberal arts peeps. Sometimes "popular science" treatments just don't work. OTOH it gets cited a lot because nobody in liberal arts really tried writing a statistical risk book until the guy wrote it. If you ignore everything else about it, all the controversy, its a weird yet strangely intriguing book; imagine trying to write a book about differential equations without using numbers or graphs, or trying to write a programming book without a single line of source code. Its like code golf but for sociologists and engineers.
Not to rip on it too much but another guy in my class described it similar to Shakespeare tried to write a book about actuarial science and everything turned out exactly like you'd expect. I'm not as harsh on it. I wouldn't buy it but it might be worth checking out from a public library, maybe.
(Score: 0) by Anonymous Coward on Wednesday September 25, @12:40AM
So what was the point of you trying to give this guy's work more undeserved mind share?
(Score: 1, Interesting) by Anonymous Coward on Monday September 23, @03:12AM
" tech companies are scrambling to build data centers, which require enormous amounts of electricity to operate. Such facilities are forecast to make up a growing share of the nation's electricity use in the years to come"
We've taken decades to recover from being fucked over by Repugicans and their oil fetish.
Now, we're going to take what we gained and heat up the planet for this???
(Score: 3, Informative) by VLM on Monday September 23, @11:34AM (3 children)
Technically accurate in a strict legalistic sense but highly misleading.
For some decades now, it's been an empty building with a sign out front and its not worth the lawyers fees to adjust the paperwork to decommission just part of the overall plant.
Remember you can't simply shut off the security guard's camera to the empty reactor room because that security monitoring plan is deeply integrated into the security plan for the currently operating plant. Likewise you can't simply shut off the radiation monitoring system in that part of the plant because its decommissioned because its part of the active monitoring system on the other side of the plant. In a similar manner you can't shred or deep archive blueprints because some apply to one reactor, some apply to the other, some apply to both, and it would be a legalistic and safety nightmare to have to 100% accurately classify everything 100% perfectly, so its safer and cheaper to just keep the old wiring diagrams laying around even if the old wiring was literally shipped out and disposed of decades ago. The nightmare scenario would be something "happens" in the working plant and then someone has to write up an explanation of how they decided to throw away the first aid kit or the pipe leak kit or whatever because it was "part of the old reactor so we threw it out" although it wasn't.
It is, as I understand the situation, an empty-ish building. Everything "reactor" was shipped out and disposed of a long time ago.
I wonder if they store stuff for the other reactor in the empty building. It's a secured monitored area, may as well take advantage of the free space. In that way unless someone here has a security clearance and site access we don't 'know' if the building is literally empty, someone could have dropped a pallet of equipment in there last night. But it is correct that all the old stuff was physically removed decades ago, so whatever is in there this morning was moved in long after the accident.
(Score: 0) by Anonymous Coward on Monday September 23, @12:12PM
There are different approaches to take in decommissioning a reactor. For multi-reactor facilities, they typically put each unit in a "Safe Storage" (Saftor) mode where they get out most of what they can, and they sit on the rest and let it naturally decay while the other unit(s) are still in operation. This way they can do the final removal of everything at the same time and have the place re-habitable on the order of a decade or two. TMI Reactor 2 was done in this manner, and unless they change their minds, presumably its decommissioning is now further delayed since they are going to turn Reactor 1 back on.
Safstor can take many many decades [bbc.com]. But I'd wager that it is the most common method because it allows the company running the reactor to kick the can down the road for many decades and make it someone else's problem.
(Score: 3, Interesting) by Anonymous Coward on Monday September 23, @01:57PM (1 child)
> In a similar manner you can't shred or deep archive blueprints because some apply to one reactor
I'll bet there are some blueprints that the TMI owners would like to shred.
As I've posted before, an engineer friend worked for a company that had quite a bit of equipment in Three Mile Island. At the time of the accident, he was sent down to track their equipment, make sure it was working as designed.
It turned out that some critical sub systems that were supposed to have on-site backups (things like pumps, air-handlers), and even had storage rooms called out on the blueprints, were never ordered. Nice way to save on costs, eh? To repeat, safety critical backup equipment that was part of the original design was never ordered, and the storage rooms for it were empty.
Draw your own conclusions. My conclusion is that the non-engineers that run electric utilities are not competent to operate nuclear power plants.
(Score: 3, Insightful) by RS3 on Monday September 23, @10:39PM
> non-engineers that run
electric utilitiesany technical operation are not competent to operatenuclear power plantsthat technical operation.FTFY.
(Score: 3, Informative) by SomeRandomGeek on Monday September 23, @04:12PM (2 children)
Committing to bring new nuclear power online in 2024 is just insane. Here's a nice chart that shows that solar energy now costs a third of what nuclear does, and is dropping like a rock, while the cost of nuclear is actually increasing: https://www.vox.com/climate/372852/solar-power-energy-growth-record-us-climate-china [vox.com]
Here's an article about how the cost of solar is dropping like a rock: https://www.vox.com/climate/372852/solar-power-energy-growth-record-us-climate-china [vox.com]
It takes years to get a nuclear plant on line, and it takes decades to pay itself off. By the time they get this thing up and running, will it cost 5 times as much as solar? 10 times?
I know what you're going to say: "But there is no solar at night. Once you factor in the costs of storage, it is more expensive." So, let me just skip to the part where solar with storage is now the cheapest option in most of the world. See the article I linked.
(Score: 2) by DannyB on Monday September 23, @05:51PM (1 child)
How can you say that ?!?
This is Microsoft who is doing this.
Some kinds of payoffs happen rather quickly and quietly.
Don't put a mindless tool of corporations in the white house; vote ChatGPT for 2024!
(Score: 0) by Anonymous Coward on Monday September 23, @08:44PM
> Some kinds of payoffs happen rather quickly and quietly.
Along the same lines, Three Mile Island is downwind of Redmond, what do the Microsofties care if there is another meltdown...this time with radiation leakage.