Hacker News new | past | comments | ask | show | jobs | submit login
HH70, the first high-temperature superconducting Tokamak achieves first plasma (energysingularity.cn)
233 points by zer0tonin 5 months ago | hide | past | favorite | 261 comments



Insane factoid (post from Feb 27, 2022) ... this was funded by a Chinese gaming company and built in 2 years for relative pennies??:

MiHoYo, the developer of Genshin Impact, has led a $65m funding round in Shanghai based Energy Singularity which is a company involved in nuclear fusion technology, tokamak devices and operational control systems.

The company plans to build its own Tokamak device by 2024.

https://x.com/ZhugeEX/status/1497957735337443331


Unrelated to what you are citing, but I believe a “factoid” is something that looks like a fact but is not. Like how a planetoid looks like a planet but isn’t one.

I only realized this myself decades after using the term factoid due to pages in highlights for kids.


This is a British vs American English thing. In British English a “factoid” is something that looks true but isn't. In American English “factoid” is a synonym for trivia--something that is true, but of minor importance.


> In British English a “factoid” is something that looks true but [ ... ]

may or may not be true.

Wikipedia has it as "an item of unreliable information that is repeated so often that it becomes accepted as fact." after the original USofAmerica coinage by Norman Mailer.

In Commonwealth countries (Australia, Canada, UK) two decades past we used it on intelligence forums as the name for atomic snippets of information released by companies via stock exchanges, company reports, PR .. each nugget being an atomic fact like paragraph linked back to a source that asserted that fact to be true, but to be taken as potentially incorrect.


Literally the very next line: "Since the term's invention in 1973, it has become used to describe a brief or trivial item of news or information."

The intended meaning by Norman Mailer never took on in the states.


Literally you asserted:

> In British English a “factoid” is something that looks true but isn't.

I responded that

In British English a “factoid” is something that looks true but may or may not be true.

.. there's a difference.


So, a factoid being sometimes true, but not always... Is a factoid.


That’s arguable.


This sounds like a factoid to me.

Jokes aside, what do we actually do in this scenario, when the same word has opposite meanings?

In my opinion, it’s always best to err on caution and use another word if possible (“short fact” instead?).

Because I have seen this factoid discussion before…


A wonder how different people will interpret "a couple of factoids" then!


This is only true if “American English” means English spoken by people of low education or as a second language.


The other meaning is a small or trivial bit of (true) information.


I thought the second definition came about from continual misunderstanding of the word, like how literally no longer means literally.


BTW, what is the new word to use when one literally means literally?


There is none. The word has been misused to the point of ambiguity being an accepted part of its definition, and we are all worse off for it. The language is now less expressive, and you need to use more words to add context and remove ambiguity when you really do mean "literally" in the literal sense.


‘Actually’ is what I’ve heard most often.


Use literally. It still means literally. Language has all kinds of things like sarcasm, exaggeration, and metaphor that change the way a sentence should be interpreted, but the meaning of each word remains the same.


You add “quite” before “literally”.


Just prefix the sentence with “literally literally (not literally literally)”


Gen Z uses the very awkward "unironically".


that too, will often get used ironically


Sure, but that’s how language works. Lots of words that we use in modern English have drifted away from their original meaning.

Language is the shared meaning between people, so if lots of people understand something the same way… then thats what the word means now


The curious thing is that Norman Mailer coined the term about 1970. Is drift accelerating, or do words so new lack the stability of the old?


I don’t think it’s that. I think it’s that new words are less stable than old ones.

In the same way that if you want to predict which authors will be well known in 400 years, your best bet is on authors that we currently know from 400 years ago. Better to bet on Shakespeare and Aristotle, than e.e.cummings and T.S. Eliot

A word coined in the 1970s won’t be nearly as entrenched in its meaning with the public as an older word.

So, that’s my suspicion. New words are more prone to drift than old words


Quite that factoid. How do we know which it is? ;p

Guess it goes both ways... which is kinda worse.


I really like to call those factlets, but that's probably just me.


Other examples I like to trot out: Android is not really a man, Asteroid is not really a star, Meteoroid is not really a meteor.


Factoids (true or not) seem to have special appeal for people who like to socialize with others by knowing things - the Cliff Clavens of the world. It has an overtone of superficiality along with triviality.


“A factoid is either an invented or assumed statement presented as a fact, or a true but brief or trivial item of news or information.”


Literally a useless word on its own now that the definition evolved this way... many such cases unfortunately.


Funny that the word "literally" have evolved in a similar way


Until they've made a billion dollars I'd assume the situation is not that rosy. It is easy enough to do a cool science experiment for $65 million. That being said, I applaud anyone achieving any result when it comes to energy.

My first test for Chinese success is "would this have been legal in the US?". I'm not sure who regulates tokamaks but I assume they have a similar risk profile to nuclear reactors (nuclear process releases a vast amount of energy in a tiny space) and so it would be normal if building one commercially was prohibited.


But they don't have the stored energy density of fissile nuclear reactors; they have the stored energy density of a big particle accelerator. Shutting down a particle accelerator (either temporarily or permanently) is way easier than shutting down a fission pile, because a particle accelerator or fusion reactor would just dump the plasma into a graphite bed and dump the stored magnetic field energy into a bunch of large copper bars acting as big resistors.

EDIT: If you shot a hole in a fusion reactor, the cold air would immediately quench the plasma down to room temperature.

https://www.fusionindustryassociation.org/nrc-decision-separ...

Supporting letter from Helion Energy: https://www.nrc.gov/docs/ML2224/ML22243A083.pdf


Fission has a very different risk profile than fusion.

Additionally making a fusion plant isn't a stepping stone to making a nuclear bomb


> making a fusion plant isn't a stepping stone to making a nuclear bomb

In theory a fusion plant can use the neutrons to irradiate the right chemical element to produce Plutonium-239 or Uranium-233.

  It has been estimated that each 14.1 MeV fusion neutron could be used to produce up to 0.64 plutonium or 233U atoms [4] assuming a TBR of 1.06. This corresponds to 2.85 kg plutonium per MW-year of DT fusion power production, assuming that all of the neutrons are captured in the blanket.
https://duckduckgo.com/?q=proliferation+risk+fusion for more info


> making a fusion plant isn't a stepping stone to making a nuclear bomb

What about a hydrogen bomb?


How would you make a bomb out of a tokamak? It is barely stable enough to generate the little heat required to generate energy, any disruption to that will just put out the reaction it wont explode.

Fusion bombs requires fission bombs as a fuse to have enough heat to explode, fusion reactors wont even come close to that.


Their risk profile is basically zero.

I doubt (nuclear) regulations are stifling much innovation here.


certainly not zero: fusion reactors still contain a decent amount of toxic materials, some of which radioactive, and when working make all kinds of highly radioactive elements inside the structure of the reactor. Probably less of a risk than fission, though (it's pretty unlikely a failure will result in an explosion or runaway reaction which spreads those radioisotopes far and wide).


Why would you assume fusion has the same risk profile as fission? They are opposite ends of the periodic table :)


China's best coding LLM, which beats GPT4 on coding benchmarks ( https://github.com/deepseek-ai/DeepSeek-Coder-V2 ) was trained by a random Chinese hedge fund.


I doubt there's much in terms of brainpower, compute or financial resources to differentiate a hedge fund and an AI research lab.


I just see some great market politics from Chinese leadrs


The old saying about absolute power corrupting absolutely clearly has parallels in all other fields: Absolute money corrupts vision and focus.

Tesla: "We did it. We have become profitable and created a real product people want. Now we can laser focus on making it better and more reliable and cheaper for everyone!" "haha nope! lets put it all into crypto and humanoid robots and impregnating as many CEOs as possible, let that bet ride bayyybeeee!!!!"


I don't think this is a corruption of focus - MiHoYo has had "Tech Otakus Save The World" as their slogan long before Genshin made its first billion dollars.


Not to mention that they can make back $65M in just a few weeks from one of their two mobile games and they are about to launch a new one. This is basically pennies to them.


One of their three. HI3 and HSR aren't the same game.


Don't forget their best game tears of themis


I've always wondered what they meant with that slogan, but now it makes sense.


Also Tesla: drive the price of EVs down to parity with ICE cars while delivering a superior product, built out the nations charging infrastructure (and got everybody to switch to NACS), and oh yeah: made self driving available to everybody for next to nothing.


> Tesla: drive the price of EVs down to parity with ICE cars while delivering a superior product

Tesla was more than willing to jack up their prices and maximize profit when they could. What drove prices down on Teslas was real competition from the incumbent manufacturers. And inflation cooling people's willingness to blow a bunch of money on expensive cars. And CATL making batteries less expensive. And even then, their cars are only at parity right about now, with a $7500 tax credit. And also only if you are fairly loose about what features you need to consider 'parity' achieved.


Let’s just ignore that oil companies owning patents killed evs for years my gripe here is Tesla’s “self driving” isn’t actually self driving? It is basically advanced cruise control and requires supervision, Tesla is not liable for it running into things, and there is no indication of that changing anytime soon?


Basic Autopilot that is free on every model is advanced cruise control + lane keeping.

FSDS (Full Self Driving, Supervised [for now]) can handle the vast majority of driving scenarios, from A -> B. I currently intervene once per 10 drives, usually due to a routing issue (never safety critical). It is rapidly improving, and will drop the human requirement once it surpasses most drivers.


FSD still requires you to pay full attention. The name is still a lie and Elon has been claiming the human requirement will be dropped “soon” for what now? Almost a decade?


> for next to nothing

Their cars are certainly out of my price range. Plus, openpilot has been doing it for free for years.


you should try driving with openpilot and compare it with FSD to see if they're equal since you are making this comparison


So, most other new cars are as well?

OpenPilot works well in limited scenarios, but even the founder George Hotz openly admits Tesla is significantly ahead and has the right approach.


“Self driving”


Not sure where an unreliable, very expensive to fix, poorly QC'd cars are a "superior product" but it's not where I live. The charging infrastructure is a great feat though.


Even more reason to dislike Genshin Impact.


Sorry, how does contributing to an energy project cause dislike for a barely related video game?


Spending money in that casino-cum-game funds energy research in a foreign adversary nation.


Giving me “nobody can make the world a better place more than me/us” vibes. I’d be happy if my leisure spending went to causes that benefits humanity, no matter which country is making the contribution. That said, I agree that gacha/casino game mechanics should be made illegal


Tokamak energy did this back in 2015[1,2] (the article is wrong)

[1]: https://tokamakenergy.com/about-us/#trackrecord

[2]: https://royalsocietypublishing.org/doi/full/10.1098/rsta.201...


ST40 does not use HTS magnets. The magnets are made from copper and LN2 cooled. The company is demoing HTS magnets but has not used them in a working tokamak.


ST25 HTS did as far as I can tell — unless TE are lying, but that seems improbable.


ST25 was a "desktop tokamak" measuring 25cm in diameter.


Am no physicist, but was wondering about that. Thought this was all done before by different fusion companies?


The point is the high temperature superconducting magnets. Creating a tokamak plasma is not novel.


When talking about the price of energy produced by fusion, various estimates put it at 'probably about the same as nuclear fission, maybe a bit higher, but it won't have the proliferation risk/contamination risk of fission'.

However, because the tech was '50 years away', it never made sense for private sector investors, so most investment was from governments.

However, with solar and wind now far cheaper than nuclear due to no need for massive capital investments in concrete and steel upfront many years before production starts, does it even make sense for governments to go down this route?


AFAIK it's not at all clear that solar and wind are really cheaper when making up a substantial part of a large-scale power grid that meets our current expectations of 100% consistent and reliable power everywhere, no matter what.

The unreliability of solar and wind requires either hot (constantly running and spinning) non-renewable backups or grid-scale power storage (has never been done so ? on cost to build and upkeep) to guarantee reliable voltage and AC frequency. The cost of that should be factored into determine the true cost of these power sources.

The stability of the grid is dependent on the collective physical inertia of the many tens of thousands of huge and heavy spinning turbine-generator sets that make up the majority of the current generating capacity. Most current solar power sources rely on grid-following inverters, which are not stable without a grid stabilized by a preponderance of large spinning turbines. There has been some work on grid-forming inverters that are less impacted by this, but AFAIK there aren't currently any that can replicate the grid stability provided by that physical inertia.

I'm less certain about wind turbines, but I think they have this problem too. I don't think they're controllable enough to be mechanically synced to the grid frequency.

I'd love to be wrong about this, please prove me so if you can! But I don't often hear these points addressed, and we're not helping anything by ignoring the complexity of the real world.


Inverters can easily replace physical inertia, it just requires technology developed within the past 30 years, and most grid folks haven't thought about new technology for far longer than that.

As more and more intermittent renewables get pushed onto grids, they become more reliable. Most outages are from single points of failure from large generators or transmission. Dealing with highly distributed renewables means that grid ops get used to acting fast, and there's greater redundancy instead of so many SPOF. Kind of how cloud services got reliable by expecting there to be failure and designing it into the system.

Storage is advancing super quickly, is super fast to deploy, and can replace a lot of more expensive things like transmission upgrades.

We have all the tech to replace fossil fuels on the grid with the above. The only question is the final cost. It's likely to be far far lower than using existing "hard" energy, because by the time we can deploy 50 TWhs of storage, it will have gotten so cheap. We don't know when costs will stabilize, but they have a loooong distance to fall.

And we have all sorts of other technologies that will make all this far cheaper: enhanced geothermal, enhanced geothermal with temporal storage based on injection pressure and release, iron air batteries, flow batteries, thermal storage for industrial process heat, etc. etc. etc.

For every area of the energy economy, there are two to three solutions that look promising. Fusion and fission look promising for none. That's not to say that they can't have some serious innovation and start dropping their costs, but nobody currently operating in the field has demonstrated a path. Yet.


> As more and more intermittent renewables get pushed onto grids, they become more reliable

This is called grid firming, and it’s massively expensive.


I don't think that's the same thing, but what expenses are you thinking of specifically? The German grid for example got much more reliable with additional solar.


Grids were designed to operate in tight tolerances. Cycling power levels up and down a lot, while handling the frequency variation lots of renewables inputs bring, wears down the grid without protective measures. Those measures are called firming [1].

Not sure what Germany did (or plans to do—you can run an unfirmed grid until stuff starts failing for several years).

[1] https://www.gevernova.com/gas-power/applications/grid-firmin...


"Capacity firming" will be carried out by legacy gas turbines as they run less and less, and eventually by batteries.

Batteries are also much better than gas at frequency regulation, and even at the prices a decade ago, completely took over the market for frequency regulation in the PJM market in the US. But frequency regulation is very very tiny in terms of power needs, it only takes a very small number of grid batteries to completely solve that problem.

The amount of batteries waiting in the interconnection queue completely dwarfs gas. There will be no "firming" coming from new gas turbines, unless old-school corrupt utilities are able to sneak it by PUCs by creating some sort of crisis and tricking them.


> "Capacity firming" will be carried out by legacy gas turbines as they run less and less, and eventually by batteries

At least among the American TSOs, there are zero I know of that plan to do this. Do you have a source for one that does?

Trillions have already been spent on gas. That infrastructure will need to earn its return through the 2040s at the very least, and that precludes running them exclusively for firming. To the extent retrofits are being discussed, it’s as an add-on amidst full peaked functionality.

> Batteries are also much better than gas at frequency regulation

Limiting solar and wind by utility-scale battery capacity means scaling back EV adoption or solar and wind deployment. The math simply doesn’t work. (Again, in America. Without significantly raising rates. Not sure elsewhere.)

> it only takes a very small number of grid batteries to completely solve that problem

Frequency regulation is one component of firming. Batteries are good at some components, marginal at others. (As a system. Technologically, they're fine.)

Apart from de-industrialised grids, a batteries-only approach has been practically abandoned through the 2030s. It's why we're building so many turbines and abandoning nukes.


If every (second) house would have a powerwall, wouldn't that make the grid stable?


> every (second) house would have a powerwall, wouldn't that make the grid stable?

At 131mm American households [1] and $11.5k per PoweWall [2] that’s over $750bn at 50% loading.

[1] https://www.statista.com/statistics/183635/number-of-househo...

[2] https://www.thisoldhouse.com/solar-alternative-energy/review...


China is starting to mass produce NaCl batteries. They will be cheaper.

And also only Powerwalls produced in this quantity would have way lower prices. My point was batteries are getting cheaper every day.


It could be less stable, if each and every powerwall is slightly out of phase.


But this is something, one can avoid by only allowing well tuned batteries to the grid? Or is this a serious problem to get right?


A powerwall is $12k installed.

That's about 5 years worth of power bills for most people.


What new technology changes things to not require spinning turbines?


Integrated circuits, basically. The term if "grid forming inverter" and the standards are somewhat new. I'm not an expert, but here is one standard, I don't know if it has been adopted or if others are preferred:

https://www.energy.gov/sites/default/files/2023-09/Specs%20f...


And they have been used where? Everything I can find suggests they are theoretical, not in use even on micro grids yet, and many have huge concerns about them being able to support a grid on their own.

Also the technology required has been around for decades it’s not new and no one’s done it yet.

So again what’s changed because it seems like for now there is no “I have a grid and need something now” solution it’s a “maybe one day”


Massive improvements to battery technology.


That’s a non answer - what improvement


It is an answer, in that batteries were used for frequency regulation far before they were used for energy arbitrage.

But the real innovation is communication networks and IC control of the inverter. It's completely possible to create a waveform that modulates and responds to variation in frequency in the same way a large spinning mass would. And if reactive power is for some reason not enough, synchronous condensers are very old technology to solve that.


> grid... grid-scale power storage... stability of the grid ... grid-following inverter ... grid stabilized by ...

The problem isnt solar, or wind, or storage ... the problem is the grid. Were running on a system that was never designed to do what were asking of it, and yes its going to be a number of problems to solve. All of those are jobs, economic action and improvements to reliability and quality across the board.

> I don't think they're controllable enough to be mechanically synced to the grid frequency.

Google, there are a number of ways this gets addressed.

> There has been some work on grid-forming inverters

Yes, we know how, and the race to build them is on... this isnt a hard problem it's just a problem.

> grid-scale power storage (has never been done so ?

Already deployed in a few places with battery systems (hati, Australia both have them. Possibly Hawaii too). We're doing quite a bit of this. Again a quick google will give you a sea of sources.


Battery backed renewable energy with grid upgrades is cheaper today and getting cheaper.

https://www.csiro.au/en/research/technology-space/energy/Gen...


> Most current solar power sources rely on grid-following inverters

There are now inverters that simulate the rotational inertia. They simlpy shift the phase of the generated waveform just a bit if the frequency starts dropping.

And it doesn't require any expensive additional hardware.


I think the solution will come from storage but also from a massive grid-wide ability to shed non-critical loads on-demand. The current grid is built on early 20th century principles, before we had real-time digital communications.

As a though experiment - imagine a 19th century world suddenly getting all of our current digital tech and wind farms and solar power - there would be no point in trying to create a "static" grid where producers and consumers weren't communicating with each other. Every consumer would negotiate power availabity based on momentary price.


Batteries can mimic inertia better than physical spinning objects.

An operator in Australia has seen massive success and profits over the past few years using batteries to out-compete other grid stabilization. IIRC, they have already made enough to pay off the upfront costs. Even better, Australian government was super against the change, but now most places are positive on them because of the obvious success.


Regarding stability, this physical inertia is also present in rotating wind turbines. But I guess exploiting this at a meaningful scale would recquire a level of interconnectivity which boils down to the same issue of cost.

I am however optimistic about grid-scale storage. There is a long term trend of rapidly dropping battery prices, and with recent developments in sodium ion batteries there is no fundamental reason this won´t continue. Another enabler could be advancements in lifespan. This could allow storage being installed inside or near wind and PV, cutting down on space and installation costs. Even then however, grid improvements would be needed.

Some problems still need to be solved indeed, but in my (mostly uneducated) opinion, they seem easier than achieving economically viable fusion. But they do still require large investments in R&D and manufacturing capability.


Wind turbines don't get the same feedback like traditional generator turbines do.

Wind turbines are designed to run on unstable wind speed -- this meant it have to somehow decouple from the main grid


Wind turbines do not provide grid inertia.

They do not spin at 50/60hz, they are deployed with frequency converters to match optimum generation to the grid and spin at whatever speed they can achieve.

They're essentially another type of solar plant.


> grid-scale power storage (has never been done so ? on cost to build and upkeep)

This is out of date. Grid scale battery storage has recently become economic in lots of cases and is ramping up quickly.


> Grid scale battery storage has recently become economic in lots of cases and is ramping up quickly

Being economic and being cheapest are worlds apart.

Solar or wind + utility-scale storage come in at 46 to 102 and 42 to 114 $/MWh, respectively, in terms of LCOE [1]. That does not include grid firming costs [2], which could raise the upper end of those figures to $120 or more, and is based on current storage prices; if everyone tries to build at once, it rises. (On the other hand, there are further economies of scale to be realised.)

Fission clocks in around 141 to 221 $/MWh, which is why we aren’t building it, but $31 at the margin, which is why closing working plants is stupid. SMR focus on lowering capital costs through economies of scale. Fusion by reducing compliance costs. In all likelihood, the solution is fusion SMRs baseloading solar, wind and geothermal energy with peaker industrial processes running during the day. (Hydro can come too.)

[1] https://www.lazard.com/media/2ozoovyg/lazards-lcoeplus-april... slide 2

[2] https://www.gevernova.com/gas-power/applications/grid-firmin...


This is "the solution" on what time scale?

This all definitely sounds like a great way to meet all our energy needs without carbon emissions ... in like 30 years? maybe 20?

We haven't invented "fusion SMRs" yet, let alone commercialized and scaled them. We're still in the extremely early stages of commercializing geothermal at scale. I'm curious what "peaker industrial processes" you're thinking of that can be profitable while running only during high power supply periods?

I think if we build a bunch of batteries, and then they turn out to no longer be useful after all this other stuff scales up, that's totally fine, far worse things have happened!


SMRs are based on the wishful thinking that, for the first time in history, making an industrial facility smaller increases economic efficiency. It's just not going to happen.


I'm pretty skeptical of the SMR thesis, but I think there are plenty of examples of modularization and miniaturization being economically advantageous.

For instance, I just the other day came across an article about how in the first wave of electrifying the textile industry, they replaced big centralized steam engines with big centralized motors. But then they realized that motors made it possible to mass produce small machines and put one at each station, and they turned out to be an advantage.

I think there are a number of specific reasons to be skeptical of SMRs, but I don't think the entire concept is per se flawed.


We also haven't had a big disaster with battery storage yet, which is probably inevitable as the facilites are built out.

A 10MWh battery storage facility, if it were to release its energy all at once would be something on the order of the Chernobyl explosion (sans radioactivity) so certainly capable of destroying a building and killing people nearby. I'm not sure what is "normal" for a utility scale storage facility but 100 times[0] that doesn't seem out of the question

(From Wikipedia[1], the explosion "was estimated ... to be at 40 billion joules" and from unitconverters.net, 40 billion Joules is about 11 MWh[2].)

[0] https://www.energystoragejournal.com/worlds-largest-utility-...

[1] https://en.wikipedia.org/wiki/Chernobyl_disaster

[2] https://www.unitconverters.net/energy/joule-to-megawatt-hour...


That's not a useful comparison. The power of the explosion at Chernobyl, while deadly to the one person immediately next to it, was not the problem that made Chernobyl the catastrophe that it was. That was the radionuclide contamination that it spread and the remaining latent power in the fuel that, if released uncontrollably, would have spread it even further.

A grid scale lithium ion battery, even completely burned up and vaporized into the atmosphere, is not dangerous in a comparable way.


not to mention that battery fires don't release all of their stored energy at once, unlike explosives. Granted, they might burn uncontrollably for a long time, and difficult to extinguish.


Solar and wind are not "unreliable" any more so than any other power source we've used in the past. Just like transformers have been replaced by power electronics in many applications, the reliance on flywheels for frequency stability will be replaced by grid-forming power electronics. There isn't anything magical about this technology.


Another thing that's easy to miss is that, unfortunately, renewables tend to be correlated. An event that reduces insolation over a large area, for example, will affect solar and wind over a large area. So simply over-building is a lot more expensive than it seems when you need to be able to handle tail risks.


> does it even make sense for governments to go down this route?

For past 50 years, we had ["fusion never" level of funding](https://imgur.com/u-s-historical-fusion-budget-vs-1976-erda-...). Because of climate change, there is a sleuth of nuclear startups.

I wouldn't hold my breath for any of the startups. None of them (at least non-state backed ones) seem to have realistic way to the goal.

I remember reading a post from one of startups after rejection from NRC. It read like a blog post after being dumped by a girlfriend written at 3 AM, drunk.

On the other hand, it's not like nuclear is going away, e.g. Uganda and Kenya are planning on nuclear reactors. Maybe we should have a better option to offer than the light water reactors.


No love for Commonwealth Fusion? They seem to have solid backers, technologists & approach.


I’ve been following them since they were giving promising lectures at MIT and I absolutely think they have the most solid approach! Tokamaks are well understood, they supposedly have the same plasma physics as ITER which has been heavily scrutinized and supported by work at JET, and their concept is simple - Tokamak but with very high field superconducting magnets using technology that wasn’t available when ITER was conceived, and apparently higher field strengths mean a smaller reactor for the same power gains. As a lay person the story is simple and that’s good! Then they demonstrated their magnets and got $2B in funding and now they’re deep in the construction phase for SPARC.

I encourage anyone curious to look up videos on SPARC on YouTube. It’s very encouraging! It seems honestly very reasonable that they will see sustained net energy gain for their entire power plant before 2030 (tho SPARC is still a demonstrator not designed for continuous service, so “sustained” means like one minute).

Here’s some videos:

8 years ago:

https://youtu.be/KkpqA8yG9T4

2 years ago:

https://youtu.be/KkpqA8yG9T4

Latest update posted yesterday:

https://youtu.be/w3Giq6NuPYs


As context, they're aiming for first plasma in 2026

https://www.axios.com/pro/climate-deals/2024/05/01/commonwea...


Wonderful, thanks for the context! I knew they originally had plans for mid-decade, but I wasn't sure what their current timeline was.


I have a lot of love for CF. But when I talked to someone who actually knows about the stuff, the business plan of all fusion startups is basically to sell know-how/IP, once a state actor decides to go at it.

That's a good plan, but ultimately, it's going to be a state backed (that's why I have "non-state backed ones" qualifier). CF is going to have a reactor with fusion with Q>1, but commercial product?

China is working on MSR. It has employs something like 700 Phds and 700 support personel for over a decade and has only recently made a research reactor. That's what I consider a serious effort (and that's for far simpler technology).

In my opinion, people underestimate how brutally hard it is to make new technology to work reliably. E.g. Superphenix, sodium cooled reactor had a capacity factor of 7.9% over a decade of production. That was after they had a demo reactor Phoenix with capacity factor 65%.


I think these guys have a viable approach - https://xcimer.energy


> However, with solar and wind now far cheaper than nuclear due to no need for massive capital investments in concrete and steel upfront many years before production starts, does it even make sense for governments to go down this route?

If we would like to stop polluting the air, the future of maritime shipping is nuclear (fusion or fission). China understands that, and invests in R&D necessary to make it happen.

Plus, on ships, there's no competition with solar or wind. And nuclear will actually be quite cheaper than bunker oil, if executed correctly.


I’ll eat my hat if cargo ships go nuclear. Even the US Navy stopped using nuclear for all but carriers. Shipboard nuclear is on another level to regular power plants for many reasons.


This. If you could make ship-sized nuclear reactors easy and affordable, the US navy would be knocking down your door. There's no lack of DoD funding, no lack of operator expertise, and no nimbyism from dolphins, so the fact that the USN doesn't have a reactor in every single Arleigh Burke is purely because it's not economical.


They could put out a big production line of cheaper reactors, but the problem is that their navy boats are a target for missiles, which means larger risks of bad incidents. So they pick carefully

Cargo generally isnt a target in the same way


As much as I'd love to see nuclear powered cargo ships, they do still have to give consideration to the possibility of getting damaged.

Even putting aside exceptional situations like with the Houthis, we tend to get one or two highly public ship accidents per year. It would not be nice to have an incident like that involving a nuke ship every few years.

I feel like the solution for decarbonizing shipping would be carbon capture. Have the ships store the combustion products rather than exhaust them out, then reprocess them back into fuel on land using some other energy source (say, nuclear).


one of the things being looked into is syngas from the carbon in the air, to produce net-zero hydrocarbons.

if it worked out, it would at least be relatively easy since pretty much all our infrastructure is designed around hydrocarbons.


That’s not why they don’t do it, it’s because it’s way too expensive. That fact alone precludes it from being used on cargo ships.

Now ask yourself this: do I want vessels flagged in the countries with the least regulations and the most corruption to be run by a for profit maritime shipping company that skimps on maintenance budgets and crew costs to be running nuclear reactors with highly enriched uranium (weapons grade) anywhere they want around the world, even through pirate territory?

Fuck No I don’t. I barely trust the nukes running them in the USN!



Submarines are boats.


> I’ll eat my hat if cargo ships go nuclear.

Would you like some ketchup or ranch sauce?

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


Go nuclear means the future. Past experiments failing is why I said that ships won’t GO nuclear, implying the future. I was a navy nuke, I know a bit about shipboard nuclear reactors.


Does that count as "going nuclear"? Four have been built, and as of now they've all been decommissioned.


I think it makes a lot of sense. You could probably seal the engine compartment for decades at a time.

I read somewhere that running on bunker fuel was the equivalent pollution of 50m cars.

https://sustainability.stackexchange.com/questions/10757/doe...

I think it was russia? that had nuclear powered ice breakers. Made sense as the constant power demands must be phenomenal.


>I read somewhere that running on bunker fuel was the equivalent pollution of 50m cars.

For SO2 and NO2 pollution, not CO2. They are the most efficient way of transportation in terms of CO2 emissions. Ironically reducing their sulfur dioxide emissions is likely what caused the uptick in global temperatures the last two years. https://www.nature.com/articles/s43247-024-01442-3


I don’t really buy this argument. Maritime alternatives like hydrogen fuel cells and biodiesel seem like far more realistic plays than installing nuclear reactors on thousands of vessels.


Fuel cells don't scale well to multiple megawatts when compared with combustion technologies. Hydrogen is tricky to store. Most likely option is ammonia in steam or gas turbines or large slow ICEs; next most likely option is liquid hydrogen in the same engines.

Biofuels is also severely limited in supply and will in the future most likely be reserved for aviation, which is a lot more constrained than shipping etc. when it comes to which fuel options can be retrofitted on existing systems.


Ammonia is simply nonsense. It's not going to happen for a variety of reasons. Liquid hydrogen is an even bigger nonsense.

Realistic fuels that are being used now: 1. Methanol. 2. Liquid methane.


How difficult would it be to use nuclear power to make synthetic hydrocarbons?


If using electricity, it's "easy", first you split water into hydrogen and then use the Sabatier reaction. Of course, any electricity is fine.

One could get (much) higher efficiency by using the heat from a nuclear power plant directly (never producing electricity) but I guess that would have to be a completely custom design.

https://en.wikipedia.org/wiki/Sabatier_reaction


Nuclear power by itself? It's useless. It can only produce low-grade industrial heat.

If you have spare electricity (from any source), it's easy. Just capture some CO2 and react it with hydrogen with specific catalysts and at a high pressure. You can get methanol directly this way.

It's more expensive than fossil fuels at the current prices, so nobody cares.


Both of your options have significant CO2 emissions, so they are a no-go in just a few years.

Liquid methane is essentially the same as LNG, which is rapidly becoming the most popular fuel for newbuild ships today. But it's about as environmentally friendly as building natural gas powerplant to replace coal - a temporary solution at best.

Future solutions need a carbon-free fuel, period.


There's nothing wrong with CO2 emissions, as long as they remain carbon-neutral. So if you capture carbon dioxide from the atmosphere, and then use it to synthesize methanol or methane, then there are no problems with that.

Methanol is slightly preferred because methane can leak, and it's a more potent greenhouse gas than CO2. However, even most of the CH4 leaks happen near the drilling wells, and in pipelines. It's unlikely that synthetic CH4 will have to be transported over long distances.


Nonsense or not, major companies are literally building ammonia fueled ships right now.

https://gcaptain.com/aet-orders-worlds-first-ammonia-dual-fu...


Sure. Ammonia was used to power buses during the WWII, diesels can burn pretty much anything that burns (within reason). It's not a problem of technical feasibility.

Ammonia fueling infrastructure does not exist, and its failure scenarios are just not going to be acceptable. Meanwhile, LNG fueling infrastructure is rapidly getting built out.

What's worse, ammonia is also produced from natural gas, it's used for process heat and as a hydrogen source. There's pretty much no "green ammonia". So instead of round-tripping through ammonia production, it's easier to just burn the LNG directly.

In future, we can switch to green ammonia, but then we also can use power-to-gas or power-to-methanol instead. Both are more efficient than ammonia synthesis.

Methanol production, in particular, can potentially scale down to very small facilities. In theory, large utility-scale solar or wind farms can have a methanol synthesizer unit, that will produce it when there's more electricity when needed. It can then be transported by regular tanker trucks.


Exactly.

Proliferation will always be a risk with nuclear reactors. We will never have nuclear powered civilian ships, as long as there exist pirates out there. Sure, Russia operates nuclear powered ice-breakers, but there are no pirates in the Arctic Ocean, plus, for Russia the distinction between civilian and military is not all that clear.

As for hydrogen, I think ships are the killer app. High pressure tanks or cryogenic tanks benefit from the square-cube law. If you want them to be economical, they need to be really large. They will never make sense for cars, or even trucks, but they can make sense for trains, and certainly for ships.


> Proliferation will always be a risk with nuclear reactors.

Wasn't one of the promises of thorium reactors a much lower risk of non-proliferation? (Here's a fun question, can one make a pebble bed reactor design with pebbles designed such that if a ship sank, could a special magnetic sphere of a 'correct' size pull in the pebbles but keep a safe distance? IDK but trying to think outside the box here...)

I think it's worth remembering that for the sake of many ships, we do not need the power-density of an SXX or even an AXX per-se.

> As for hydrogen, I think ships are the killer app. High pressure tanks or cryogenic tanks benefit from the square-cube law. If you want them to be economical, they need to be really large. They will never make sense for cars, or even trucks, but they can make sense for trains, and certainly for ships.

The bigger the tank, the more rigorous the inspection has to be to avoid risks due to hydrogen embrittlement.

I'll admit, I'm -less- worried about that property on a train than a ship, but on a ship I think we'd first need to see good evidence we can maintain things of such size on ground safely.


As a lay person, it seems like trains are pretty much always suited to electricity. Adding a power line alongside the existing right of way seems like it’s a pretty straightforward option. What are the conditions in which on-board power storage is preferable?


Hydrogen fuel for merchant ships isn't going to happen. Despite some issues with toxicity and pollution, the industry seems to have settled on ammonia as the main replacement for fossil fuels.

We might actually get more "nuclear powered" civilian ships, in a way. The reactors will be on land, where they can be properly guarded. And the heat and power will be used to manufacture carbon-neutral liquid fuel.


Not to mention modern sail


what modern sail



ehhhhhhhhhhhhhhhhhh?

We gotta remember what a lot of the Marine world really looks like, under the covers.

That is, lots of them will use HFO aka Residual Fuel oil or 'bunker fuel'.

Switching to Biodiesel? Probably the 'cheapest' of the options, not sure what if any implications exist from the switch (lots of ships will stop burning HFO in ports and switch to more common diesel/etc, however not sure if there is a difference in some engines with doing so long term)

Hydrogen Fuel cells are likely as much of a 'refit' from a labor standpoint as switching over to a nuclear reactor; Also the general issues of hydrogen embrittelment and the like have not yet been solved AFAIK especially for the volumes needed for large ships, also not sure if there have been a lot of studies as to whether the hydrogen embrittlement problem could lead to larger structural integrity issues on such a vessel.

Nuclear, OTOH, has had at least a few 'non-military' ships (mostly nuclear icebreakers) with good success.

The current 'whitewashing' strategy of cruise lines is LNG, for whatever -that- is worth...

Edit: finger slipped and hit post too early, so a bit was added, apologies!


Why is it impossible to use wind and solar for ships? I mean, most of our history, ships used wind.


Sails are great, but they are incompatible with the way we load and unload ships now. Ports are designed around unobstructed access from the top. Maybe you could make it work with tankers, but people are risk averse with those. Also sail ships need a lot of crew to handle the sails.

Some shipping companies are experimenting with other ways to use wind. You can deploy kites to pull the ship, but that brings some operational challenges. The more promising idea are probably flettner rotors [1]. Those look like big spinning columns and work on the Magnus effect (how wind puts a 90 degree force on spinning objects). Their limited footprint makes them easy to integrate into existing designs, and since all they do is spin they are easy to use with the small crews of todays ships.

All of those modern ideas are mostly for reducing fuel consumption though, not replacing the engine entirely.

1: https://en.wikipedia.org/wiki/Rotor_ship


I suspect we're quite close to ships switching to wind simply because it's cheaper.

Huge kilometer square kites would be pretty cheap compared to the fuel budget of a ship, and clever routing and control systems can probably mean they reduce fuel consumption 80% for the same travel speed.


> The kite in question has been named Seawing, and may help ships reduce their fuel emissions by between 10 and 40 percent

Not KM but 822m seems pretty close. I think you’re grossly overestimating the benefit from the kite. Seating’s current website says:

> A 1000m² sail surface to harness the power of the wind and tow ships. Based on modelling and preliminary land-based tests, Airseas estimates that the Seawing system can reduce fuel consumption and greenhouse gas emissions by an average of 20%.

I don’t think better routing will increase that to 80% even if you combine it with next gen tech that knows wave patterns and when a slot will be available to minimize speed and energy loss.

We need a path to remove fossil fuels from ships (& planes). There’s also industrial applications that need high heat that solar can’t really accomplish. Finally, solar & wind need insane battery capacity which when included pushes the economics strongly back in favor of fission and fusion.


One idea for high heat industrial requirements is to move those factory locations to places with geothermal power (like Iceland).

We won't see discussions on that until we're serious about cutting fossil fuels.


On its face that seems like a pretty ridiculous suggestion when the alternative is to just build nuclear power plants which don’t have any real geographic considerations and thus can be built next to existing factories that are already built around such things.


Why not hydrogen?


Liquid hydrogen is impossible to work with at large scales, it causes embrittlement, leaks like crazy, has poor volumetric energy density, requires storage in vacuum-insulated tanks, etc.


Molecular hydrogen does not cause embrittlement (neither gaseous, nor liquid). This is a concern in certain chemical reactions that produce atomic hydrogen, but not in any storage applications.


????

It certainly does. The higher the pressure, the worse it gets. And it absolutely applies to storage.

There are companies that sell various technologies for hydrogen-resistant coatings for pipes, for example.


It will have the same proliferation risks.

A fusion reactor is an extremely intense source of neutrons. The neutrons can be used to transmute elements, e.g. to transmute cheap natural uranium or depleted uranium into plutonium 239, which can be separated easily (in comparison with enriching uranium) and it can be used to make nuclear bombs.

Besides producing plutonium for nuclear bombs, it is also easy to use a fusion reactor to produce any kind of dangerous radioactive isotopes that could be used in terrorist activities.

So no, a fusion reactor that uses the fusion reactions that are possible today will not be any safer than a fission reactor, from the point of view of the proliferation risks.


Neutrons are free. You can make a fast neutron source for a few million bucks today. Making a bigger one that can produce usable power is no big deal. The existence of fusion reactors makes absolutely zero difference to the question "who should have fissile material?" You cannot start or stop ignoring that question whether you have or don't have fusion reactors.


> with solar and wind now far cheaper than nuclear ... does it even make sense for governments to go down this route?

If this works without the sun shining then, yes, it makes sense. It is always good to have multiple sources of energy even if only as a form of redundancy. Our world depends on power.


> If this works without the sun shining

HVDC lines are already mature enough that the cheapest route is to just wrap the Earth with them to form a planetary grid.

The sun always shines somewhere.


Is this true? What are the costs per km for HVDC?

Perhaps you're just talking about the Eurasian continent? What do the people of Western Europe do? Connect to the US?

Even then, we've seen with the recent Russia-Ukraine war that control of energy is a useful geopolitical tool with Europe being softer on Russia because of their reliance on their gas.


> reliance on their gas.

... and nuclear fuel and fuel rods from Russia. Which are still not being sanctioned btw. It's peanuts compared to the natural gas, admittedly, on the order of 700 million Euro per year.


Bunch of projects in the works. But building HVDC lines is not cheap.

https://en.wikipedia.org/wiki/List_of_HVDC_projects


Same goes for nuclear power plants really.

A 2.5GW undersea HVDC line costs $2.5mln/km.

The cheapest nuclear power plant in Europe is the Ostrovets power plant in Belarus, the cost of which was $11bln for a 2.4GW plant.

For that money you could buy a 2.5GW HVDC line spanning the entire EU.


If the earth was a uniform sphere without oceans and mountains, sure.


For general power delivery to the grid I think renewables make a whole lot of sense. But for specialty industrial processes that require very large levels of constant power, I think nuclear fusion is very interesting. I worry about environmental effects of mass industrialization but at the same time, I wonder what we could achieve if we had 100x more power available for this or that industrial process. Would it be helpful in decarbonizing steel refining or other metallurgical work?

I think if we develop the technology we will find a use for it and be grateful that we have it, even if it’s hard to predict today what those uses will be.


I am excited for how much bitcoin we could mine if we had a dedicated nuclear fusion plant.


If you mean "we" as in "humanity", then exactly the same amount. Bitcoin is created per unit time, not per unit energy.


I think fusion has one major advantage compared to other renewables.....its much less resource intensive.

While plasma confinement is currently done via supercooling of electromagnets(from last time I was looking into fusion) that's the major resource sink that I can see. We have massive fusion chambers, but I know some universities have built much smaller scale chambers. And we also can address the helium shortage if we solve fusion.

I'm not sure if fusion will ever get solved or if we will she commercial adoption. I also don't know what the life cycle of a fusion plant would be but its got to be cheaper than the big turbine blades, and more ecofriendly the photovoltaic cells.


One can’t power Tokyo (metaphorical) with sunwind


> However, with solar and wind now far cheaper than nuclear

They are not cheaper. They produce very low-quality electricity. If you want them to provide any supply guarantees, their price skyrockets.


It makes a lot of sense, nuclear is nearly 100% reliable. Weather (wind) has wild swings. Solar is -pretty- good but can still swing around a lot and we simply don’t have the grid scale level of batteries that need to smooth it out. I’ve seen estimates that we need battery tech with 10-20x energy density(at current cost levels) what we currently have to make a viable replacement for classical energy sources (coal, natural gas, nuclear)


the cost of solar/wind depends on how much solar/wind is actually deployed.

1kWh of solar delivered midday, when there is 20% penetration? easy peasy.

1kWh of solar delivered at 2AM, when there is 65% penetration? much much more difficult.

These types of price comparisons are always unfair, always apples and oranges, because they always compare a 2AM kWh of nuclear with a midday kWh of solar, and of course solar wins that comparison.


The price of fission will hopefully come down over the next decade via reduced regulations and higher production rates (of smaller reactors).


A bit tiring to see the price of solar and wind being compared to nuclear. Nuclear can produce electricity on demand. Solar and wind cannot. You need to pair them with either some humongous energy storage facilities (and then you need to also over-provision), or some other on-demand source of electricity. Once you factored those costs, then you are not comparing apples and oranges.


Nuclear really isn't anywhere close to 'on demand' at least if you consider unit economics. It really wants to be just 'on' instead.


1) even if it was, over producing electricity isn't really the problem

2) it isn't. Modern reactors are designed to do load following. The French do that nationwide on a daily basis.


Technically you can but you spent however many billion euro and aren’t utilizing the capacity. Maybe it still makes sense vs keeping coal and gas underutilized, I don’t know.


If you allow yourself to use carbon energy (coal and gas) then it absolutely makes sense to use them to compensate for the variability of wind. That's what the UK does. Their cost is pretty much proportional to their utilisation so it makes sense to switch them on and off.

You also have hydro but it's a fairly limited (there are only so many valleys you can flood and so much water you can capture - plus historically it's the source of energy that killed the most people).

But if you truly decarbonise, and in absence of an economical way to store vast amounts of energy for a long time (wind can be down to pretty much zero for weeks on a typical year), and I don't see any such facility being built at scale, I am not sure what else than nuclear you can use to compensate for the volatility of wind. And because nuclear costs the same whether you use it or not, you then might as well save yourself the construction of a wind farm.

That's why I don't understand why we are spending billions building those gigantic wind farms. They only make sense if the intention is to keep using carbon. Otherwise they should spend that money on nuclear.


> However, with solar and wind now far cheaper than nuclear due to no need for massive capital investments in concrete and steel upfront many years before production starts, does it even make sense for governments to go down this route?

Cheaper per watts generated, which aren't constant. Cheaper for a constant output? Reliable to actually power a full grid through downturns such as storms, winters, etc? No, not really. There are exactly zero currently available widely usable grid scale (being able to have enough capacity to power the grid for up to days at a time) solutions. Pumped up hydro is the only one coming close, but it's expensive and it requires specific geography. Just saying "batteries" or "supply and demand by load shedding" doesn't magically solve this problem.


We don't have enough production of basic materials like steel to scale solar and (especially) wind to cover our entire energy needs, regardless of energy storage. Fission and fusion will become inevitable in a decade or two.


Why do we have to make solar panel infrastructure (grilles, consoles etc.) from steel? I'm sure more common materials can be used.


I don’t know if the statement you are replying to is correct “we don’t have enough steel” but what I can say is steel (well, iron) is about the most common material on earth. I’m surprised to see that aluminum is slightly more abundant, but they are similar.

https://en.m.wikipedia.org/wiki/Earth%27s_crust

However this chart shows that iron represents more than 94% of all metals mined. That is, iron (used to make steel) is the most commonly mined metal by far. So actually more common materials can’t be used as no more common metal exists.


Earth is almost entirely iron, but Aluminum floats in iron so it ends up being a large amount of the crust. But yeah, if we're ever like "gee we don't have enough iron" then we've far surpassed all other possible natural resource limits of the planet.


The current production of 1.9 billion tons of steel per year is something you consider insufficient?

I don't know how much steel we need per square meter of PV (e.g. frames can be made from wood), but I do know the area we need for the current global electrical demand of 2 TW even after accounting for capacity factor and not just cell efficiency, and that our current production in each year is sufficient to put a contiguous 2 mm layer behind all of it:

http://www.wolframalpha.com/input/?i=%281.9e9%20tons%20%2F%2...

Given the panels are supposed to last 25 years, even at steady-state replacement rates, and assuming zero growth in the steel sector, and assuming none of that steel gets recycled when the cells themselves need refurbishment or replacement, that doesn't seem to be a real problem to me.


Depends on whether we want to reach a qualitatively different (and better) level of civilization, or at best stay at the current level (but in a carbon-neutral way).


Dumb question, but is the basic idea that you need to harvest more heat energy from the plasma than is needed to maintain the magnetic field?

Also, very dumb question but the plasma means that fusion is actually occuring, right?

And does anyone know how this one collects the heat and converts it into electricity or whatever?

Or any other fusion device, how does it actually collect or output energy from the fusion. And how much do they make, and how far off is that from matching the input power?

Maybe it was some protons escaping from the plasma and hearing something external or something.


1. Yes, sorta, but it's more than just the magnetic field. You're also heating the fuel, so you have to offset that too. Plus there are pumps which circulate coolant to carry heat away from the plasma and towards a turbine, so you have to offset their power. Probably a few other things as well.

2. I don't think plasma == fusion. You can get plasma just by heating a gas beyond a certain point. Plasma cutters, for instance, operate on super heated air, no fusion anywhere nearby.

3. I think the wall of the reaction chamber heats up because they're being bombarded by radiation.

Most of the radiation incident on the reaction chamber walls is infrared, radiated from the hot plasma, but there are also more exotic things like stray neutrons also crash into the sides of the thing. These cause the metal to deteriorate over time (and become somewhat hazardous), but they also they impart additional heat energy.

So you have to have two cooling systems, one to keep the magnets actually cold so they they remain superconducting, and another to keep the housing below the point where it melts. It's this second one that let's you pull heat away from the hot metal donut that is a tokomak and use it to make electricity.

Between the magnet coolant and the chamber coolant and the reacting plasma you have some of the steepest thermal gradients anywhere in the known universe.


Thanks..right I know about plasma in general, I just assumed in this case it was caused by the fusion. Maybe not. But they have fusion right? Just not recovering any/enough energy to make up for power requirements.


The article is light on details. It doesn't mention an operating temperature or Q factor.

I would hazard to guess that no - they did not achieve fusion. They achieved plasma which is a precursor to fusion. Controlled plasma, at a high enough temperature, is an environment in which fusion can occur. All this article says is they created controlled plasma. Crucially, they did so with high temperature magnets which is fairly novel.

https://en.wikipedia.org/wiki/Fusion_energy_gain_factor You might also be interested in reading this. Q factor is what's used to discuss whether a fusion device is generating net positive energy.


No tokamak, even one intended to achieve fusion, would first be operated on D or DT. They'd first extensively test it with ordinary hydrogen.


I doubt they have achieved any fusion reactions. They don't state any numbers on density or temperature so it's impossible to know. But in general plasma is never "caused by" fusion. Creating a plasma is quite easy compared to getting it hot and dense enough to fuse.


I'm musing about the phrase:

> plasma is never "caused by" fusion

Which do you suppose comes first in a gravitational confinement scenario, plasma or fusion? It sorta seems like a chicken/egg scenario. I mean you gotta get those electrons out of the way, but where does he heat come from to do that, if not fusion?


Definitely plasma. Look at star formation in nebulae. Most of the hydrogen in them are in glow mode plasma. It takes a very small amount of energy to ionize plasma and a a huge amount to fuse it. So when stars are forming they might start as cold hydrogen, but they get progressively warmer and more dense until they ionize, then get even warmer and more dense, until they're burning.

https://en.wikipedia.org/wiki/Star_formation


They likely can have some fusion reactions (if they use fusible fuel, like D-D). Fusion is not that hard to achieve, you can do that on a table-top scale (Farnsworth Fusion).


> Dumb question, but is the basic idea that you need to harvest more heat energy from the plasma than is needed to maintain the magnetic field?

No, since creating and maintaining the magnetic field in principle consumes no energy. All the energy put into a superconducting magnet (1/2 L I^2) can be recovered.

What is needed from a physics point of view is for fusion energy production to comfortably exceed the energy put into the plasma. And there's also a whole host of engineering and economic issues beyond that.

Energy is recovered from DT fusion by stopping the neutrons in a blanket, converting their energy to heat, and taking that heat away in a fluid.


> plasma means that fusion is actually occuring, rigth?

As mentioned plasma is just another state of matter[1], where a significant portion of the electrons and ions a separate rather than combined as atoms.

Fusion happens when you overcome the electrostatic repulsion of nuclei, bringing them close enough together so they can fuse[2]. Typically, in reactors like this, that means you confine (compress) a sufficient amount of material ("fuel") to a small volume and heat it up sufficiently. Both are needed to make it possible for the nuclei to come close enough to fuse. The heat required is so great the material will turn into a plasma.

> And does anyone know how this one collects the heat and converts it into electricity or whatever?

This depends somewhat on reactor design, including fuel used. However they're all fancy steam generators in the end, so not unlike a traditional nuclear power plant in that regard.

From what I know, typically the "surplus heat" of a fusion reactor comes in the form of energetic neutron radiation[3]. This radiation is ionizing and as such shielding is required, and this shielding will heat up as it slows down those energetic neutrons.

In the ARC reactor[4] for example, a liquid shielding "blanket" surrounds the fusion chamber. As the neutrons heats up the liquid, the liquid gets pumped through a heat exchanger to produce steam to run a steam turbine.

edit: I found this talk[5] from one of the folks behind ARC to be very illuminating in how fusion power works and the challenges involved. It's from 2017, but the basics haven't changed.

[1]: https://en.wikipedia.org/wiki/Plasma_(physics)

[2]: https://en.wikipedia.org/wiki/Nuclear_fusion#Requirements

[3]: https://en.wikipedia.org/wiki/Neutron_radiation

[4]: https://en.wikipedia.org/wiki/ARC_fusion_reactor

[5]: https://www.youtube.com/watch?v=L0KuAx1COEk


> the plasma means that fusion is actually occuring

No. Plasma simply means a specific state of a matter. E.g. the fluorescent lamps (the long tubular lights that flicker on start) have a plasma inside when it produces light


Your reply implies that in this specific case there is no fusion. I know that plasma can occur without it, but this discussion is about the specific machine.


You make the plasma before any fusion can happen.

Just there being plasma there means nothing, you inject it on the machine already that way.


In the case of this machine it implies that they got plasma by fusion. Which means the fusion is working. It's a milestone, albeit one of many.


You don't ever create plasma via fusion, fusion occurs in plasma that has reached a certain temperature and density threshold.


I dont believe magnetic containment would contain heat, so just run a liquid through the reactor and use it to heat up water to make steam and drive a turbine. Nuclear plants do this.


Well it's a torus right? So you put a turbine in the middle? I don't think I've heard that explanation before.

Or maybe it can go in the outside. I guess it's like, you need a huge amount of electricity to make the magnetic field strong enough, right? So the question is, how do you collect enough heat without melting key components?


No unless you want your turbine to be neutron activated. (You don't.)

You would pump water through the reactor and use a heat exchanger to a secondary water loop which powers the turbine. Maybe you can do without the secondary loop altogether, not sure; this ITER document suggests only one loop, but it's super vague: https://www.iter.org/sci/MakingitWork


No one has figured out how to actually do this yet. Which is why it is vague. The radiation levels and difficulty maintaining the magnetic confinement make this essentially impossible right now.

Another reason why fusion is always 50 years away. It’s really hard (outside of a nuclear bomb or star, anyway).


Great questions.

The difference between energy harvested and the energy necessary to maintain confinement is the difference in denominators of Qscientific and Qengineering. Q is power out / power in.

Qscientific is a figure of merit used to know close to a burning plasma a machine is (how many fusion reactions it can do vs. how many it would need to do to be a working reactor).

Qengineering is power put on the grid / parasitic power needed to keep the machine running. Every electrical power source has an analogous concept (keep the lights on, fuel pumped, inverters operating, etc.) There are some noisy non-experts who claim that focusing on Qplasma is deceitful, but it's akin to complaining that engineers are focusing on engine efficiency instead of car efficiency before the engineers have finished making the engine. At the end of the day the scale of parasitic loads scales much less than the power output of a reactor, so the reactor size chosen will be at the economic minimum between "bigger machine is more expensive to make" and "smaller machine produces less power / lower Qengineering / other difficult scaling law things like neutron bombardment on plasma facing components (maintenance schedule)".

https://x.com/JB_Fusion/status/1506964692627034118

Yes, to have a real measure of Q you need to be doing fusion. In many research cases not a lot of fusion is happening and the neutrons are not actively being measured. What is typically done is to measure plasma performance metrics with protium or deuterium then say what the Q would have been if they used deuterium-tritium based on known plasma-performance to Q conversions (Lawson criterion).

https://en.wikipedia.org/wiki/Lawson_criterion

https://x.com/swurzel/status/1534556521744457731

Heat collection is done via neutrons. In D-T fusion 80% of the energy is released as a 14.1 MeV (17% speed of light, like a bat out of hell). The remaining 20% of energy is an acceleration of a He4 nucleus (fused byproduct). This He4 nucleus is a charged particle, so it stays in magnetic confinement and imparts its energy on fuel via collisions, helping to self sustain the reaction. The neutron has no charge so it flys straight out of the machine. You can model this as a small ring on the innermost core of the donut shooting neutrons in all directions. So you wrap a neutron-absorbing blanket around the vacuum vessel to slow these neutrons down via collision and heat up coolant in the blanket. You run this coolant through a heat exchanger to make pressurized steam to spin a turbine to... you get the idea.

https://en.wikipedia.org/w/index.php?title=Deuterium%E2%80%9...


"HH70 has independent intellectual property rights, with a localization rate exceeding 96%."

what...does that mean?


I would guess that it means that 96% of the components come from within China. Self-sufficiency is important in China right now, and it's doubtful that 'localization' refers to just the company itself.


That means when they inevitably appear on sanction list of US government, they won't have to close the shop.

It's a Chinese project.


Though that depends on what the remaining 4% is. Curious about that. (E.g., for an aircraft the engine being local is more important than the seats being made locally.)


If I were to speculate, I'd say hitting 96% indicates that 100% is an important target to them. That would imply that the remaining 4% are pretty difficult to replicate "in-house" on a short timeframe (though obviously not impossible given the will and time to do so). All speculation though.


Not relying on licenced IP? 96% in-house? That's my guess but I'm just a dude on HN.

Also, this is kinda like SpaceX getting a Falcon 9 to orbit the first time but in fusion land.


It’s a bit difficult to parse the analogy since you’re comparing something that has never been done (and is a notoriously difficult technology to crack) to something that had been done by many others, many times. But, even so, and despite the lack of specific information about the test/achievement, I have a feeling you’re over selling this by quite a bit. If you want to compare to spacex, I’d say it’s more like the first time they demonstrated that they could control a re-entering booster stage with grid fins—a notable step to booster reuse.


The analogy is apt. Many, many, many fusion reactors have achieved first plasma. This is comparable to a rocket achieving orbit.

This company's ultimate goal is commercial fusion power, which has never been done. SpaceX's goal is landing people on Mars, which has never been done. The milestones being discussed are just stepping stones.


> SpaceX's goal is landing people on Mars, which has never been done

Cheap, frequent flights on reüsable rockets would seem to be space’s commercial fusion power threshold. Colonising Mars is like fusion SMRs at a fraction of solar’s cost.


Yeah that makes sense when you explain it like that.


That chinese -> english machine translation still has some way to go.


No, localization rate is the right translation, you just need to understand the context. They've been on a mad dash to domestically source techonology parts and intellectual property, ever since all the sanctions. Foreign suppliers are seen as unreliable now.



"Localization" in many countries means local supply chain. How regional "local" is depends on context...can mean support local community as in farmers market or give jobs to locals, or in projects like this in the more strategic sense i.e. all of supply chain is in country i.e. chinese.


Silly question but say fusion works and we use massive amounts of it. Wouldn’t that increase global temps?


No. Whatever amount of waste heat is released into the atmosphere would be so small compared to insolation that it would be impossible to measure.

If it displaced all remaining coal and natural gas burning, temperatures would stabilize.


No. Assuming it's using steam generators like most powerplants (gas, nuclear, and coal), then the only heat released to the atmosphere would be the inevitable entropic losses. This is the same amount of heat lost by nuclear powerplants (although you could argue nuke plants release decay heat that fusion plants wouldn't but that's negligible). Gas and coal plants release that heat as well, but then also release greenhouse gases that heat things up further


The author wasn’t asking about production entropic losses, but whether using massive amount of energy will heat up the planet, and it definitely will. This is why it’s a technosignature. Heat dissipation is the ultimate limit on growth on a planet, but we’re talking about trillions of people living in luxury.


Yes, but the direct heating effect is quite small compared to the effect of greenhouse gases. The world primary energy consumption is around 19 TW, whereas radiative forcing (difference to pre-industrial values) is estimated to be around 1 PW.


Ah that’s helpful. Numbers give a sense of scale. Thanks


Also note that 19TW is not even mostly heat but useful work: turning motors, pumping fluids or gases, generating light and sound, etc.


This is incorrect. Almost all primary energy is turned into heat eventually. Only a microscopic fraction is turned into light that escapes into space.


The stored potential energy of a concrete block on the 15th floor of a building takes a very long time to turn back into heat.


For someone who doesn't know much about physics, what significance does this have?


I'm brave enough to remember LK99

does anyone know how this differ from outside temperature?


High temperature superconductors don't have to work at room temperature. As it doesn't require liquid nitrogen cooling, it's a lot easier to maintain and run.


Humorously, they don't give a temperature, but they do mention "high-temperature superconducting" 14 times.


If they're hitting 25T on a bore larger than a few cm then they're using supercritical 8K helium to cool ReBCO superconductors OR they have a super secret new superconducting material that the rest of the world doesn't know about and hasn't been used in other practical application (exceedingly unlikely, drunk uncle conspiracy theory tier).


If this isn't a Sputnik and/or an Apollo 11-level wake up call to the western leaders I don't know what has any chance of working.


I think "western leaders" should be more worried about another thing: Constituents(?) exhibiting totally uncritical acceptance of a literal corporate press-release.


I don’t have a problem with that if it results in funding fusion research properly for once


Except you won't get to that point, because first the crowd will be incited against it on the basis that fusion power is heretical against the Sun. God and also causes cancer and migraines and belly-button lint.


I don't understand. They built a thing. We also have similar things, don't we?

Wouldn't the Sputnik moment require actual energy generation? It doesn't sound like they're any closer than we are.


> We also have similar things, don't we?

Yes. Commonwealth Fusion and MIT are building a superconducting fusion reactor at Devens, Massachusetts right now. It's called SPARC and the site has been under construction since 2021. The plan expects to achieve first plasma sometime in 2026.


It's impressive but there are multiple startups in the US working on fusion power.


With currently half the western population pushing a significant anti-science agenda (even greater than half if you consider that there’s also not insignificant anti-science ideals in various left wing groups as well, albeit not usually to the point of ripping kids out of education), that seems like a nearly impossible proposition unless there’s a significant political awakening.


The space race has shown that America hates losing more than it hates science. We're kind of early but e.g. the race back to the Moon is on. If there's a credible threat of losing a race to fusion power I'm pretty sure US politicians would be able to sell that threat to the public in a single afternoon.


I'm not sure why you are being downvoted. The original question is political in nature and polls run by both major sides of the political spectrum in the USA support your argument.


Anything political or feeding the flames is not encouraged on HN. Especially if it appears bipartisan, personally. And many of us are in other countries so the whole subject is usually annoying.


Why?

Assuming this is real and not exaggerated propaganda, does China think IP theft is a one way street?


> Why?

IP theft is a thing and yet China can't make Nvidia GPUs and I can bet $10 it won't be able to in 2030. I don't see why the west could 'just' copy a Chinese energy-positive tokamak even if it had all the plans. (Yes I know this one isn't that.)

The wake up call is for the west to be able to do that at the very least.


To be fair the US also can't make Nvidia GPUs and neither can anyone outside of Taiwan. Agreed that IP isn't everything but the chip shortage during covid sure as hell WAS a wake up call to the west.

Now they are finding that actually its going to take a decade to reproduce what the chip fabs in Taiwan have built even with their help.


Fusion is never, ever going to be economical. The fuel is basically free, which is great. Meanwhile, the reactors themselves are arguably the most complex and expensive machines ever built, and they are essentially disposable due to the nature of fusion reactions.

There's a reason that the wise engineers who built our only working fusion reactor put it about 1 AU away from us. Much cheaper and easier to just catch the energy it sends us.


Given the impact the world has already seen because we let two companies tie up LiFePO4, after we let a few companies tie up other battery patents for hybrids before that...

TBH I would judge the world if they just went ahead and 'stole' it vs RAND licensing...

At the same time, I can see it being one hell of a hypothetical 'carrot' for lots of things, and of the current major powers, China is the only one with enough overall (political+humanpower+etc) will (at this time, anyway) to possibly make Fusion happen sooner than ITER can.

Strategically speaking, it would 'make sense' for them to pursue... Would the European union force NL's hand, to make ASML sell machines for whatever comes after EUV, in exchange for Fusion tech? Or all sorts of other fun things for the right Q factor?

Things become murkier.


And here’s why it’s irrelevant and inconsequential…


If you have a substantive point, please make it thoughtfully; if not, please don't comment until you do.

https://news.ycombinator.com/newsguidelines.html


Please do go on? I won't say irrelevant, but when compared to SPARC project, it seems kind of underpowered?

HH70: major radius: 0.75 m, magnetic field 0.6 T SPARK: major radius: 1.85 m, magnetic field 12.2 T

HH70 has the advantage of actually existing and working, but to my completely layman eyes, it doesn't seem that using high temperature superconducting magnets brought expected increase in parameters.


The real hair-raising thing is that HH70 is out of the blue. Three years is a very short period of time to put together a company, supply chain, and working HTS tokamak of any scale. The plasma performance metrics are nearly irrelevant. They've knocked down a lot of difficult questions and paved the way for bigger machines on short timescales. I'm not sure what the next out-of-the-blue headline will be in 3 years, but there's a good chance it will be "burning plasma". The West runs a real risk of being left way behind here.


> e real hair-raising thing is that HH70 is out of the blue.

I wish them best of luck and China speed. It doesn't matter who develops the technology, in either case it's a win for humanity. 7 out of 8 billion people are not in the "west".


I think this is a reference to the normal pattern in the tortuously slow development of practical fusion power. Despite all of the significant milestones, fusion remains about 20 years down the road, for the last 50 years.


Considering the funding for past 50 years has been below "fusion never" level, I think they made a great progress.

See fusion budget vs expected timelines: https://imgur.com/u-s-historical-fusion-budget-vs-1976-erda-...


Wow, I never imagined that fusion funding was that paltry. Considering the insane things that have to be built to make it work, it is very impressive what has actually gotten done.


To be fair, that's budget for magnetic confinement fusion. US has always been more interested in inertial fusion (i.e. shoot it with lasers). Likely because of synergy with military application of lasers.

The thing it, inertial confinement seems to be a dead end and has been for quite a while. At least rest of the world has decided to fund magnetic confinement (plus few oddballs with z-pinch), so I assume it's more promising approach.


It seems that the investment in fusion is incredibly tiny relative to the potential payoff and compared to other trivial or even destructive pursuits?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: