Google commits to buying power generated by nuclear-energy startup Kairos Power

edit: to be clear, 1GW of wind or solar is $1B. Build 3GW for overcapacity and you’re still at just 17% of the cost of 1GW of nuclear, and you technically have 3x more capacity. Now figure out how many megapacks you can buy for the $14B/GW you saved https://www.tesla.com/megapack/design (answer: 16GW/68GWh)

It's no conspiracy why nuclear never gets traction these days -- maybe it was cost-effective 10-30 years ago but renewable technology has gotten relatively cheap. (Shutting down active nuclear reactors earlier than needed is a whole different issue though.)

Here's the report for 2023: https://www.eia.gov/outlooks/aeo/electricity_generation/pdf/...

There is no report for 2024 because they are building a new model to take into account even newer technologies: https://www.eia.gov/pressroom/releases/press537.php

-- Edit --

To clarify, "Nuclear is a terrible investment for private industry in 2024." However, I understand why nation states (and their equivalents) would want a diversity of power sources. There many be non-economic reasons why nations want to build nuclear over solar+battery+wind.

History is not really against them. Our current reactors (mainly pressurized water reactors) are the way they are because Admiral Rickover determined that PWRs are the best option for submarines. He was not wrong, but civilian power reactors are not the same as the reactors powering submarines.

PWRs are expensive mainly because of the huge pressure inside the reactor core, about 150 times higher than the atmospheric pressure. For comparison, a pressure cooker has an internal pressure about 5 times higher than the atmospheric pressure, and such a cooker can explode with a pretty loud bang.

The Kairos Hermes reactor design is based on a design that was tested in the '60s, the Molten-Salt Reactor Experiment [1]. While such a reactor can be used to burn thorium, Kairos decided to go with the far more conventional approach of burning U-235. The reactor operates at approximately regular atmospheric pressure. This should reduce considerably the construction costs.

Of course, there are unknowns. While the world has built thousands of pressurized water reactors, it has built maybe 10 molten salt reactors. For example one quite unexpected effect in the MSRE was the enbrittlement of the reactor vessel caused by tellurium, which shows up as a fission product when U-235 burns.

The Nuclear Regulatory Commission is a very conservative organization, and they don't have much experience with molten salt reactors because nobody has. It took them 6 years to give NuScale an approval for a pressurized water reactor, design that they knew in and out. My guess is that they will not give Kairos an approval without at least 15 years of testing. But Google's agreement with Kairos is quite crucial to keep this testing going.

[1] https://en.wikipedia.org/wiki/Molten-Salt_Reactor_Experiment

Or alternatively, radioactive dust could be released into the atmosphere such as THTR-300 did.

INL did a gap analysis in 2011 between what was known and what needed research. The german AVR reactor had technical issues that weren't expected -- dust being one of them.

https://inldigitallibrary.inl.gov/sites/sti/sti/5026004.pdf

From what I can tell the dust issue is still a point of contention.

Getting Google in line as a customer is absolutely huge for Kairos.

If a technological solution is optimistic and remains vaporware possibly forever, then it maybe "virtue signaling" is if there more nonfunctional desire for it that outstrips practical or economic utility. A better term would be "vaporware" when there is less social puritanism involved, and I don't think coal or nuclear signal anything of redeemable greenwashing value compared to cheaper renewables combined with PES and distributed grid storage.

I believe this is referring to rather small quantities (kilograms) imported for pH control in light water reactors.

If a nuclear power plant-owning VC-backed startup cuts corners because their shareholders expect 3 cents more of profit YoY on their earnings report, we get a nuclear meltdown that shuts off an area the size of a small city for a century.

These startups can barely keep passwords secure, I don't trust them and their VC psycho buddies for a second with something as potentially powerful as nuclear energy.

I have heard this before, but is this just the physical waste's volume? Isn't that a useless metric? What would happen if you included the volume of the containers required to safely house it?

but the great thing about next gen reactors is that the waste solution does not need to be addresed; any waste from next gen reactors will simply go wherever the final solution for existing waste engines lands.

https://insideclimatenews.org/news/10102024/inside-clean-ene...

https://coal.sierraclub.org/coal-plant-map

https://www.vox.com/climate/372852/solar-power-energy-growth...

https://www.dnv.com/news/eto-energy-related-emissions-will-p...

https://news.ycombinator.com/item?id=41602799 (citations)

"The best time to plant a tree was twenty years ago. The second best is now."

Like, 1960 itself clearly belonged to the 1950's, the same way 1980 still belonged to the 1970's -- culturally, that is.

Obviously, the question of what year a decade "really" started in, allows for endless argument. :)

Without evidence this is just a repetition of an old, unproven meme for internet points.

Caveat. I did pencil in data centres using nuclear as highly probable some years ago (I worked in geophysical energy+mineral exploration for decades then moved to resource intelligence presentation).

Won't I don't see is the "inadvertently saves us from catastrophic climate change" part .. sure, that might happen but it doesn't follow that it will - the more probable outcome is that overall we humans just consume even more energy with some of that addditional energy coming from nuclear.

And the TMI accident didnt really make any land radioactive and Fukushima seems cleanup-able.

Also claiming certain knowledge of the financial outcome of a large engineering project before one has even a prototype is pretty much the definition of hubris.

The podcast also says that there are 92 reactors operating in the US now and that one of the reasons why the US is so far behind other countries in the nuclear industry is that 'no four are the same'. So why not pick the most successful one and build a copy?

The word cheaper is doing a lot of work here. How much cheaper?

I'm not arguing that modularity is bad just that there doesn't seem to be any hard evidence yet that they will be truly modular, certainly not in the same sense as washing machines coming off a production line are. Look at how many rockets SpaceX built before they got it right.

Even production lines for much simpler things generally take some time to settle down. The idea that the second one will be noticeably cheaper than the first seems rather unlikely because it will have to incorporate fixes for the things that didn't quite work right on the first one.

Also a 300 MWe reactor is still very large machine and has to dump huge amounts of heat somewhere so the site is still important so making the individual reactors cheaper does not make the whole project correspondingly cheaper.

[1] https://www.wipp.energy.gov/

No. This is a false dichtomy pushed, from what I can tell, by the gas lobby. It's solar and wind + nukes or gas.

Batteries work in theory but not in practice: production doesn't scale fast enough, and that was before LLMs brought a new and growing source of power demand to the table. (I'm ignoring that grid batteries compete with transport electrification. A combination of economies of scale and common bottlenecks in construction of battery plants, irrespective of chemistry, links the pursuits.)

For energy we obviously need all the options available.

If a major volcano goes off up and darkens the sky with clouds and high winds make wind farms unsafe to operate, then nuclear is probably our only reliable power source left. It's not like there weren't multiple ice ages and warming events in the history of our planet.

There is a reason sailboats were obsoleted by the steam engine: it could tug forward in windless waters and stll make it fast enough to deliver the mail. The base load power station is the steam engine. The sailboat is the wind turbine or the PV array. Most of them need a gas fired power plant to compensate for windless or cloudy days, like newer sailboats need an engine. We could use a load following SMR in place of the gas fired plant.

Nonsense.

Such things can be regulated, but my point is that solar and wind are perfect for h2 generation. The sun shines? Produce. The wind blows? Produce.

The variability is irrelevant, and the result is the creation of a fuel source that can be stored.

Even better, we already have an immense network of Ng pipes, and there have been many tests and studies on injecting h2 into Ng lines, and pulling it out at the other end with molecular filters. There is no molecular reaction either.

The means low cost, massively deployed infra already exists.

And this massive network of Ng lines, with h2 injected, can in effect be an immense storage tank of h2.

We don't need some unified "batteries only" group think, but instead having multiple clean sources of energy is a boon. Just the cost of adding 3x the power transmission capacity, distribution is daunting, h2 can let a faster rollout of clean transport occur.

We should embrace all paths which the market can endure amd which can be green.

The Germans ended up focused on one only.

My point? H2 is perfect for solar.

https://en.m.wikipedia.org/wiki/FLiBe

That isn't the only worry. If the fuel is smuggled out... https://spectrum.ieee.org/high-assay-low-enriched-uranium

(Sadly, we have to consider that long because the current climate change is also unprecedented on these timescales.)

I am appalled though that anyone is seriously considering the idea of container-sized reactors, for which we can be nearly certain, if this seriously takes off, and we make million(s) of them, that hundreds if not thousands will end up abandoned in random places, and weathering will pierce them before they become radilogically inert.

It’s not there yet. CAES is cheaper than hydrogen today at grid scale, generally speaking : https://www.ctc-n.org/technologies/compressed-air-energy-sto...

How does that follow?

How does using nuclear for some of our energy needs bias the rest of our energy sources towards fossil fuels? As opposed to renewables or even more nuclear?

Chernobyl is a bad example.

The Soviets knew it was an inherently unsafe design and built it anyway.

When you play stupid games, you win stupid prizes.

Fukushima is a better example.

That's a failure mode common to all large centralized power plants, not exclusive to nuclear. For instance, an unexpected failure (a triple fault on one of the transmission lines coming from the power plant) of a single hydroelectric power plant caused a country-wide power outage in two countries (https://en.wikipedia.org/wiki/2009_Brazil_and_Paraguay_black...).

A better question is why nuclear needs that much regulation, it's all just political difficulties. Nuclear technology itself don't need this much.

I thought the answer was going to be a coal plant until I read the rest

And I don't want to hear the Fukushima partial-meltdown was operator error and we just need people to not make mistakes. I'd rather be told that accidents will happen, radioactive substances will leak, and we have ways to deal with that.

Are they the same as new large LWR, as in, must withstand a 737's impact?

This isn't the defense or retort that you think it is. This is from ONE incident. The industry recognizes the long tail of catastrophic failures is so large that laws have been passed to limit nuclear power liability in the US. I'm talking specifically about the Price-Anderson Act [1].

This severely limits nuclear power liability to ~$500 million per incident. That's a lot less than $1T or $750B or whatever figure you prefer. So who picks up the tab in a catastrophic event? Taxpayers. It's another example of how nuclear power can only exist with government subsidies. Yet we apparently want to trust private corporations to manage nuclear power plants and take the profits while shifting the risk to the public.

Additionally there's a self-insurance fund, but it would be completely inadequate to cover an incident like Fukushima. This is recognized [2].

> The cost of burning fossil fuels is estimated to be well, well over $1T.

So this is a strawman on two fronts. First, you're comparing the cost of a single massive failure by one plant to the entire fossil fuel industry. Second, I never even said fossil fuel power. I said solar.

[1]: https://crsreports.congress.gov/product/pdf/IF/IF10821

[2]: https://thebulletin.org/2020/02/the-us-government-insurance-...

Nuclear power has almost unlimited downsides and fairly limited upsides.

It’s at least as expensive (and mostly more) to maintain a nuclear power station As any other form of (non-carbon) power generation - and the costs of catastrophic failure and orders of magnitude higher.

This is just back of the envelope maths. Cost to maintain a power station of X GW for 100 years is X, cost to maintain solar panels of X GW for 100 years is Y. Cost of total catastrophic failure in fifty years of solar panels because the country fell apart is small y. Cost of total catastrophic failure of fission reactor is huge great X.

The simplest analysis just comes up with huge downsides.

Look I take the train past Battersea powerstation most days. It was built 100 years ago at the height (?) of the British Empire. It became disused as Britain fell into bankruptcy in the 70s and was left to fester for decades before people realised a vast shopping centre in the middle of London was quite nice.

If it was nuclear it would still be sealed off, any slacking of maintenance, any cost saving too far, would fuck up the world’s greatest city .

And if you think the worlds most powerful and richest country could always afford the very best maintenance - let me introduce you to political decision making in the 1960s, industrial policy in the 1970s and human beings who tend to hope as a strategy.

It’s not hard to pretend the obvious won’t happen, and if you take the risk sometimes you will be right. And the cautious man will look silly.

But in the end Warren Buffet looks more sensible than Dick Fuld.

And even Warren is aware he pays far less tax relative to his maid. But it’s upto us to fix that just as it’s upto us to not make bad investment choices as a society that we will pay costly annual fees for centuries to come.

1. We want to generate electricity with minimal carbon output. 2. Nuclear is part of this equation (along with solar, wind and tidal. Maybe one day fusion) 3. Nuclear has large capital upfront, a maintenance cost that requires us to always be on the A game, and the cost of catastrophic failure is fucking huge. 4. The other options have downsides of course, but their ongoing maintenance is basically lower because the catastrophe cost is much much much lower. 5. It’s really hard to quantify things like “major urban area made uninhabitable”, because it has almost never happened. But it can and it will if we keep chucking risk around like this.

6. The way to stop this, no the way to align investment, is to correct price externalities - positive and negative.

If we want to re- start places like Indian Point (a relative well Managed successful nuclear plant whose history reads like a series of disasters) then we ask what if Indian point failed like Fukushima.

That’s Westchester, and most of Manhattan that suddenly looks like a disaster movie. No Fukushima was not actually as deadly as feared (1 person kinda), but the evacuation and knock on effects. Try that on the Hudson and see what the cost of evacuating New York is - I mean, shipping, finance everything.

Honestly I struggle to see what’s crazy anymore.

How about every bond raised to fund a nuclear plant has a 100 year lien attached that no payments can be made till a century of safe operation and closure has occurred.

If the financials make sense after that I will take another look.