"zeolite based cat litter?" As in carbon sequestration zeolite? I mean, cat poo as a carbon sink? (Lights go on, then off, then on, then ...)
A New Nuclear Technology Discussion Thread
- Thread starter UserJoe
- Start date
Started to hear more about this design. I don't recall paying particular attention to it previously but it seems like a really good balance of size, passive safety, and low novelty. A good set of tradeoffs if you want to get something built ASAP.
The BWRX-300 hasn't had as much publicity as some of the other SMRs because it hasn't gotten DOE money to develop it. There does seem to be a fair amount of interest in building them. It looks like Poland and TVA in the US are going to build them in addition to Ontario Power Generation in Canada. The technology for the design has a high level of technological maturity and it will use standard BWR fuel that could be procured from multiple vendors. There was a recent announcement that BWXT Canada will build the reactor pressure vessels for the BWRX-300.
OPG was also looking at building X-Energy Xe-100 high temperature helium cooled reactor before they selected the BWRX-300. It looks like the first Xe-100 will be built by DOW in the Gulf Coast region. They will use the 4 Xe-100 units for a combination of process heat and electricity.
This is intuition more than anything else but I think I've shifted to the view that a lot of the SMRs went too small. Whereas some of the other sub-GW reactors didn't save enough on the construction costs. It seems like the BWRX-300 is really at a sweet spot because the interest in construction seems pretty good for a new reactor. I mean that's not a huge number of firm orders but it's still pretty damn good compared to the state of the rest of the industry. In particular my layperson's impression is that they have done a very good job of reducing the civil construction required while still achieving a good level of passive safety and making good use of existing technology.
Also, and correct me if I'm wrong here, but it seems like the BWR technology is well suited to adaptation to SMRs in ways I had not recognized previously. I remember previous discussion on natural circulation vs relying on pumps to get better power density. But in a core that's designed to boil continuously you can just let the phase separation do that job. And then in shutdown the core that's designed to boil water just keeps boiling water. As long as the condensers can keep up, which they easily can because they are oversized, the water will be condensed fast enough to keep the pressure under control.
Natural circulation BWRs get close to the same power density as pumped flow BWRs. The RPVs need to be taller to get enough natural circulation driving head to provide the core flow. You get more driving head in a natural circulation BWR than you do in a natural circulation PWR for the same height because of the bigger density difference of the two-phase mixture on the hot side of the flow loop in the BWR compared to just hotter water in the PWR. I don't think BWRs are necessarily better suited for being SMRs than PWRs are. PWRs can have passive natural circulation heat exchangers too.
GE sold small natural circulation BWRs back in the 60s that was about 60 MWe. There weren't very many built. One of them was at Humboldt Bay in California. I think if the first BWRX-300 gets built at close to the build time and cost there could be a lot of demand for them. There's a lot of coal plants that will be retiring and those sites can be repowered with SMRs. The grid connections are already there, and they might even be able to reuse the generator sets and save money that way.
Last edited:
Those sites would still need NRC licensing, though. That'll be a huge time and cost burden, unless GE can convince them that the BWRX-300 is so safe that siting isn't a big deal.
There is a significant amount of work to get early site permits but the process has been exercised multiple times now and it is now a known and tested process. There is the matter of the environmental impact in addition to the safety analysis since the plants have to meet National Environmental Policy Act (NEPA) requirements.
Last edited:
It's been a long time coming but, Vogtle Unit 3 was connected to the grid on Saturday April 1st. It will now go through testing is supposed to go into full commercial operation in May or June. Vogtle 4 started hot testing recently and is supposed to go into full commercial operation by the of this year or beginning of next year.
Don't know if it was already linked, but here's MIT's analysis of AP1000 LCOE and what happened at Vogtle.
Seems like a rather unique situation, with AP1000 design being revised after construction basically started, then Westinghouse fell over, then Covid. Next AP1000 should have a much lower cost, numbers thrown around are:
A paper also plotted expected cost for different tech per kWe, from Capital Cost Estimation for Advanced Nuclear Power Plants
MMNC = NuScale, NC-SMR = SMR-160.
But we all know that large plants always go over budget for a random reason in US/EU. We don't build enough of them in the same way to really get to lower pricing, the Chinese copy-paste method of building a ton of AP1000 and the evolved CAP1000, HRP1000, etc. got them to the right side of the chart (HPR1000 claimed cost of $2450/kW, first 3 units operational and 10 under construction).
Hopefully SMR gets to scale and learn quicker than chonky units like AP1000, building 12 smaller reactor vessels should bring you to the 10th plant cost quicker than 2 big ones per site, making the cost drop for MMNC / NC-SMR even steeper than the chart above and with less crazy cost went up 3x oopsies.
As with any prospective study, the only way to lend credence to something like this is to do a proper review after the next unit is delivered. So whenever the next reactor would be built, let's see if these estimates come close.
And of course the standard criticism of HPR1000 is: we don't actually know their costs. So far nobody has been able to independently verify their claims, and there are big incentives to lie. We saw the same happening when China exported their high-speed rail and other infrastructure; suddenly HSR in Kenya cost three times as much as reported cost in China was supposed to be.
And of course the standard criticism of HPR1000 is: we don't actually know their costs. So far nobody has been able to independently verify their claims, and there are big incentives to lie. We saw the same happening when China exported their high-speed rail and other infrastructure; suddenly HSR in Kenya cost three times as much as reported cost in China was supposed to be.
But it's just steel and concrete and it follows the modelled projections in papers well; everything gets cheaper per copy built, and they build a lot of them.
HPR1000's predecessor CPR-1000 has 18 operational copies built since 2002 (first construction) and the evolved ACPR-1000 has another 6 units operational. They build a staggering amount of reactors compared to the rest of the world and go off the chart of cost vs copy #. And every ~10 units, they discover some new tricks and iterate on it (4x French M310 900 MWe -> CPR -> ACPR -> HPR / Hualong One -> Hualong Two).
Compared to AP1000 -> err, the CAP1000 I guess, the Chinese revision of it, and the bigger CAP1400 and CAP1700. They have 4x AP1000 operational in China (out of 6 total, 2x in Vogtle), and another 4x CAP1000, 2x CAP1400 under construction, so iterate they do, just a bit slower than with the French M310 based design line that have way more copies operational.
Oh I think that paper did use historical data for Fig 1 (left), "cost of copy x" vs "first of a kind cost"; It would be really nice if China would publish that for their reactors but as you said, numbers are fuzzy there, but it should be at least cost effective if they have 21 reactors and 7 different models under construction right now.
[edit]
Oh, South Korea is also cooking. System 80 -> OPR-1000 (12x operational) -> APR-1400 (6x operational, 6x commissioning/testing/under construction). System 80 -> AP1000 is the Westinghouse evolution line.
This is one of the data used for cost vs copy number in the paper.
Last edited:
The OPR-1000 is smaller than a System 80. It came came from an earlier Combustion Engineering design but it might have some system 80 features. The APR-1400 is a descendent of the System 80+ which is an evolution of System 80.
The AP1000 came from the AP600. The only thing it borrows from the System 80/80+ is the large steam generators. The AP600 used W 4 loop plant steam generators. I think the AP300 will also probably use 4 loop plant steam generators.
DOW and X-energy announced a site for the 4 unit Xe-100 hight temperature helium cooled reactor plant they are going to build. It5 will supply both electricity and process heat to a site in Texas. They want to start construction in 2026. Dow has a cost sharing agreement with DOE under the advanced reactor development program.
Interesting news. I've always had a soft spot for pebble bed technology after doing my undergrad final paper on them.
I'll just stop you there. The Noord-Zuidlijn is just steel and concrete as well, and I think you can see where I'm going with this ;-)
It's steel, concrete and politics and economics. Those last two make everything more complicated than it seems. Not saying that's all bad, I think regulation is a good thing in general, but it's not something that makes projects cheaper.
Good thing NIMBY doesn't get you very far in China, and you tend to build your nuclear reactors not in the middle of a city with mud as foundation.
After copy 30 they probably know what to do
After copy 30 they probably know what to do
The other interesting pebble bed technology is the Kairos Power molten salt cooled pebble bed reactor. They are hoping to start construction later this year on a small 35 MWt technology demonstration reactor for their 320 MWt plant concept. It would also be good technology for combine electricity and process heat applications.
It's really hard to go down this line of reasoning, because this is exactly how nuclear power in the US, Russia and France started as well. They were seen as projects for public good and/or maintaining nuclear weapons, so they were essentially subsidized and fast-tracked through means that have since become inaccessible, heavily contested and partially or completely rolled back.Good thing NIMBY doesn't get you very far in China, and you tend to build your nuclear reactors not in the middle of a city with mud as foundation.
After copy 30 they probably know what to do
The fact that China may be doing the same now doesn't mean any of this is transferrable to anywhere else, even if they'd export a mature 50+ unit experience reactor design. Like, that IS the AP1000, right? That's the result of decades of iteration on common designs by a very experienced nuclear reactor design company and all the companies it has absorbed through the years. Same with the EPR, another true disaster of a reactor design through no lack of experience. And decades of neglecting due diligence on engineering and system design are now biting France really hard, making their 'cheap' nuclear fleet extremely expensive at the most inopportune moment.
It is the AP1000+++ (using Intel's terminology) yes, but as seen in the cost paper, learning also needs to be done by the builders as most of the labor is done on site. With many copies of each plant design being built, the builders get valuable experience that makes it cheap and it probably is minimally transfereable to the west, they won't accept "belt and road" style project bids including crew.
The problem with AP1000 is that the first two copies went to construction hell and containment building design is changed past project start, then delays stacked up and bankrupted Westinghouse. Two in progress builds in the US were also cancelled due to cost and Westinghouse going out of business.
Only SMR may change that, standardize the reactor and safety requirements and allow only mimimal changes to site / containment can let it finally scale. Start with smaller steps and who knows if we can follow China in getting better at getting from blueprints to actual working things. Keep paying First Of A Kind cost for nuclear will never be profitable.
The problem with AP1000 is that the first two copies went to construction hell and containment building design is changed past project start, then delays stacked up and bankrupted Westinghouse. Two in progress builds in the US were also cancelled due to cost and Westinghouse going out of business.
Only SMR may change that, standardize the reactor and safety requirements and allow only mimimal changes to site / containment can let it finally scale. Start with smaller steps and who knows if we can follow China in getting better at getting from blueprints to actual working things. Keep paying First Of A Kind cost for nuclear will never be profitable.
lol
Absolutely! But, like, there's much more going on in this regard. The nuclear 'renaissance' of the late 90s/early '00s ground to a halt on the back of multiple structural issues. The 'construction hell' (I like it, like production hell in media production) happened in multiple stages, from federal support commitments not materializing early on to 9/11 the credit crisis to fukushima. All these things happened and caused things to change drastically for this project, and that's not something you can ignore. That's only 10 years of world history, and even the most optimistic construction timelines for nuclear construction now - including SMR - stretches at least that long if you include permitting.
I agree with everything you say, I even agree with what most nuclear proponents say, but I'm arguing that - essentially - results from the past DO inform expected returns in the future. Say you start construction on a bunch of nuclear projects with all the right intentions and systems in place right now, I guarantee that in the next 10 years there will be at least 3 more giant world events that will reshape each of those projects profoundly and cause redesigns, regulatory changes, construction halts for a year or more, etc. etc..
tokamak of less than half meter radius reaches 100 million degrees
the hype
https://www.sciencealert.com/a-compact-fusion-reactor-barely-3-feet-across-has-hit-a-huge-milestone
The science
https://iopscience.iop.org/article/10.1088/1741-4326/acbec8
the hype
https://www.sciencealert.com/a-compact-fusion-reactor-barely-3-feet-across-has-hit-a-huge-milestone
The science
https://iopscience.iop.org/article/10.1088/1741-4326/acbec8
Not immediately clear to me why this is important. Just the temperature in itself doesn't get you much, and much higher temperatures are possible with benchtop devices and have been for decades.
It might be doing something interesting from an engineering perspective, and some of the concepts developed might be interesting for larger devices, but just in terms of plasma confinement as far as I can tell this falls well short of existing devices like JET. As far as I can tell the triple product is roughly where fusion experiments were in the 1980s.
I believe the highest performance Tokomak at this time is JT-60U, which is not equipped for using tritium but I believe most people think it could reach scientific breakeven if it ran with D-T.
Last edited:
The benchtop devices you speak of do not have thermal plasmas of that temperature. They have electrostatically accelerated ions that have the energy of a thermal particle of that temperature. There is a big difference between that and having a large thermalized plasma. The key statement in the article of why this is important is:
ST stands for spherical tokamak. It is a much smaller and less expensive machine than the large tokamaks. Tokamak Energy is one of the better funded private fusion companies and the current machine is a proof-of-principle for the spherical tokamak. Their next machine, the ST80-HTS is a superconducting magnet ST that they project will get better plasma parameters than any currently operating or past tokamak. It is supposed to be completed in 2026. The machine beyond that in the early 2030s is a demonstration power production machine that will deliver a net of 200 MWe to the grid.
I know, that example is to illustrate why a temperature bereft of context is useless and in a broader sense how flawed this type of clickbait science communication is.
My understanding is that it is not unique in that regard, eg MIT SPARC.
Recent progress in superconductors allows higher field strengths, which shrinks the overall size of the tokomak, which makes everything about it cheaper and faster to deal with, with the result they may end up not that far behind ITER in spite of getting started decades later. It is an interesting instance of the wait calculation.
Is there anything unique about the spherical configuration that makes it particularly advantageous?
Cheaper to build, cheaper to run in certain aspects, more efficient and stable D-shaped plasma.
Maybe, in theory, probably not. The ST is a device with lots of theoretical advantages if they would work, but there are scaling and in particular magnet strength issues (both mechanical and field strength) that have so far made spherical tokamaks a bit of a laughing stock among experts.
The whole idea of a spherical tokamak is to take a known scalable design (the regular tokamak) and improve its plasma aspect ratio (the shape of the plasma cross-section, the most stable plasma is about as tall as it is wide, regular tokamaks have elongated plasma cross-sections that are inherently less stable). That's done by reducing the hole in the torus as much as possible, and that in turn is done by making the thing that punches through the middle of a reactor - the superconducting magnets - as small as possible.
That has ramifications for the reactor in that you can't push as much current through the magnet so your field strength goes down, which makes plasma density lower. Also the coils overheat and can't have as much shielding. Maybe most importantly you can't scale up the design much, so you can't get as much of the plasma scaling benefit that tokamaks have. That is a biggie, because Q is very much dependent on the size and resulting reactivity of your plasma. Bigger is better.
Then again, Q of an ST is inherently higher already, so you can 'get away' with a smaller reactor. What ST proponents are hoping for, is that the scaling downside isn't as bad as the efficiency upside of a ST over a traditional tokamak.
But of course all of this has already been figured out long ago, and it's been consensus that a spherical tokamak just won't work. Unless we get much better at plasma physics, there is no way to scale an ST to power generation before having much better tokamaks. Any technology that will make one work, will make the other work even better. Maybe theoretically there is a chance the smaller size of an ST causes it to first get this tech (as it's much more expensive to fit to a $10B tokamak than a $100M ST), but... we'll see if that plays out. Lots of SMR concepts tout the same benefits for fission reactors but so far it's all just been expensive - but small - boondoggles.
And that's only the really high-level stuff. There are much more damning critiques of STs if you get into the details. Everything about them is harder in some way, and it reintroduces engineering problems that have been solved for tokamaks already (like needing super high flux neutral beam injection, maybe not even being able to work with superconducting magnets at all, etc.)
----------
By the way, despite me being a Debbie Downer on all of this, that is why the announcement of 100M K temperatures is actually a big deal in some way! It sorta debunks my point about some of the details being so damning. Two of the really big detaily critiques is the need for so much secondary heating and the impact of high temperatures and neutron fluxes on the central column under fusion conditions. Having demonstrated a sustained (eh, 150ms) plasma temperature at anywhere near decent plasma densities is a genuine milestone and proves there may yet be some legitimacy to the concept.
The whole idea of a spherical tokamak is to take a known scalable design (the regular tokamak) and improve its plasma aspect ratio (the shape of the plasma cross-section, the most stable plasma is about as tall as it is wide, regular tokamaks have elongated plasma cross-sections that are inherently less stable). That's done by reducing the hole in the torus as much as possible, and that in turn is done by making the thing that punches through the middle of a reactor - the superconducting magnets - as small as possible.
That has ramifications for the reactor in that you can't push as much current through the magnet so your field strength goes down, which makes plasma density lower. Also the coils overheat and can't have as much shielding. Maybe most importantly you can't scale up the design much, so you can't get as much of the plasma scaling benefit that tokamaks have. That is a biggie, because Q is very much dependent on the size and resulting reactivity of your plasma. Bigger is better.
Then again, Q of an ST is inherently higher already, so you can 'get away' with a smaller reactor. What ST proponents are hoping for, is that the scaling downside isn't as bad as the efficiency upside of a ST over a traditional tokamak.
But of course all of this has already been figured out long ago, and it's been consensus that a spherical tokamak just won't work. Unless we get much better at plasma physics, there is no way to scale an ST to power generation before having much better tokamaks. Any technology that will make one work, will make the other work even better. Maybe theoretically there is a chance the smaller size of an ST causes it to first get this tech (as it's much more expensive to fit to a $10B tokamak than a $100M ST), but... we'll see if that plays out. Lots of SMR concepts tout the same benefits for fission reactors but so far it's all just been expensive - but small - boondoggles.
And that's only the really high-level stuff. There are much more damning critiques of STs if you get into the details. Everything about them is harder in some way, and it reintroduces engineering problems that have been solved for tokamaks already (like needing super high flux neutral beam injection, maybe not even being able to work with superconducting magnets at all, etc.)
----------
By the way, despite me being a Debbie Downer on all of this, that is why the announcement of 100M K temperatures is actually a big deal in some way! It sorta debunks my point about some of the details being so damning. Two of the really big detaily critiques is the need for so much secondary heating and the impact of high temperatures and neutron fluxes on the central column under fusion conditions. Having demonstrated a sustained (eh, 150ms) plasma temperature at anywhere near decent plasma densities is a genuine milestone and proves there may yet be some legitimacy to the concept.
I think the only thing giving some hope here is that the scaling is not perfectly understood, so it is ambiguous which design performs better at realistic power plant scale. Various tokomak geometries, possibly even stellarators, etc. Toroidal tokomaks with a D-shaped cross section with the flat side facing in are the farthest along but it's so expensive to experiment that we don't really know how other approaches would scale.
Read about them a little yesterday and apparently the neutron flux in the center is so severe that some people are talking about a resistive magnet made of molten metal just to be able to get the heat out.
infernal666
Ars Praefectus
Molten metal magnets,that combination of words hints at some very exciting failure modes.
Yeah, I've read that too as I was factchecking my own musings in the previous post. It sounds, like a lot of physicsy engineering in these kinds of megaprojects, really cool but also like it couldn't possibly work?
Liquids aren't stable forms of matter under a vacuum, they will always create a partial pressure of gas, especially at elevated temperatures. I'd assume a jet of molten metal running through a reactor with a few hundred MW/m2 of radiation will heat up plenty of that molten material to mess up your vacuum, right?
But the more ontopic conclusion to be drawn from this is: having to re-solve a bunch of engineering problems that we've already spent decades on is the central theme of most novel fusion ideas.
Liquids aren't stable forms of matter under a vacuum, they will always create a partial pressure of gas, especially at elevated temperatures. I'd assume a jet of molten metal running through a reactor with a few hundred MW/m2 of radiation will heat up plenty of that molten material to mess up your vacuum, right?
But the more ontopic conclusion to be drawn from this is: having to re-solve a bunch of engineering problems that we've already spent decades on is the central theme of most novel fusion ideas.
I think a jet of molten metal exposed directly to the vacuum vessel is probably not what they mean for the reasons you state. It would instead probably be the usual tungsten or whatever walls and pressure vessel, with some channel behind the walls for the molten metal. If that was lithium it could serve as field coil, and breeding blanket, and primary circuit working fluid. You can see why they'd see that as an appealing idea. You replace all the problems of your magnets having to deal with all that neutron flux with useful heating and tritium breeding. Possibly get the coil closer to the wall because no need for shielding.
Moreover at a big enough scale tokomak a lot of people seem to think resistive magnets are viable, and if the magnets are resistive molten metal is one way to get the needed cooling. Resistive magnets for that kind of field strength would require extremely aggressive cooling, and if it wasn't molten metal it would probably have to be copper or silver with liquid nitrogen cooling.
The engineering tradeoffs there are well beyond me and I know only just enough to know it would be complicated. You're dealing with neutron susceptibility, how close you can get to the walls if you need shielding, electrical conductivity being higher for copper/silver at low temperature but more cross sectional area being available to molten metal, and so on. Not clear to me the molten metal resistive magnet idea is a good idea, only that at first blush there are appealing aspects to it.
Last edited:
Here is the case for spherical tokamaks and how the new high temperature superconductors might make them feasible for a reactor. I haven't looked at any of the conceptual designs recently, but I don't think any steady state tokamak is really feasible without superconducting magnets. There would already be a lot of required recirculating power in a tokamak power plant even with them.
Here's a front-page article about new experimental results from an inverted D shaped tokamak. It gave unexpectedly good results and would have some advantages for reactor engineering.
People have done a fair amount of exploration of different plasma shapes in tokamaks. TFTR had a circular plasma cross section and the GA tokamaks looked at some weird shapes. The D shape was widely settled on because of better MHD stability properties.
I am aware, but this is one of those counter-intuitive scaling things.
https://www.osti.gov/servlets/purl/2684
For a given field strength, fusion power is proportional with plasma volume, but resistive magnet power dissipation is proportional to the cube root of plasma volume. This means that while existing experiments are superconducting and it would be a huge energy cost to go resistive, at power plant scale it may be the case that resistive magnets would be a minor loss.
Moreover, with copper or silver cooled to cryogenic temperatures it may be possible to get stronger fields than are possible with superconductors, because unlike superconductors, resistive magnets do not have a critical field strength. A stronger field would allow higher density plasma, all else being equal, and hence shrink the overall size of the machine.
I only realized this reading about the molten metal central solenoid proposed for spherical tokomaks, but if you make your resistive magnet from molten lithium you are getting neutron shielding, tritium breeding, power generation coolant loop, magnet conductor, and coolant for the magnet conductor from a single fluid, and because combined into a single fluid, you can put it closer to the reactor walls and get a much higher cross sectional area for the conductor (to make up for lower conductivity of high temperature lithium vs cryogenic copper).
That seems like it would be highly desirable from an engineering perspective and important to the economics even if it's a significant loss term from a power generation perspective. Some of the energy would be recouped because it would be the same fluid as power generation, though obviously with efficiency well below unity.
I do recall that. As I said there may be advantages but it is extremely difficult to test different confinement concepts at the needed scale. There is probably a reasonable scientific case for half a dozen ITER-scale experiments for different tokomak variants, stellarators, etc, but the money just isn't there. Or rather it is there but it is more important to spend it on tax breaks for the wealthy and stealth bombers rather than technology that would contribute to keeping the planet habitable to humans.
The study shows a design that is well beyond anything that could be built. The "small" example was 30 GW of fusion power. A plant of that size could probably destroy itself in a disruption event among other problems. he scaling economics are based on the old low temperature niobium-tin superconductors. It probably wouldn't hold with the new high temperature superconductors.
One of the best funded private fusion companies, General Fusion, is building a pulsed power spherical tokamak. It has a liquid metal first wall that will compress the plasma to ignition.
I don't think there is a need to build lots of ITER scale machines. Ignition physics could have been studied in a smaller and cheaper machine that did not use superconducting magnets. ITER got morphed into something well beyond the initial purpose which is too bad because a more limited scope machine would have been built and demonstrated ignition already.
DOE awarded $46M in total to 8 private fusion companies this week. From the article:
Commonwealth Fusion Systems (tokamak)
Focused Energy Inc (laser fusion)
Princeton Stellarators Inc
Realta Fusion Inc (tandem mirror)
Tokamak Energy Inc
Type One Energy Group (stellarator)
Xcimer Energy Inc (laser fusion)
Zap Energy Inc (z-pinch)
Edit: Some notable companies absent from the list are Helion Energy (FRC), TAE Technologies (FRC), and CTFusion (spheromak).
The companies getting the awards are:
Commonwealth Fusion Systems (tokamak)
Focused Energy Inc (laser fusion)
Princeton Stellarators Inc
Realta Fusion Inc (tandem mirror)
Tokamak Energy Inc
Type One Energy Group (stellarator)
Xcimer Energy Inc (laser fusion)
Zap Energy Inc (z-pinch)
Edit: Some notable companies absent from the list are Helion Energy (FRC), TAE Technologies (FRC), and CTFusion (spheromak).
Last edited:
Seems like no more than a token amount of money? Especially as CFS, Tokamak, Zap, etc. have already gotten hundreds of millions in VC and angel funding.
Hm, that's the folks behind SPARC. I would say that is probably the most credible private effort out there. Conventional tokomak design using newer (but already commercially available) superconductors to achieve a 20T magnetic field (vs ITER's 12T).
It does seem like a small amount for some of the companies. I think CFS raised close to $2B in their latest round of funding a couple years ago that let them start building the magnet factory and SPARC. CFS is so well funded that they were supporting Realta Fusion (short video if you want to see some information about them) although it's not clear if they are providing cash or just superconducting magnets. The article does say this about the funding:
It doesn't say what the milestones are and how much additional money they could get for meeting them.
CFS is the best bet to hit scientific breakeven, but tokamaks are probably the worst concept in terms of overcoming engineering challenges needed to build a power plant.
You'll have to expand on that. Just due to the size requirements?
Because from like a 'tick off this list of engineering and science issues'-perspective, tokamaks (and arguably stellarators) are the furthest along. Alternative approaches pretty much all have a larger number and arguably less certain-to-succeed hurdles to overcome.
(and in all of this I'm only talking about the power plant goal, i.e. going from Q=1 to Q>100 and running continuously)
That's an interesting take, I am curious what you think the shortcomings are and other approach is better positioned in terms of being commercially viable?
My thinking is that inertial has very difficult problems in terms of fuel fabrication and achieving reliable operation due very difficult transients if you were to ramp up to several shots per second, and of the various magnetic or other approaches, tokomaks are the only one close enough to scientific breakeven that we can be confident of scaling into breakeven territory. Stellarators are arguably not that far off, but then it looks a lot like they have a much tougher job in terms of fabricating the magnets etc which would be relevant for a commercial design. Other magnetic confinement or other eg electrostatic confinement concepts are either unproven at similar scale or outright exude all the hallmarks of snakeoil.