How big are the commitments here? I’m having trouble finding actual dollar amounts. Does this actually represent an infusion of money into these SMR efforts, or are these “commitments” tied to so many missable targets that it’s actually meaningless?
Oklo in particular seems to be total vaporware, I can’t find a single technical picture anywhere of anything this company’s reactor is seeking to do. They seem to raise money based on a rendering of a ski lodge.
A huge, concrete investment in TerraPower would be more interesting, but as a molten salt SMR which has never been built, this also looks extremely non-committal.
SMRs in general seem like a dead end, we’ve heard about them for decades and they don’t seem to be any closer to making nuclear power buildouts less expensive.
Everything that makes proven nuclear power plant design expensive seems to revolve around the same drivers of expense for all long-term construction: large up front capital requirements, changing regulations, failure to predict setbacks, and pervasive lawsuits. SMRs purport to tackle a couple of these (shorter-term builds, fewer setbacks), at the cost of considerable efficiency, but so far this seems like an inferior alternative to “just get better at building proven nuclear plant designs”.
Nuclear is never getting cheap [1]. Nuclear reactors need to be large to scale [2]. As for why SMR persists? Because someone makes money selling the idea. That's it.
And SMRs get sold is the very idea you state because it sounds compelling: the more you build, the cheaper it gets.
Nuclear seems like it should work. But there are massive unsolved problems like the waste from fuel processing, processing the spent fuel, who can be relied upon to run these things, who can be trusted to regulate them and the failure modes of accidents. Despite there being <700 nuclear reactors built we've had multiple catastrophic failures. Chernobyl still has a 1000 square mile absolute exclusion zone. Fukushima will likely take a century to clean up and cost upwards of $1 trillion if not more.
Yet this all gets hand-waved away. Renewable is the future.
> who can be relied upon to run these things, who can be trusted to regulate them and the failure modes of accidents.
I personally trust the Nuclear Regulatory Commission. I also trust the Canadian Nuclear Safety Commission, and the regulatory bodies in the UK and the EU.
Why?
The failure modes are not binary. A reactor is not just operating fine or going boom. There are multiple small failures that can happen, and you can get an idea if a country's nuclear fleet is run with safety in mind or not.
Chernobyl happened during a safety exercise, an exercise that was attempted 3 times before and failed 3 times before. In principle the plant should not even have been allowed to operate until the exercise had been completed. The exercise was supposed to demonstrate if in case of reactor emergency shut-down the cooling water can be kept circulating in the core for one minute, the amount of time it took for the Diesel generators to ramp up power; it was an essential exercise to perform before starting full power operations. The fact that the plant was allowed to operate for 3 years without completing this exercise - no, actually, while failing this exercise multiple times, tells you a lot about the safety mentality of the nuclear industry in the Soviet Union.
In the US, the NRC performs a lot of monitoring, and the results are published. For example, here's [1] a dashboard of performance indicators. There are 17, such as: Unplanned Scrams per 7000 Critical Hours, Unplanned Power Changes, Residual Heat Removal System, Reactor Coolant System Leak, etc. Out of about 100 reactors, you can see only green, with the exception of one yellow; that yellow is for the Palisades plant that is not currently operating, it is in the process of restarting operations, and I am sure it will not be allowed to restart until all the performance indicators are green.
I more or less agree with your comment but feel it should be pointed out the CSIRO economic feasibility study is specific to Australia.
The arguments made there; why Australia is better to pursue renewables now rather than hope for nuclear eventually have no bearing on, say, China's use of nuclear for 20% of Chinese baseload.
A large part of the CSIRO argument is the greenfield standing start no prior expertise massive upfront costs and long lead time to any possible return.
China, by contrast, has an existing small army of nuclear technologists, multiple already running reactors, and many reactors of varying designs already in the design and construction pipeline.
Even China who committed to significant nuclear capacity and wanted to ramp up their nuclear percentage to 20% (IIRC) is slowly moving away. The percentage of nuclear has in fact reduced over the last 5 years and initial commitments/projections of nuclear capacity are likely not going to be med. The whole reason being that solar (and to a lesser degree wind) have become so cheap that nuclear just doesn't make economical sense even for China.
China is a special case. In fact, it's the one country on Earth I'd actually trust to build, maintain and regulate nuclear power.
I don't believe China is convinced (yet) of the long-term viability of nuclear power (fission or fusion) but, like with many things, they're hedging their bets. In the US? It's just another opportunity to transfer wealth from the government coffers to private hands through a series of cost overruns, massive delays and under-deliveries.
China's advantages here are extreme. They have the manufacturing base, would likely use the same plant designs in multiple places (rather than a separate procurement process in every city or province) and they have a bunch of existing infrastructure that gives them options, like they're pioneers in UHVDC transmission lines that might make it more viable to build a nuclear reactor away from populated centers. Even UHVDC development was to solve a largely China-only problem: the power generation is mostly in the west part of the country whereas the people are in the east.
And yes the CSIRO report is Australia-specific but the timeframes for building nuclear power in the US are similar: 10-15 years. Starting today it's unclear if such a plant would be online by 2040. Yet we can build solar in months.
That's the other part of this: if we're just looking at data centers, theyh can be placed anywhere. You can ignore where fiber runs. You just build more fiber if you have to. DCs need power and water, basically. The Southwest is very efficient for solar [1] but light on for water. There's the Colorado River but that's been tapped beyond its limits already.
Along the Mississippi is another option. Not as efficient as the Southwest for solar but water is plentiful. Inclement weather is an issue though, both tornadoes and the winters.
- Spent fuel is a solved problem, we just store it securely
- Who can be relied upon: who do you rely upon to run your drinking water?
- Failure modes of accidents: have been extensively studied and essentially designed out
- Multiple catastrophic failures: sounds bad until you realize that you can name only two:
1. Chernobyl: old flawed reactor design, basically impossible today, a few unfortunate deaths among first responders in the cleanup, that's it
2. Fukushima: no radiation deaths. You would get a higher dose of radiation flying to Japan to visit Fukushima than from drinking the irradiated leaked water there.
> upwards of $1 trillion if not more.
Where are you getting this number? According to https://cnic.jp/english/?p=6193 it was estimated at JPY 21.5 trillion (roughly USD 150 to 190 billion).
> Spent fuel is a solved problem, we just store it securely
This is simply untrue. Depending on the type and enrichment of the fuel it will need to be actively cooled for some period, possibly decades. After that you can bury it. You need facilities for all of this. You need personnel (done by the NRC currently) to transport and install new fuel, remove old fuel and transport it to suitable sites as well as manage those sites. Before they even make it to storage sites they'll typically be stored onsite or in the reactor for years.
> Who can be relied upon: who do you rely upon to run your drinking water?
Given the current administration, almost nobody. The state of drinking water in places like Flint, MI is a national disagrace. The continued existence of lead pipes that leech lead into drinking water in many places is a national disgrace. The current administration gutting the EPA and engineering the Supreme Court to overturn things like the Clean Air Act and the Clean Water Act are just the cherry on top.
A significant ramp up of nuclear power would necessitate a commensurate ramp up of the NRC in all these capacities.
> Failure modes of accidents: have been extensively studied and essentially designed out
Like I said, hand waved away.
> Where are you getting this number?
Multiple sources [1][2]. Fukushima requires constantly pumping water to cool the core. That water needs to be stored (in thousands of tanks onsite) then processed and ultimately released back into the ocean, which itself is controversial. Removing the core requires inventing a bunch of technologies that don't exist yet. The decomissioning process itself is something most of us won't live to see the end of [3].
The $1 trillion and a century for 1 nuclear plant. Pro-nuclear people will point to the death figure because it suits their argument. It's economically devastated that region however.
And as for Chernobyl, billions of euros was spent building a sarcophagus for the plant, only to have the integrity of that shield destroyed by a Russian drone.
The issue with spent fuel has to do with the long term (essentially permanent) storage part and is purely political. It's a solved problem except for getting approval for the solution.
The other fuel issues you mention are already dealt with today as a matter of course. It's just the final part that remains up in the air.
You are the one hand waving about failure modes. As with aircraft, as failures have happened we've learned from them. New designs aren't vulnerable to the same things old ones were. All the mishaps have happened with old designs.
Personally I think the anti-nuclear FUD that the climate activists push is unfortunate. We would likely have been close to carbon neutral by now if we'd started building it out in the late 90s.
That said, I'm inclined to agree that solar might be a better option at this point in environments that are suited to it. The batteries still aren't entirely solved but seem to be getting close. In particular, the research into seasonal storage using iron ore looks quite promising to me.
Yes, because others were mostly not affected by the Fukushima disaster despite being in the impact area. Why? Because they took safety precautions. Onagawa was closer to the epicentre, but they built on a high embankment and did not flood and lose power.
Anti-nuclear people conveniently ignore, because it suits their argument, that Japan is restarting their nuclear energy program. They finally understood that there's no other viable option for energy security, price, and achieving decarbonization goals.
> The combination has had a toll on Japanese automotive (and other) exports. Barring Fukushima’s impacts, one would assume a return to pre-2008 fiscal meltdown exports by now. But basically they’re static. That’s in the range of $200 billion in lost exports just for the automotive industry.
>
> It’s likely fair to attribute $20 to $50 billion of that to irrational fear of radiation.
Like, are you serious? This is the most bizarro accounting I've ever seen.
> ...that’s about $100 billion in extra fuel costs.
And now it's counting as part of the cost of Fukushima the fossil fuels needed to replace it. Even more wacky accounting.
> another $22 billion for unexpected health costs due to burning extra fossil fuels.
It continues to get even more wacky, if that was possible, by attributing this cost to the Fukushima disaster. These are costs that would be avoided with a strong nuclear electricity generation program! These are arguments in favour of nuclear! It's not cost-effective for Japan to cover their land mass and offshore areas with solar and wind arrays! They have regular earthquakes and typhoons which would knock these vast arrays offline and take massive amounts of time and money to get back online!
You said: 'Fukushima will likely take a century to clean up and cost upwards of $1 trillion if not more.' The sources you provide don't provide the numbers or, if they do, they include bogus numbers that actually make the case for nuclear.
They should focus research on thorium reactors as they are supposedly cleaner than what we have today, and afaik you can actually use the fuel waste again and again, so it drastically reduces the problem of nuclear waste and what to do with it.
The promise of thorium is that it requires external energy to be added to maintain the reaction. The theory is that it is safer because of this as it's far less likely that you get a runaway or out-of-control reaction.
The reality is more complex [1].
Molten salt reactors are another active area of research but they have been for decades as well.
> But there are massive unsolved problems like the waste from fuel processing, processing the spent fuel, who can be relied upon to run these things, who can be trusted to regulate them and the failure modes of accidents. Despite there being <700 nuclear reactors built we've had multiple catastrophic failures. Chernobyl still has a 1000 square mile absolute exclusion zone. Fukushima will likely take a century to clean up and cost upwards of $1 trillion if not more.
sigh same low-tier non-issues brought up over and over again by people with no idea what they're talking about.
Look up some hard data before you speak.
- A nuclear reactor produces a tiny amount of waste per unit of power generated and it's all solid. Most sites just store it on-site because why not? Containment of small amounts of solid waste is as big of a non-issue as can be, obviously.
You realize our current energy generation revolves around burning up coal and gas and dumping the waste products into the atmosphere right? Right? And that those waste products include radioactive materials that you're so fake worried about?
You're out of your mind, completely gone in terms of what's actually happening right now vs what you're worried about. Detached from reality.
- Who can be trusted? We've had nuclear reactors for 50+ years, so... the same people that are already doing all that? What sort of a question is this? You're asking how to do something we're already doing.
- As for accidents, again, look up any data in existence. Nuclear is the safest energy production method by far, and yes, it's safer than e.g. solar. The fact that all you can point to are two accidents that have barely cost any lives at all proves that.
The very tsunami that caused Fukushima in the first place claimed 20 000 lives and all you can speak in regards to the plant is economic damage. Laughable.
You're displaying insane levels of ignorance. Look up data before you speak. Even consulting an LLM would have been better than just making stuff up.
China are building dozens simultaneously, and even with their questionable workers rights, safety and environmental practices, they cost $7 Billion a pop.
A dozen $7B nuclear plants is $84B, which is incidentally almost exactly the estimated cost of the SF-Gilroy-Palmdale plan for California's high speed rail. If you count all of phase 1, the P50 estimated cost goes up to $106B. That's the equivalent of 15 nuclear plants.
China has over 28 plants in progress, which should provide a total of >32GW of capacity when they're completed. That's 32×24×365= 280TWh of electricity per year. California's total electric grid in 2024 produced 216TWh.
Which is to say, $7B is a huge sum. But as far as infrastructure goes, China is currently building 130% of all of California's generation capacity that'll be complete within a decade or so, for much less than double the estimates for a high speed rail system that'll serve almost nobody by 2038.
$7B is a lot of money. But it's actually a very reasonable amount of money because the projects are actually happening. 28 $7B projects in the US are actually probably closer to a trillion dollars in investment for far less net public good over five times the timeline.
Whether they run over budget (or whether this is an under inflated figure) is yet to be seen, but it would seem that China is bringing the cost down, and substantially.
I'm not a nuclear expert by any means, but from the reading I've done, they're largely designing and building the reactors themselves these days. And it seems that to help keep the cost low (among other reasons), they're also helping other countries build them.
There was a recent study from Chain where they assess their own 4th generation Nuclear Reactor programme as being at least 10 - 15 years ahead of the west, and specifically said even that number is conservative estimate.
I wouldn't be surprised if they accelerate their time line and building target.
Yes, China have a good shot at doing it because they are building 33 simultaneously now and they have questionable workers rights and environmental policies.
As I said, if a developed country can do half what they’re doing (ie twice the price and double the construction time) in the next 20 years it would be a miracle.
It's not really a fair comparison though, is it? Is a questionable environmental policy worse than a bad electric grid? America has a dirty grid that has fairly limited capacity. How many fossil fuels will we burn (producing electricity, and powering non-EVs) because we aren't building nuclear? The environmental benefits of having nuclear power probably largely make up the difference (if they don't exceed it), and that's over the time scale of a century or more where we'll need to catch up.
Workers rights I have no real knowledge on. But China isn't known for their track record on any kind of rights, and arguably US blue collar workers have a pretty awful quality of life that the government largely doesn't take the blame for (because we don't have state-run healthcare and minimum wage doesn't keep up with the cost of living). China has forced labor, America has legalized slavery in the prison system. Plenty of American industries rely on the unethical use of migrant labor while the state disappears those same people to "alligator alcatraz" or overseas prisons. I don't know the full extent of how bad things are in China for the kinds of workers who build these plants but I am hesitant to overlook how bad things are in the US.
It's also a country that doesn't seem to care if the project is not cost-effective from the PoV of western companies. This is always a salient point missing from most conversation about the US and by extension the Western world; the advocacy of cheap energy are hiding the argument that nuclear power is both more consistent in power delivery and cleaner (arguable with the nuclear waste ofc) than any alternative currently available.
Remember those systems are non-intermittent and have lifespans of 50 years or more. Server farms are not amenable to load shifting, they expect round the clock power. Trying to power them with intermittent sources would need very hefty power banks.
If this nuclear plant has 2 GW of power output, were talking about 2.4 billion dollars to store 12 hours worth of the plant's output assuming $100 per KWh of storage.
Your numbers are off. Korea + China ~ 2500 $/kw, USA ~ 6-9000 $/kw. ` GW ~ $2.5B. A large portion of that is dealing with archaic regulations and very long timelines. Important to have regulations that are functional protect the public but also don't inhibit industries growth (which were the design of Nuclear regulations in the 80s).
Cheap-er, not cheap. They’re still fundamentally massive complicated constructions. They will never be as amenable to mass production cost reductions as things like solar and battery
Can we please not have these "slightly improved language" comments? You're arguing against something I didn't say and making a meaningless nitpick on word choice.
you literally said "cheap" and the comment said "cheap-er not cheap". I think the comment is correct and you are wrong. China is building the same design again and again and again. And it's still not cheap.
i'm sorry it came across that way. let me rephrase.
"cheap" to me implies it is affordable in a relative sense, compared to other options. It will almost certainly never be cheap - even if we make it cheaper through more production, it is going to remain in the group of the least affordable power generation technologies.
tbh i don't think either the original or improved language post is presenting effectively because they both just give a conclusion without any nuance, explanation or support. "cheap" cheaper who cares? $/kwh matter. transmission costs matter.
If you have credible figures then present them with citations. Otherwise you're just hand waving.
I don't think anyone will dispute that the initial build out for solar is far far cheaper. That much is self evident to everyone. The devil is in the rest of the details.
>I don't think anyone will dispute that the initial build out for solar is far far cheaper.
OK.
>The devil is in the rest of the details.
Now, this is "hand wavy" instead of answering my question and pointing to sources who support the up thread claim that nuclear will be "cheap" v. alternatives.
Do you have an LCOE study showing nuclear as "cheap"?
It's an elegant rendering trick, but if their worlds are represented as a torus, then I expect this would make rotation on a spherical globe view unintuitive.
One example of this: I would expect each location would not have a single antipode (opposite coordinate) but would instead have three. If you were to start at location A, rotate travel 180 degrees along the latitudinal axis to location B, then 180 degrees around the longitudinal axis... on a sphere you would expect to be back at location A, albeit upside down. But on a torus, you are in a completely different location, which is the 'C' antipode. Rotating 180 degrees latitudinally from here will bring you to point D, the last of the antipode set.
I don't find it to be a problem as planning a route from A to B isn't done by looking directly for it, but by subconsciously referring to and plotting a path through landmarks along the way that aren't close to antipodal.
One of the worlds i played on had road planning from the start and a set of roadways covering the entire world in a 4x4 square grid. Pathing to point D was just a matter of going 2 blocks in one direction, and 2 blocks in an orthogonal to it.
In a world without such roadways, you'd look for landmarks such as oceans and continents instead.
Ultimately, you don't care if somewhere is antipodal or not because you never see the antipodes to where the globe is currently looking at without rotating the globe.
Wonderful write-up of attempting to tackle this problem. I believe there must be a significant number of people who have played both Minecraft and Super Mario Galaxy, and had something like this sequence of thoughts - although you have followed it all the way to an actual demonstration, and written up your thoughts along the way so clearly.
The vertical distortion is the biggest issue IMO, there are a few reasonably satisfying ways to approach the horizontal tiling of each “shell”. For example, you can make your world a donut instead of a sphere, and now you have a perfect grid at each level! Of course, this introduces a level of distortion between the interior and exterior, so you also twist the donut once, and now you’ve both solved your distortion problem and invented the stellarator fusion device.
Many of the levels in that game take place on tiny planetoids with spherical surfaces and central gravity. "Spherical" sells it short, there were some truly wild topologies around which Mario could run and jump.
Rust uses the traits “Send” and “Sync” to encode this information, there is a lot of special tooling in the compiler around these.
A type is “Send” if it can be moved from one thread to another, it is “Sync” if it can be simultaneously accessed from multiple threads.
These traits are automatically applied whenever the compiler knows it is safe to do so. In cases where automatic application is not possible, the developer can explicitly declare a type to have these traits, but doing so is unsafe (requires the ‘unsafe’ keyword and everything that entails).
There are tantalizing ways to create fusion which don’t require these precise conditions. For example, a simple farnsworth fusor device gets fusion reactions just by causing atoms to cross paths at super high speed until they collide - they simply don’t collide often enough to release anywhere near a net energy gain.
Inertial confinement fusion, such as the National ignition facility, does generate comparable pressures and temperatures to the core of the sun within the fuel pellet for an extremely small moment during an implosion. This is done by focusing a lot of energy on small target.
Plasma confinement techniques don’t utilize high pressure to create fusion; they rely on extreme temperatures which are significantly hotter than the core of the sun, which can produce fusion events in a plasma which is only pressurized to around 1 atmosphere (they also rely on different fuel types than the sun which fuse much more readily). The key is once again focus, a large amount of energy is put into a small amount of gas. The obvious issue with this is that the extreme temperatures would destroy any physical container rapidly - but given the electromagnetic nature of plasma, it can be contained using a strong magnetic field without reaching the surface of its physical container.
This would work very well for medicine/procedures which are known to work very well, to both doctors and consumers. This includes medicine which is already OTC (pain relievers), but also probably anything they can do at an urgent care: x-rays for broken bones or sprains, throat cultures, antibiotics (drug resistance is complex for this but people generally know they work).
Where costs will inevitably get complicated are:
1. emergency medicine, where the purchaser is in severe pain or possibly unconscious.
2. conditions without cures, or possibly even well-established treatments, and there is thus active experimentation and disagreement
Both of these are unpredictably expensive to an extraordinary degree, and the second category is sometimes rare enough that economies of scale don’t come into play for individual conditions.
I think government coverage of emergency medicine, aka ERs for severe injuries, is relatively uncontroversial due to its nature of treating unconscious patients.
However, that other category is very large in modern medicine. It includes all chronic conditions without cures, for which many options are available and improved techniques are constantly sought - and it includes complicated conditions where treatment has risks involved, which is basically a huge range of surgeries.
The problem in these areas is that the consumer does not have adequate understanding of the efficacy of what they’re buying, yet they’re driven to buy it strongly by pain and suffering. They are likely to want to do whatever a doctor or hospital tells them to do.
What is needed here is a consumer advocate with medical knowledge to keep prices consistent. In the US, this is provided by a mixture of regulation, medical malpractice lawsuits, and insurance companies.
Insurance companies are now failing in that role, but removing them entirely without any sort of replacement is going to leave the courts as the major vehicle to manage the costs - that isn’t a system renowned for efficiency.
I work with this algorithm quite a bit in a personal project. (I've changed from calling it WFC to Model Synthesis; Paul Merrell earned the naming rights, unless we discover an even earlier description.)
I mentally divide it into two important components:
1. Cross-correlation (the 'wave function')
2. Label selection (choosing where the 'collapse' occurs)
The "cross-correlation" is exactly what it sounds like - given an incomplete model and a set of example patterns, this data structure represents the cross-correlation of every place that any of the patterns would individually fit in the incomplete model. I think this is the most useful structure in the algorithm, and is a form of iterative 'feature detection' over the incomplete model.
I think there are many interesting things which can be done with this which aren't in the basic algorithm, such as:
- Having larger indivisible 'patterns' and collapsing such patterns all at once. Although example patterns are always broken down into discrete cells, larger patterns can exist where each cell gets a unique label and very limited transition set. The cross-correlation will detect where they can be placed, and if one is chosen the entire pattern could be chosen at once.
- Partial seeding of labels, e.g. selecting portions of the incomplete model and artificially limiting the available labels in that area, without collapsing it.
- Addition of information to labels which aren't easily expressed in an example pattern, such as rudimentary counting of distances.
The 'label selection' portion (in WFC terms, this is the algorithm used to choose which space to 'collapse') is where all of the random generation comes into play. If using Model Synthesis for full world generation, this is what should be focused on for generating more interesting types of environments.
However, given the common label selection algorithms I've seen and used, I think that Model Synthesis excels mostly at adding small-scale, largely self-similar patterns onto constraints provided by a larger structure. I think that other proc-gen methods better lend themselves to generating large-scale structures. A combination of the two - generating a large-scale structure to serve as a constraint, and allowing model synthesis to fill in interesting detail - seems like a good path to explore.
> Partial seeding of labels, e.g. selecting portions of the incomplete model and artificially limiting the available labels in that area, without collapsing it.
It seems like you if you had this, you could do some really interesting things with multi-scale collapse. For example, you could first use collapse to lay out biomes (e.g. beach needs to border sea, desert needs to border mountains, etc.) and then use those as constraints for the next scale level, which decides the details within those biomes, and the next level for cities and towns within those environments, etc.
> Partial seeding of labels, e.g. selecting portions of the incomplete model and artificially limiting the available labels in that area, without collapsing it.
I've done that in my toy WFC to let it generate an output which matches[0] the colours and transparency of a supplied PNG. Handy for generating text / shaped output. Does a pre-pass which only keeps the best N match options for the colour in that cell.
[0] As best it can - my input tileset only really has green, light green, slightly less green, a bit reddy-green, fractionally blue-green, ...
> I've changed from calling it WFC to Model Synthesis; Paul Merrell earned the naming rights, unless we discover an even earlier description.
Wang Tiles, 1961. The WFC algorithm is just one way to generate a tiling. I haven't done a literature search but I'd be surprised if there isn't a materially identical paper with decades of priority.
Good point - I just linked to the first relevant explanation. pvg linked to a much longer explanation that isn’t specific to, or about, Show HN: https://news.ycombinator.com/item?id=23071428
This release has Observer and Hooks which in my opinion really elevate the ability to express complex, reusable structures in the ECS.
For example, one of the well-known challenges in the paradigm is working with hierarchies or graphs - observers are a powerful tool for communicating between specific entities, and should make this a lot easier to express.
Hooks offer the ability to enforce cross-component consistency in a way that wasn’t previously available.
I think there’s some possibly good investment advice to extract here, if you’re able and willing to invest in fusion startups: based on the need to compete with renewables alone, commercial success implies that highly complex reactors simply may not have a market based on construction cost, even if they do generate power.
Of the fusion startups mentioned in the article, I’d say that makes Zap Energy the one worth gambling on (if you’re a gambler that is), as its success apparently depends on exploiting a fluid dynamics effect which was not well known in the past (“shear flow”). If this sufficiently solves the confinement problem, the resulting device looks ludicrously simple in comparison to contemporaries.
Of course it may not work at all, I sure don’t know if it will; but if you had to invest in one of these, that seems like the one where successful power generation actually creates a marketable product.
Oklo in particular seems to be total vaporware, I can’t find a single technical picture anywhere of anything this company’s reactor is seeking to do. They seem to raise money based on a rendering of a ski lodge.
A huge, concrete investment in TerraPower would be more interesting, but as a molten salt SMR which has never been built, this also looks extremely non-committal.
SMRs in general seem like a dead end, we’ve heard about them for decades and they don’t seem to be any closer to making nuclear power buildouts less expensive.
Everything that makes proven nuclear power plant design expensive seems to revolve around the same drivers of expense for all long-term construction: large up front capital requirements, changing regulations, failure to predict setbacks, and pervasive lawsuits. SMRs purport to tackle a couple of these (shorter-term builds, fewer setbacks), at the cost of considerable efficiency, but so far this seems like an inferior alternative to “just get better at building proven nuclear plant designs”.
reply