Since the article opening sentence and headline don’t say it:
The breakthrough is the plasma “triple product,” literally just the plasma temperature (in keV) times particle density times (confinement) time. The Lawson Criterion. https://en.wikipedia.org/wiki/Lawson_criterion
A useful fusion power plant needs a triple product of at least about 3e12 keV * s * m^-3.
They weren’t fusing things (at least, not much). This is a figure of merit that allows you to compare, across all the different fusion methods, how well you would be able to fuse the plasma if you were using burnable fuel such as deuterium and tritium (isotopes of hydrogen that have one or two extra neutrons).
So on this graph they're at about 0.2e20, but it also says they need 3e21 (and the graph on Wikipedia agrees)... So are they 150x off the target? 3e12 is a typo I guess?
Yeah it's 3E21 in SI units. The wikipedia graphs also highlight how nonlinear performance scales across machines. W7-X isn't 1000x larger than T3, yet performs ~1000x better. Confinement field strength (a little expensive) and major radius (very expensive) are knobs that turn these from research machines into power plants.
I love these guys and gals. Just knocking down the engineering goals one after another. It's been a lot of fun watching their progress over the years. If they told me "we're going to build a energy producing stellarator in 5 years" I'd actually believe them. :-).
But they somehow avoid making a visible cooperation reference on the main page, like they do for other research institutions like Jülich. (Or I missed it.) But sure, they definitely reference it in the text, also on the main page.
Kudos to the science but also the artistic beauty of that view inside the vacuum vessel. I can't even fathom the engineering that produced that structure.
I have a feeling ASI will follow similar trajectory as fusion, with the critical intelligence explosion always 2 years away. AGI by Turing’s definition is here. But fusion my whole life has been just around the corner…
Is it? AI is impressive and all, but i don't think any of them have pased the Turing test, as defined by Turing (pop culture conceptions of the Turing test are usually much weaker than what the paper actually proposes), although i'd be happy to be proven wrong.
I have a rather specialized interest in and obscure subject but one which has a physical aspect pretty much any person can relate to/reason about, and pretty much every time I try to "discuss" the specifics of it w/ an LLM, it tells me things which are blatantly false, or otherwise attempts to carry on a conversation in a way which no sane human being would.
> pop culture conceptions of the Turing test are usually much weaker than what the paper actually proposes
I've just read the 1950 paper "Computing Machinery and Intelligence" [1], in which Turing proposes his "Imitation Game" (what's now known as a "Turing Test"), and I think your claim is very misleading.
The "Imitation Game" proposed in the paper is a test that involves one human examiner and two examinees, one being a human and the other a computer, both of which are trying to persuade the examiner that they are the real human; the examiner is charged with deciding which is which. The popular understanding of "Turing Test" involves a human examiner and just one examinee, which is either a human or a computer, and the test is to see whether the examiner can tell.
These are not identical tests -- but if both the real human examinee and the human examiner in Turing's original test are rational (trying to maximise their success rate), and each have the same expectations for how real humans behave, then the examiner would give the same answer for both forms of the test.
Aside: The bulk of this 28-page paper anticipates possible objections to his "Imitation Game" as a worthwhile alternative to the original question "Can machines think?", including a theological argument and an argument based on the existence of extra-sensory perception (ESP), which he takes seriously as it was apparently strongly supported by experimental data at that time. It also cites Helen Keller as an example of how learning can be achieved through any mechanism that permits bidirectional communication between teacher and student, and on p. 457 anticipates reinforcement learning:
> We normally associate punishments and rewards with the teaching process. Some simple child-machines can be constructed or programmed on this sort of principle. The machine has to be so constructed that events which shortly preceded the occurrence of a punishment-signal are unlikely to be repeated, whereas a reward-signal increased the probability of repetition of the events which led up to it.
Not really - at least at current goals, population size, etc -, even with the very high energy expenditure of, say, LOTS of AI hardware to run the "Skynet" we're driving ourselves into, we're talking the order of magnitude of 30TWh (human generation today per Wikipedia).
Imagining a future: with a ~3% growth, so let's say fusion is deployed and everything is electric in the next few years (not happening that fast though), with AI data-centers everywhere so individual-level AI runs (say "Her"'s movie personal OS-level stuff) per human and we reach the out-of-my-buttocks figure of 500 TWh/year in 10 years time, which is crazy shit ... well, that would not "boil the world"!
The Sun delivers ~170,000 TWh per year. So 500 TWh still would not be that significant, and within the Sun's yearly delivery fluctuations.
The problem with energy generation today is that it's releasing gases, and these gases are disrupting the planet’s energy balance - especially how Earth gets rid of the massive energy it receives from the Sun. We do need to restore the balance between what comes in and what goes back out - fusion can help tackle that problem specifically, so it's beneficial overall even if it eventually adds a fractional percentage to the overall planetary energy bill.
I picture that fusion would be a complementary source, not the only one, and, once/if deployed, would help close some of key the loopholes that prevent solar (and other renewables) from being deployed 100%.
You can stop worrying, because fusion energy from this kind of reactor will be anything but cheap. It will likely be more expensive than energy from current generation fission power plants.
The goal posts on AGI would be superluminal and somewhere back in the 1400s if they were physical objects. I’ve never seen or heard of a field so deeply in denial about its progress.
For every major trouncing of criterion we somehow invent 4 or 5 new requirements for it to be “real” AGI. Now , it seems, AGI must display human level intelligence at superhuman speed (servicing thousands of conversations at once), be as knowledgeable as the most knowledgeable 0.1% of humans across every facet of human knowledge, be superhumanly accurate, perfectly honest, and not make anything up.
I remember when AGI meant being able to generalize knowledge over problems not specifically accounted for in the algorithm… the ability to exhibit the “generalization” of knowledge, in contrast to algorithmic knowledge or expert systems. It was often referred to as “mouse level” or sometimes “dog-level” intelligence. Now we expect something vastly more capable than any being that has ever existed or it’s not “AGI” lmfao. “ASI” will probably have to solve all of the world’s problems and bring us all to the promised land before it will merit that moniker lol.
"I remember when AGI meant being able to generalize knowledge over problems not specifically accounted for in the algorithm… "
So do we have that?
As far as I know, we just have very, very large algorithms (to use your terminology). Give it any problem not in the training data and it fails.
> “ASI” will probably have to solve all of the world’s problems and bring us all to the promised land before it will merit that moniker lol.
People base their notions of AI on science fiction, and it usually goes one of two ways in fiction.
Either a) skynet awakens and kills us all or
B) the singularity happens, AI get so far ahead they become deities, and maybe the chosen elect transhumanists get swept up into some simulation that is basically a heavenly realm or something.
So yeah, bringing us to the promised land is an expectation of super AI that does seem to come out of certain types of science fiction.
Same question got asked of Bob Bussard when he visited Google to talk about his whiffle-ball design. It's not that we lack models, it's that they're effectively incomputable at the scale we'd need them to be.
In a fluid, effects are local: a particle can only directly effect what it is in direct contact with.
In a plasma, every particle interacts with every other. One definition of a plasma is that the motion is dominated by electromagnetic effects rather than thermodynamic: by definition, if you have a plasma, the range of interactions isn't bounded by proximity.
This doesn't apply quite so much to (e.g.) laser ignition plasmas, partly because they're comparatively tiny, and partly because the timescales you're interested in are very short. So they do get simulated.
But bulk simulations the size of a practical reactor are simply impractical.
Putting a bunch of much more viscous radioactive material within proximity of each other is simpler than squishing and maintaining confinement of plasma under extreme conditions.
Fission reactors are relatively "easier" to simulate as giant finite element analysis Monte Carlo simulations with roughly voxels of space, i.e., thermal conductivity, heat capacity, etc. I happened to have been involved with one that was 50+ years old that worked just fine because of all of physicists and engineers who carefully crafted model data and code to reflect what would be likely to happen in reality when testing new, conventional reactor designs.
The problems with fusion are many orders-of-magnitude more involved and complex with wear, losses, and "fickleness" compared to fission.
Thus, experimental physics meeting engineering and manufacturing in new domains is expensive and hard work.
Maybe in 200 years there will be a open source, 3D-printable fusion reactor. :D
The difficulty is in the details. Small differences lead to bigger differences, like in chaos theory [0] What if the model says this coil needs to be 23.1212722 centimeter? Or two coils need to be 37.1441129 centimeters apart? How do you build that? Mathematics is always much more precise than engineering.
Maybe stellarators will be the common design in 2060 once fabrication tech has improved, but for the near future I think its going to be one of the first two.
It'll be really funny if we get a commercial fusion device before ITER has even been turned on.
I'm sure they developed some really useful technology in the process of building the thing, but I suspect they would have made more progress faster if they had taken a more iterative approach.
How isolated are the paths it could have taken? One major outdated choice ITER took is cold magnets, rather than "warm". Could they have switched this point in an iterative project?
I doubt Helion will work. By their own paper using simplified model their device allows theoretically for Q no more than 2 (2 times energy produced versus energy spent). They claim with their like 90% efficient capture they still get net energy gain. But typically reality way messier than model and I will be surprised if they archive Q=1.
Tokamaks main problem is plasma instabilities. While Commonwealth may archive high Q briefly, nobody knows how plasma will behave at those conditions and long operations may not be possible.
Stellarators on the other hand do not have plasma stability problems. So my bet is on those.
If the goal is viable commercial operation, Helion has vast benefits over the other approaches when it comes to the economics of turning the fusion energy into electricity.
All approaches have huge hurdles to overcome. Helion may have bigger challenges on the Q side, but all-in-all I think the probabilities of being viable ends up similar.
All other fusion power plants are thermal power plants. I suspect all thermal power plants will end up being economically unviable in the world of renewables, for various reasons. They’re just too bulky and slow, and require special consideration wrt cooling. It’s one of the reasons why gas power is king these days.
If we think really far ahead, the scaling of thermal power plants is limited by the heat they put out. It ends up contributing to global warming just from the thermal forcing they apply to the environment. The effect of the ones we have today are already surprisingly significant. Helion is a path to being able to produce a huge amount of energy with fairly limited impact on the environment (eventually limited by the thermal energy they dump, but perhaps they can use thermal radiation panels that dump the waste heat energy directly to space)
They've already demonstrated very high efficiency recovery of the plasma energy that they inject when the plasma is compressed. The additional plasma energy from fusion products would be recovered just the same way.
Good list, I'm also keeping an eye on Tri Alpha Energy and First Light Fusion. TAE recently announced [1] initiating a field reversed configuration with no plasma injectors, only neutral beam injection, which is a pretty big deal in simplifying the design.
Thea Energy is working on a stellarator that doesn't require the complex shaping coils that W-7X is using. I'd put them above Helion and below CFS, but in a couple years they might take the top spot.
They had to abandon their first approach because it didn't work in at least three different ways. And now they've had layoffs because additional funding didn't show up.
Since the article opening sentence and headline don’t say it: The breakthrough is the plasma “triple product,” literally just the plasma temperature (in keV) times particle density times (confinement) time. The Lawson Criterion. https://en.wikipedia.org/wiki/Lawson_criterion
A useful fusion power plant needs a triple product of at least about 3e12 keV * s * m^-3.
They weren’t fusing things (at least, not much). This is a figure of merit that allows you to compare, across all the different fusion methods, how well you would be able to fuse the plasma if you were using burnable fuel such as deuterium and tritium (isotopes of hydrogen that have one or two extra neutrons).
So on this graph they're at about 0.2e20, but it also says they need 3e21 (and the graph on Wikipedia agrees)... So are they 150x off the target? 3e12 is a typo I guess?
Yeah it's 3E21 in SI units. The wikipedia graphs also highlight how nonlinear performance scales across machines. W7-X isn't 1000x larger than T3, yet performs ~1000x better. Confinement field strength (a little expensive) and major radius (very expensive) are knobs that turn these from research machines into power plants.
Here's the statement on the official website if you prefer that:
https://www.ipp.mpg.de/5532945/w7x
Archive and alternate source:
https://archive.is/OHy4l
https://phys.org/news/2025-06-wendelstein-nuclear-fusion.htm...
I love these guys and gals. Just knocking down the engineering goals one after another. It's been a lot of fun watching their progress over the years. If they told me "we're going to build a energy producing stellarator in 5 years" I'd actually believe them. :-).
Give them the funding and they would love to start ..
But there is a german fusion startup about to build a stellarator.
https://www.proximafusion.com/about
(I assumed there was some sort of cooperation with Wendelstein, but found no mentioning of such on a quick look now)
They do cooperate and Jorrit Lion, co-founder of Proxima Fusion, was studying in Greifswald for Wendelstein 7-X, also see https://www.sciencedirect.com/science/article/pii/S092037962...
But they somehow avoid making a visible cooperation reference on the main page, like they do for other research institutions like Jülich. (Or I missed it.) But sure, they definitely reference it in the text, also on the main page.
Maybe there are complicated legal reasons for it.
Kudos to the science but also the artistic beauty of that view inside the vacuum vessel. I can't even fathom the engineering that produced that structure.
I have a feeling ASI will follow similar trajectory as fusion, with the critical intelligence explosion always 2 years away. AGI by Turing’s definition is here. But fusion my whole life has been just around the corner…
> AGI by Turing’s definition is here
Is it? AI is impressive and all, but i don't think any of them have pased the Turing test, as defined by Turing (pop culture conceptions of the Turing test are usually much weaker than what the paper actually proposes), although i'd be happy to be proven wrong.
I have a rather specialized interest in and obscure subject but one which has a physical aspect pretty much any person can relate to/reason about, and pretty much every time I try to "discuss" the specifics of it w/ an LLM, it tells me things which are blatantly false, or otherwise attempts to carry on a conversation in a way which no sane human being would.
Did you specifically prompt it to pretend to be a person with limbs and all?
> pop culture conceptions of the Turing test are usually much weaker than what the paper actually proposes
I've just read the 1950 paper "Computing Machinery and Intelligence" [1], in which Turing proposes his "Imitation Game" (what's now known as a "Turing Test"), and I think your claim is very misleading.
The "Imitation Game" proposed in the paper is a test that involves one human examiner and two examinees, one being a human and the other a computer, both of which are trying to persuade the examiner that they are the real human; the examiner is charged with deciding which is which. The popular understanding of "Turing Test" involves a human examiner and just one examinee, which is either a human or a computer, and the test is to see whether the examiner can tell.
These are not identical tests -- but if both the real human examinee and the human examiner in Turing's original test are rational (trying to maximise their success rate), and each have the same expectations for how real humans behave, then the examiner would give the same answer for both forms of the test.
Aside: The bulk of this 28-page paper anticipates possible objections to his "Imitation Game" as a worthwhile alternative to the original question "Can machines think?", including a theological argument and an argument based on the existence of extra-sensory perception (ESP), which he takes seriously as it was apparently strongly supported by experimental data at that time. It also cites Helen Keller as an example of how learning can be achieved through any mechanism that permits bidirectional communication between teacher and student, and on p. 457 anticipates reinforcement learning:
> We normally associate punishments and rewards with the teaching process. Some simple child-machines can be constructed or programmed on this sort of principle. The machine has to be so constructed that events which shortly preceded the occurrence of a punishment-signal are unlikely to be repeated, whereas a reward-signal increased the probability of repetition of the events which led up to it.
[1]: https://archive.org/details/MIND--COMPUTING-MACHINERY-AND-IN...
If we ever have access to unlimited cheap energy, then we are going to boil the world.
Not really - at least at current goals, population size, etc -, even with the very high energy expenditure of, say, LOTS of AI hardware to run the "Skynet" we're driving ourselves into, we're talking the order of magnitude of 30TWh (human generation today per Wikipedia).
Imagining a future: with a ~3% growth, so let's say fusion is deployed and everything is electric in the next few years (not happening that fast though), with AI data-centers everywhere so individual-level AI runs (say "Her"'s movie personal OS-level stuff) per human and we reach the out-of-my-buttocks figure of 500 TWh/year in 10 years time, which is crazy shit ... well, that would not "boil the world"!
The Sun delivers ~170,000 TWh per year. So 500 TWh still would not be that significant, and within the Sun's yearly delivery fluctuations.
The problem with energy generation today is that it's releasing gases, and these gases are disrupting the planet’s energy balance - especially how Earth gets rid of the massive energy it receives from the Sun. We do need to restore the balance between what comes in and what goes back out - fusion can help tackle that problem specifically, so it's beneficial overall even if it eventually adds a fractional percentage to the overall planetary energy bill.
I picture that fusion would be a complementary source, not the only one, and, once/if deployed, would help close some of key the loopholes that prevent solar (and other renewables) from being deployed 100%.
Right on! The energy production is not that interesting. It's the energy *blanket* we're making with the byproducts that's cooking us.
So it would seem, but we could also air condition the world and beam the energy into deep space instead.
Why beam it anywhere?
Why not capture and make use of it?
Isn't that the whole point of heat pumps? Grab energy from one locale, move it another to do useful work?
You need a temperature differential.
You can stop worrying, because fusion energy from this kind of reactor will be anything but cheap. It will likely be more expensive than energy from current generation fission power plants.
We'll be smart enough by the time we get it.
We have effectively unlimited carbon dioxide right now
The goal posts on AGI would be superluminal and somewhere back in the 1400s if they were physical objects. I’ve never seen or heard of a field so deeply in denial about its progress.
For every major trouncing of criterion we somehow invent 4 or 5 new requirements for it to be “real” AGI. Now , it seems, AGI must display human level intelligence at superhuman speed (servicing thousands of conversations at once), be as knowledgeable as the most knowledgeable 0.1% of humans across every facet of human knowledge, be superhumanly accurate, perfectly honest, and not make anything up.
I remember when AGI meant being able to generalize knowledge over problems not specifically accounted for in the algorithm… the ability to exhibit the “generalization” of knowledge, in contrast to algorithmic knowledge or expert systems. It was often referred to as “mouse level” or sometimes “dog-level” intelligence. Now we expect something vastly more capable than any being that has ever existed or it’s not “AGI” lmfao. “ASI” will probably have to solve all of the world’s problems and bring us all to the promised land before it will merit that moniker lol.
"I remember when AGI meant being able to generalize knowledge over problems not specifically accounted for in the algorithm… "
So do we have that? As far as I know, we just have very, very large algorithms (to use your terminology). Give it any problem not in the training data and it fails.
> “ASI” will probably have to solve all of the world’s problems and bring us all to the promised land before it will merit that moniker lol.
People base their notions of AI on science fiction, and it usually goes one of two ways in fiction.
Either a) skynet awakens and kills us all or
B) the singularity happens, AI get so far ahead they become deities, and maybe the chosen elect transhumanists get swept up into some simulation that is basically a heavenly realm or something.
So yeah, bringing us to the promised land is an expectation of super AI that does seem to come out of certain types of science fiction.
I have a question. How come the mathematical modeling and simulations haven't yet yielded us the perfect design that will get things right?
How come we have to build it and test it to know if it works?
Do we lack a mathematical model?
Same question got asked of Bob Bussard when he visited Google to talk about his whiffle-ball design. It's not that we lack models, it's that they're effectively incomputable at the scale we'd need them to be.
In a fluid, effects are local: a particle can only directly effect what it is in direct contact with.
In a plasma, every particle interacts with every other. One definition of a plasma is that the motion is dominated by electromagnetic effects rather than thermodynamic: by definition, if you have a plasma, the range of interactions isn't bounded by proximity.
This doesn't apply quite so much to (e.g.) laser ignition plasmas, partly because they're comparatively tiny, and partly because the timescales you're interested in are very short. So they do get simulated.
But bulk simulations the size of a practical reactor are simply impractical.
Putting a bunch of much more viscous radioactive material within proximity of each other is simpler than squishing and maintaining confinement of plasma under extreme conditions.
Fission reactors are relatively "easier" to simulate as giant finite element analysis Monte Carlo simulations with roughly voxels of space, i.e., thermal conductivity, heat capacity, etc. I happened to have been involved with one that was 50+ years old that worked just fine because of all of physicists and engineers who carefully crafted model data and code to reflect what would be likely to happen in reality when testing new, conventional reactor designs.
The problems with fusion are many orders-of-magnitude more involved and complex with wear, losses, and "fickleness" compared to fission.
Thus, experimental physics meeting engineering and manufacturing in new domains is expensive and hard work.
Maybe in 200 years there will be a open source, 3D-printable fusion reactor. :D
The difficulty is in the details. Small differences lead to bigger differences, like in chaos theory [0] What if the model says this coil needs to be 23.1212722 centimeter? Or two coils need to be 37.1441129 centimeters apart? How do you build that? Mathematics is always much more precise than engineering.
[0] https://en.wikipedia.org/wiki/Edward_Norton_Lorenz#Chaos_the...
You need to think of what happens if it's 0.001 cm too big, small, etc. Manufacturing always involves errors and engineering requires tolerances.
Doesn’t our engineering account for that? We never build things that need that level of detail in our tolerances. Like for example an airplane.
OP "How come the mathematical modeling and simulations haven't yet yielded us the perfect design that will get things right?"
We still have wind tunnels and aerodynamics is a pretty simple problem compared to fusion.
It seems like the ranking of likely success in the next 10 years is
1. Commonwealth (tokamak w/ high temp superconducting magnets)
2. Helion (field reversed configuration, magnetic-inertial, pulsed) ....
?. Wendelstein (stellarator)
Maybe stellarators will be the common design in 2060 once fabrication tech has improved, but for the near future I think its going to be one of the first two.
It'll be really funny if we get a commercial fusion device before ITER has even been turned on.
I'm sure they developed some really useful technology in the process of building the thing, but I suspect they would have made more progress faster if they had taken a more iterative approach.
How isolated are the paths it could have taken? One major outdated choice ITER took is cold magnets, rather than "warm". Could they have switched this point in an iterative project?
this discounts the likelihood that other breakthroughs are correlated with ITER in some way.
The first transistor in Silicon Valley wasn’t made by Shockley.
I doubt Helion will work. By their own paper using simplified model their device allows theoretically for Q no more than 2 (2 times energy produced versus energy spent). They claim with their like 90% efficient capture they still get net energy gain. But typically reality way messier than model and I will be surprised if they archive Q=1.
Tokamaks main problem is plasma instabilities. While Commonwealth may archive high Q briefly, nobody knows how plasma will behave at those conditions and long operations may not be possible.
Stellarators on the other hand do not have plasma stability problems. So my bet is on those.
If the goal is viable commercial operation, Helion has vast benefits over the other approaches when it comes to the economics of turning the fusion energy into electricity.
All approaches have huge hurdles to overcome. Helion may have bigger challenges on the Q side, but all-in-all I think the probabilities of being viable ends up similar.
All other fusion power plants are thermal power plants. I suspect all thermal power plants will end up being economically unviable in the world of renewables, for various reasons. They’re just too bulky and slow, and require special consideration wrt cooling. It’s one of the reasons why gas power is king these days.
If we think really far ahead, the scaling of thermal power plants is limited by the heat they put out. It ends up contributing to global warming just from the thermal forcing they apply to the environment. The effect of the ones we have today are already surprisingly significant. Helion is a path to being able to produce a huge amount of energy with fairly limited impact on the environment (eventually limited by the thermal energy they dump, but perhaps they can use thermal radiation panels that dump the waste heat energy directly to space)
They've already demonstrated very high efficiency recovery of the plasma energy that they inject when the plasma is compressed. The additional plasma energy from fusion products would be recovered just the same way.
Proxima Fusion builds a stellator. As far as I remember they were founded by wendelsteinians.
They are the only fusion startup I know of that was faster than their own timeline in the last year.
4. Acceleron (muon catalyzed)
There's huge advantages to muon catalyzed if they can get it to work. Plants would be orders of magnitude smaller and cheaper to build.
[0] https://www.acceleron.energy/
Good list, I'm also keeping an eye on Tri Alpha Energy and First Light Fusion. TAE recently announced [1] initiating a field reversed configuration with no plasma injectors, only neutral beam injection, which is a pretty big deal in simplifying the design.
[1] https://tae.com/tae-technologies-delivers-fusion-breakthroug...
Thea Energy is working on a stellarator that doesn't require the complex shaping coils that W-7X is using. I'd put them above Helion and below CFS, but in a couple years they might take the top spot.
what about general fusion?
https://en.wikipedia.org/wiki/General_Fusion
They had to abandon their first approach because it didn't work in at least three different ways. And now they've had layoffs because additional funding didn't show up.
Aren’t they running out of money?