DOE gives Microsoft partner $1B loan to restart Three Mile Island reactor
techcrunch.com
techcrunch.com
If you were to step into the control room you’d see analog phones, tiny incandescent bulbs behind plastic covers… looks like a sci-fi set from the 60s.
The expensive part of a reactor isn’t really the reactor or tech itself, it’s the government regulation from the DOE and NRC.
I worked at Areva/Framatome/B&W and IIRC they still have the archival room where hundreds of 4 inch D ring binders held the original design docs that had to be submitted for approval.
Not too disagree with the bulk of comment, but this sentence is not true. They're 90V neon indicator lamps, a technology that's really cool but also so inefficient that people rip it out and replace it.
Neat rabbit hole to do down.
> The debt facility is being made through the Department of Energy’s Loan Programs Office (LPO), which was formed under the Energy Policy Act of 2005 to foster the growth of clean energy technologies
and, more importantly:
> The Inflation Reduction Act, which passed during the Biden administration, created another pot of money under the LPO known as the Energy Infrastructure Reinvestment program. That program was created to restore existing power plants to operation provided they avoid or reduce pollutants or greenhouse gas emissions. The Trump administration kept it largely in tact, rebranding it the Energy Dominance Financing Program.
> The debt facility is being made through the Department of Energy’s Loan Programs Office (LPO), which was formed under the Energy Policy Act of 2005 to foster the growth of clean energy technologies.
> The Inflation Reduction Act, which passed during the Biden administration, created another pot of money under the LPO known as the Energy Infrastructure Reinvestment program. That program was created to restore existing power plants to operation provided they avoid or reduce pollutants or greenhouse gas emissions. The Trump administration kept it largely in tact, rebranding it the Energy Dominance Financing Program.
Congress passed the Energy Policy act of 2005 and then the Inflation Reduction Act allocating money to the DoE to make these loans.
https://www.ms.now/msnbc-podcast/msnbc/discussing-explosion-...
Nuclear is more expensive because there are extensive regulations. "Green" energy not only does not face so many regulations but it benefits from incentives.
Also, when comparing nuclear with "green" energy, most studies don't take into account the costs of energy storage.
Home grown nuclear programs will always be better than solar propped up by foreign entities.
Just because the USA squandered its early advantage in solar tech doesn't necessarily mean that it can never hope to regain a leading position.
Running away from the challenge now will all but guarantee defeat in the long term, so it's probably the wrong strategic decision to make.
- super low wages and borderline slave labor
- easy and cheap access to rare earth
- the CCP boosting your industries to flood the world
And, sure, plenty of rare earths are needed both for the drivetrains in wind turbines, and for the power electronics used by solar farms. But they'll also be needed for the steam turbines and power electronics in nuclear plants. Seems like it's pretty much a wash to me.
I don't have a dog in this fight, but it seems deeply weird for America to be refusing to even try to meet the challenge.
If you give up your sovereignty on topics like defence, energy or agriculture don't come crying in 20 years when you're someone else's bitch. Ask German's how it's going with the cheap russian gas lmao
They're cheap because they're made in massive "lights out" highly automated factories.
Access to cheap capital and automation. Those are supposed to be American strengths, and China is beating America at its own game.
The other things you are said are also true, I just wanted to provide a little historical context.
Oftentimes I worry about industrial pollution, unsafe and unfair working conditions, and a loss of relevance in high technology.
But in this case, the factories are not widely staffed - so there's reduced chances of accidents.
The future is now, and I'm the old man.
OTOH, cheap energy enabled by cheap solar panels is a huge benefit to any economy.
They stop selling? No problem, the ones you have will work for the next 2 decades. You'll be lucky to build one new nuclear plant in that much time.
Like happened in Sweden twice this year and in France at the peak of the energy crisis.
Granted, a project like this probably doesn't strictly need all $1 billion all at once, but I'd argue it's better to get whatever necessary funding upfront instead of risking having sunk a partial investment without being able to obtain the rest should the company's financial situation change.
My assumption is that the real reason this is a loan from the government and not paid directly by Microsoft has to do with other factors, like Microsoft not waiting to be in the hook for the billion dollars if the partner company folds, or the potential for loan forgiveness, or other incentives that make the effectively loan cheaper than cash.
I also think you would need more than 24 hours battery. You have to prepare for freak weather events that reduce system capacity.
I also wonder what time horizon we are talking. solar and batteries presumably have to be replaced more often than nuclear.
In general, yes. Not really in the context of utility generation for a DC, though. A DC should have onsite backup generation, at least to supply critical loads. If your contracted utility PV + storage runs out, and there's no spare grid capacity available (or it's too expensive) you can switch to onsite power for the duration. The capex for backup power is already required, so you're just looking at additional spending for fuel, maybe maintenance if the situation requires enough hours on backup.
Civilian nuclear reactors replace fuel gradually over 3 to 6 years typically. 20 year old solar panels work now. New solar panels are expected to work over 30 years. Utility scale lithium ion batteries are expected to last 10 to 15 years.
[1] https://www.lazard.com/media/5tlbhyla/lazards-lcoeplus-june-...
So Microsoft is less price sensitive than other electricity customers.
Plus they get the PR and hype boost from saying they are using nuclear, which is huge right now. Which is big enough that the other hyperscalers thought they had to announce new nuclear projects, even though it will be a decade before those new nuclear projects could ever come on line.
And unreliable energy sources routinely exclude the wildly uneconomical costs and environmental impact it would take to make them reliable.
See for example Oskarshamn 3 in Sweden having a 7 month long unscheduled outage this year.
Ringhals 4 had an 8 month unscheduled outage during the energy crisis.
Maybe modern day and tech has given "outage" a much shorter connotation than what it meant in the past.
7 months? That's almost longer than the Christmas offseason.
The greater the number and diversity of machines, as well as their geographical dispersion, the greater their availability.
In this respect, a mix of renewables (solar, wind, geothermal, biomass, etc.) deployed on a continental scale, along with storage (batteries and V2(G|H), hydro, green hydrogen...) is unbeatable (total cost, availability, risk, etc.).
For the right kind of workloads and at sufficient scale, I wonder if this is actually true. (It probably is, but it's fun to hypothesize.) I'm assuming the workloads are mostly AI-related.
AI training presumably isn't super time-sensitive, so could you just pause it while it's cloudy?
AI inference, at least for language models, presumably isn't particularly network-intensive nor latency-sensitive (it's just text). So if one region is currently cloudy... spin it down and transfer load to a different region, where it's sunny? It's kind of like the "wide area grid" concept without actually needing to run power lines.
Yes, I know that in reality the capex of building and equipping a whole DC means you'll want to run it 24/7, but it is fun to think about ways you could take advantage of low cost energy. Maybe in a world where hardware somehow got way cheaper but energy usage remained high we'd see strategies like this get used.
> Yes, I know that in reality the capex of building and equipping a whole DC means you'll want to run it 24/7, but it is fun to think about ways you could take advantage of low cost energy.
There's some balance between maximizing capex, business continuity planning, room for growth, and natural peak and trough throughout the day.
You probably don't really want all your DCs maxxed out at the daily peak. Then you have no spare capacity for when you've lost N DCs on you biggest day of the year. N might typically be one, but if you have many DCs, you probably want to plan for two or three down.
Anyway, so on a normal day, when all your DCs are running, you do likely have some flexibility on where tasks run/where traffic lands. It makes sense to move traffic where it costs less to serve, within some reasonable bounds of service degradation. Even if electricity prices are the same, you might move traffic where the ambient temperature is lower, as that would reduce energy used for cooling and with it the energy bill.
You might have some non-interactive, non-time sensitive background jobs that could fill up spare DC capacity... but maybe it's worth putting a dollar amount on those --- if it's sunny and windy and energy is cheap, go ahead ... when it's cloudy and still and energy is expensive, some jobs may need to be descheduled.
or pause it when "organic traffic" has a peak demand, and resume in off-peak hours, so that the nuclear powerplant can operate efficiently without too much change in its output.
not often and most importantly they are PREDICTABLE. You do understand why being able to control when a power plant is operating is a very important thing, right?
I guess if I knew there would be two months with less power I might design my data center to fit into 40 foot containers so I could deploy wherever power and latency are cheapest
https://www.nei.org/resources/statistics/us-nuclear-generati...
As for France's capacity factor, that has a lot to do with the presence of intermittents on the continental grid, combined with the EU's Renewable Energy Directive making France liable to pay fines if they use nuclear power in preference to wind/solar.
https://www.nytimes.com/2022/11/15/business/nuclear-power-fr...
In Sweden this year we’ve had 2 separate instances of 50% of the fleet being offline. With one reactor having a 7 month unscheduled outage.
I just don’t get where this ”100% reliable!!!!” is coming from.
2023: 93.0%
2022: 92.1%
2021: 92.7%
2020: 92.5%
Nuclear has the highest capacity factor of any other energy source—producing reliable and secure power more than 92% of the time in 2024. That’s nearly twice as much as a coal (42.36%) or natural gas (59.9%) plant that are used more flexibly to meet changing grid demands and almost 3 times more often than wind (34.3%) and solar (23.4%) plants
Nuclear power plants had a 8% share of the total U.S. generation capacity in 2023 but actually produced 18% of the country’s electricity due to its high capacity factor.
You can have a 93% capacity factor and still have short time periods with 45% of the fleet offline simultaneously.
Another example is when half the French nuclear fleet was offline at the height of the energy crisis.
Are you saying that a grid collapse is acceptable if outages correlate for nuclear plants?
Only if you have a tiny fleet with very bad management.
The US fleet might be large in absolute numbers smoothing out the average, but multiple plants in Florida having simulatenous failures won't be saved by Washington State plants having amazing capacity factors.
We still have a grid to deal with.
Are you comparing cost against what electricity currently costs or what it would cost to add capacity? I feel like Microsoft is not acting on hype here, they're going to pay a premium just because it's cool to refire a nuclear plant? Surely they've done the math to decide the feasibility of building out a few acres of solar panels.
It’s not really that farfetched, either. If the government expects a conflict in the next few decades, solar build out might become much more expensive or impossible since our domestic production might not be enough to support NATO’s growth.
PV did get spectacularly cheaper, but is not a panacea.
Nuclear is great fit for constant load, for example a cloud datacenter where relatively constant utilization is also a business goal and multiple incentives are in place to promote this. (eg. spot pricing to move part of the load off from peaks)
The article cited Lazard analysts' estimates to say this was more expensive than solar or wind. But Lazard's report said new build solar and storage could cost from $50 to $131. And new build wind and storage could cost $44 to $123.
Costs will rise over 20 years almost certainly.
And Microsoft had already large solar and wind power purchase agreements.[1] These could be affected by China's rare earth export controls scheduled to start next year. Hedging this position would be sensible.
[1] https://www.microsoft.com/en-us/microsoft-cloud/blog/2024/09...
> In 1988, the NRC announced that, although it was possible to further decontaminate the Unit 2 site, the remaining radioactivity had been sufficiently contained as to pose no threat to public health and safety.
Chernobyl (which was a far worse accident) continued to produce power at other units on the same site for 14 years after the meltdown of unit 4.
With that said, while it doesn't provide numbers, the article does say the refurbishment (costing $1.6 billion, estimated) will be cheaper than a new build. It'll also likely be much faster, projected to open in 2028.
A quick google search puts construction costs of new nuclear of a Unit 2 size in the $5-10 billion range. 3 Mile Island itself was constructed for $2 billion in 2024 inflation-adjusted dollars. All in all, refurbishing sounds like a good bargain compared to a green field build.
*Many reactors started construction in the 70s and were finished in the 80s or 90s, plus Watts Bar Unit 2 which was started in 1972 and finished in 2016 for a total of $5 billion. The US also of course builds many naval reactors.
The main problem is that things cost more per unit if you do them less. The first new reactor in decades is going to be stupid expensive because you have new people doing it who are learning things for the first time, which often means doing them over again, which is expensive. And then we didn't even get to see if the second unit at Vogtle could improve on the first because then COVID hit and made everything cost even more.
Whereas the interesting question is, how much do they each cost if you build them at scale?
https://www.sciencedirect.com/science/article/abs/pii/S03014...
Whereas what you want is multiple companies building these things on an assembly line, where you can plop the components out of a truck already approved and ready to be turned on.
> Whereas what you want is multiple companies building these things on an assembly line, where you can plop the components out of a truck already approved and ready to be turned on.
How many trillions in handouts from tax payer money to get there?
Renewables and storage deliver that dream unsubsidized today.
The modern US where a negligible number of reactors are being built, or the US in the period of the regulatory changes that accompanied the Cold War when everyone was afraid of nuclear bombs and then The China Syndrome and Chernobyl?
If you want it to mean anything you need data from when the regulatory environment isn't increasing in hostility over time during the measurement period.
> How many trillions in handouts from tax payer money to get there?
It's a loan at an interest rate higher than the government itself is paying to issue the bonds, i.e. the taxpayer is making money from this.
Also, the main issue isn't funding it, it's making it cost less, e.g. by moving most of the approvals from being needed at each and every site to the factory where components are being mass produced.
> Renewables and storage deliver that dream unsubsidized today.
Then why does the price per kWh that consumers are paying keep going up instead of down?
> It's a loan at an interest rate higher than the government itself is paying to issue the bonds, i.e. the taxpayer is making money from this.
This assumes the risk is zero. Which given e.g. Virgil C. Summer is not the case.
> Then why does the price per kWh that consumers are paying keep going up instead of down?
Depends on where you are in the world. In Europe most of the recent price increases are coming from fossil energy becoming expensive.
The ETS system is making coal power expensive and running a peaker on LNG is extremely expensive.
https://en.wikipedia.org/wiki/European_Union_Emissions_Tradi...
So now we're seeing large swings from extremely cheap renewable electricity to expensive fossil based.
Apart from the obvious labour costs difference , theres also the skills at scale.Chinese have been on a continous buildout of new plants , so at this point they have designs/skilled teams for whom this is another routine at this point(i think 30+ under construction concurrently).The US builds are almost artisinal at this point.
And yeah at $1B , given prior examples , it expect them to be late and costs to baloon.Unless they use this as a template to upskill/retrain a workforce that will lead a new buildout so economies of scale take over and put downward pressure on the costs.
[1] https://www.nytimes.com/interactive/2025/10/22/climate/china...
in a list of countries with uranium resverves 1-59 they're number 55!
https://en.wikipedia.org/wiki/List_of_countries_by_uranium_r...
And in any case Australia host a LOT of uranium and is a very close ally and is happy to sell it to the US.
And every single argument is a weird either or scenario. Like some people want Gas. JUST Gas. Some people want Solar and Wind. JUST Solar and Wind.
Nuclear power can also fill batteries. Can also fill pumped hydro storage. Ditto Gas. Nuke and Gas are good for restarting a grid when theres a catastrophe, see Spain.
Give engineers more tools, not less. Its infuriating.
They could have set up green steel and aluminum industries supplying the world. Instead they ship millions of tons of unprocessed ore and thermal coal to east asia where it's processed with CO2 intensive energy, and then the metals are shipped back in the form of automobiles and construction materials.