What Would You Do?
It is easy to say, "Nuclear should be regulated more sensibly!" But what if you had a blank slate to redo all nuclear fission regulation? How would you harness the power of fission while keeping the public safe?
There are two paths I find interesting. I'm sure there are many more possibilities, but I chose these because they also help explain a lot of the difficulty (and promise) in regulating fission power plants.
The Difficulties of Nuclear Regulation
The biggest challenge in regulating nuclear is that it's complex, and the potential impact of accidents is high. Outsiders probably underestimate the complexity of running a nuclear power plant by at least two orders of magnitude. A nuclear power plant has 20x more employees than a natural gas power plant of the same size, and it's not only because of the NRC.
Below is an excerpt from a tweet by nuclear entrepreneur Josh Payne that is a response to a hit piece on a different nuclear startup:
Due to various quirks of history, the nuclear industry is by far and away the most insular, secretive, and most difficult to approach as an outsider of any industry. It is incredibly easy to screw up reactor design and end up with a pile of junk. I've heard stories of startups that screwed up their mcnp input decks and when someone got around to checking them, we'll into the system design, it turned out that the core was far from critical. I know that INL screwed up on MARVEL by putting off shielding calcs for too long. When they finally evaluated shielding they discovered that they were going to cook the windings in their stirling generators. I have seen many many reactor designs that looked great right up until some tiny oversight either killed it outright or forced a complete redesign.
Now this wouldn't be that bad on its own, the problem is that these failures can have significant negative consequences for the other projects and the industry as a whole. On more than one occasion I've witnessed scope creep and unrealistic performance expectations based on fantastical vendor claims hinder or kill projects.
Several factors contribute to this complexity.
Managing Reactivity
Splitting uranium atoms requires neutrons to start the reaction. And neutrons interact with all matter, not just uranium atoms.
The actual design of the reactor has to allow neutrons released from one atom splitting to hit the next. One side effect of this is that small reactors typically require more fuel because neutrons are less likely to hit anything useful. The required neutron flux is not steady state, either. It depends on where the reactor is in the fuel cycle and the power plant's current output.
Power plants rely on the natural fission of uranium to initiate reactions, which requires a brake to prevent chain reactions from going too fast. A tiny change can alter the conditions and disrupt the neutron balance. Even a change in the feed water rate can increase the reaction rate by removing more heat. Modern water-cooled power plants control neutron reactivity by mixing boric acid in the reactor coolant water and/or by adjusting neutron-absorbing control rods. Other types of reactors typically use devices similar to control rods.
As you might imagine, feed water, reactor, control rod, and boric acid concentration management subsystems are very complex. They require constant maintenance, vigilance, and care. Some rarely deployed designs try to handle this balance more passively, but it almost always requires more engineered (and expensive) fuel and/or exotic materials. The market has vastly favored organizing systems around cheap fuel, water, and boron.
There is No "Off" Button for Fission Reactors
Fission reactions are messy. A neutron can go any number of directions and hit different types of atoms. When atoms split, they don't always do so the same way; there are many possible products. The reactions also release other types of radiation, such as gamma rays, of varying intensities.
Reactor designs have multiple ways to stop fission if things go wrong. Besides control rods or boric acid, water-cooled reactors stop if all the coolant boils off because there is no liquid water to moderate neutrons.
What is difficult to control are the unstable fission products in the fuel that undergo decay and release heat. Bad things can happen if heat isn't constantly removed, like in the Fukushima accident. Fuel melting, hydrogen explosions, and pressure buildup are some possible issues. Maintaining heat removal under any possible circumstance or disaster takes an incredible amount of effort in design and operations, even if the heat removal method is passive.
Off-Target Neutron Reactions
Materials in the reactor loop must withstand neutrons, avoid hindering the reaction, and avoid nasty side reactions.
Since neutrons can alter materials, many metals can undergo accelerated corrosion. Neutrons and gamma rays sometimes split water molecules into hydrogen and oxygen, gases that also cause corrosion. Many of France's nuclear power plants were offline during the 2022 European Energy Crisis due to poorly managed corrosion. Materials need to be carefully selected and monitored to prevent accidents.
The materials also can't absorb too many neutrons because it hinders the reaction. Part of the reason why the Germans were behind in building atomic bombs in World War II is that they didn't realize natural graphite has trace amounts of boron that absorb too many neutrons. Any materials must be pure enough to avoid these issues.
Neutrons can also turn materials into really nasty isotopes. A classic example is in valves. It is challenging to produce valves that don't leak under high pressure. Valve makers use a miracle metal alloy, called stellite, in valve seats to produce competent valves. Unfortunately, stellite contains cobalt. When cobalt gets hit by neutrons, it forms an extremely dangerous cobalt-60 isotope that produces high-energy gamma radiation. Little pieces of stellite rub off during valve operation, travel through the reactor, absorb neutrons, and migrate to traps where they can harm employees. What is worse, leaky reactor loop valves or random sources of high-energy gamma rays in the workplace? Not an easy question to answer or deal with!
The upshot is that neutrons and radiation add engineering complexity that drastically limits acceptable materials and fluids. Materials going into reactors can cost many times more than commercial grades, and the engineering and testing burden is many times greater.
Radiation Impact Variance
The types and intensity of radiation that a fission power plant can produce are remarkable. That makes protecting employees and the public difficult. Even worse, the method of dosing can increase the effect by orders of magnitude. Ingestion can be particularly harmful. An open Coke can or bag of chips in certain parts of the plant is a crisis and a safety failure.
Another confounding factor is that the impact of a dosage isn't always immediate, nor is it always the same. These factors can add legal and public relations challenges.
Responsibility, Non-linearity, Verification, and Incentives
One thing to lament about fission power plants is that the complexity means there is no magic software module that can handle all safety cases. One small group could put it together once and then copy it over and over. It could undergo much more significant scrutiny than other parts of the design, construction, and operation. Competing technologies, such as solar, wind, or natural gas, can abstract most safety issues to a piece of software in a manufactured piece of equipment, like an inverter or turbine control box.
Many aspects of nuclear safety are non-linear: dosages, accident interactions, and earthquake impacts. Ideally, in safety, you can use a worst-case basic calculation to set the baseline. These "worst-case" numbers are quick and easy to do. A detailed model or study can reduce the safety factor if the simple safety factor calculation is too expensive to implement. This algorithm does not work as well for non-linear effects. A model of a simple 8.0 earthquake on individual pieces of equipment might suggest everything is fine, but a more systemic model that includes equipment interactions can identify issues. We can also see this in trying to judge exposure effects from radiation, showcased in an argument between several experienced nuclear engineers and a nuclear startup CEO:
When Touran pointed out that claim could not possibly be true, Taylor responded by providing some calculations and chastising Touran, saying “Nick, a good engineer does not make claims about things he simply has no knowledge of.”
When Touran then looked at the new data, he saw that holding Valar’s spent fuel would result in a fatal dose in 90 seconds. Another nuclear engineer used a more advanced calculation method to argue the fatal dose would actually be as fast as 85 milliseconds.
Three smart individuals get answers that are all orders of magnitude apart. The radiation exposure depends on the time the spent fuel has been out of the reactor, the actual material, and the distance from the body to the source. None of these effects is linear, so answers depend heavily on starting assumptions. These three almost certainly had different starting places. Lengthy, uniform standards for important safety calculations are the norm.
The lack of abstraction, non-linearity, and the nearly infinite possible interactions within fission power plants make verification of safety incredibly challenging, not only for regulators but also for executives and other stakeholders. Verification is extremely costly, but necessary for the confidence of the public, regulators, and executives.
Challenging verification tends to breed risk aversion. Someone must authorize the entire project. Approving something that contributes to a serious accident creates extreme guilt, kills careers, puts the individual through a years-long legal wringer, and never goes away. Engineers must check their calculations repeatedly. Even the slightest concern is enough to delay.
Individual responsibility is necessary to keep skin in the game for the people designing and operating power plants, but absorbing verification efforts, regulatory risk, and accident liability is expensive. The nuclear power industry has no shortage of bankruptcies. A nuclear accident is one of the most costly mistakes a capitalist can make. It ultimately proves challenging to maintain the correct incentives.
Paths Forward
Given these difficulties, what regulatory frameworks might one try? We must be clear-eyed about risk and incentives to improve the design and engineering of nuclear power plants.
Industry-led Standards
The first approach is more conservative and applies best practices that have reduced risk over the decades of operating nuclear power plants.
The basic structure would be:
- The government in question authorizes a private regulator.
- The regulator's "ownership" consists of nuclear operators.
- The private regulator builds the regulatory framework and audits individual plants.
- It can penalize poor performers via the government's power.
The logic that underpins this path is:
-
Nuclear Power Plant Operations are Complex but Mostly Understood
Power plant operation and understanding of accident cases have improved since the 1970s. Experience and probabilistic risk analysis are behind the improvement.
-
Remove Complex and Expensive Interaction with Regulators
The gains in operational efficiency and safety are still time-consuming and hard-won, and the complexity does not interact well with a regulator like the NRC, which lacks the desire or manpower to keep up. Wild goose chases and delays are costly for the nuclear industry.
-
Private Regulation Can Improve Safety
The nuclear industry already has a private regulator in the Institute for Nuclear Power Operations (INPO). The NRC even delegates operator training standards to INPO, the most critical part of safety. Many NRC statutes contain little detail, and it is industry groups, like INPO, that determine the actual processes. It is also common for INPO and the industry to have standards that are more strict than the NRC's. Individual companies are also likely to propose fixes that go beyond standards when reporting issues to the NRC.
The incentives here are clear. Interacting with the NRC has massive costs. It is often cheaper to be stricter because the government is less likely to get involved. The operator can choose their preferred solutions.
A private regulator without the NRC would be more efficient by avoiding expensive requirements that may not add much safety value. It would have the same incentive to favor safety because whatever reappears after an accident will likely be much worse than the NRC ever was. Nuclear plants are so valuable and long-lived that they must be protected from intervention to maximize shareholder value.
The operators would have to pay for this regulator, and INPO's yearly budget is just over $100 million.
-
Shared Accident Pain
One unfortunate feature of the nuclear power industry is that an accident anywhere can trigger increased regulatory scrutiny, shutdown of viable facilities, and cost increases. Having a private regulator with agreed-upon rules, information sharing, and oversight for all parties reduces the risk of a poorly run plant ruining it.
-
Maintains Public Outreach Success
Nuclear power has significantly improved its image in the last few decades by reducing accidents and near misses while providing extensive analysis and transparency about safety measures to the public. None of these attributes would change.
The net result of the changes should be lower costs and better safety. Standards will be stricter, but operators are free to use the most cost-effective measures. Rules for licensing new facilities and operating existing ones will be easier to follow.
Maintaining Innovation?
New designs, especially those that aren't water-cooled, could be challenging to test or license if the private regulator and industry have to agree.
If the market truly favors water-cooled reactors, then there are limited opportunity costs. Performance-based regulation can still unlock many incremental improvements in water-cooled reactors through methods like qualifying cheaper materials or simpler safety systems.
There should be test parks that improve understanding of reactors with other types of coolant, fuels, or configurations, in case water-cooled reactors are not the final form.
Preserve and Expand
The primary benefits of this route are making it cheaper and safer to operate existing power plants, reducing the difficulty in building new water-cooled reactors, and pushing incremental improvements. Maintaining some semblance of entry for technologies like sodium, lead, salt, or helium-cooled reactors requires thought. These reactors have almost all been built at commercial scale and failed to beat water-cooled reactors. Many want to try again, and they should be allowed. An industry-led, incremental approach is very compatible with the history of nuclear power policy and deployment. The trick is to find balance.
Maximizing Innovation and Altering Public Nuclear Views
One of the only coherent criticisms of the current system comes from Jack Devanney, founder of the nuclear startup Thorcon, and writer of "Why Nuclear Power Has Been a Flop." Devanney designed Navy warships and ran a successful oil tanker business before entering the nuclear industry. He also wrote a fascinating book about oil tanker safety. Devanney played a part in devising a method to prevent interior corrosion on double-hulled ships and advocated for having two independent engines since most ship disasters start with the loss of engine power.
The thesis behind Thorcon is to build nuclear power plants in shipyards to reduce costs. Their proposed reactor is molten salt-cooled with clever passive safety features. They are currently applying to site the first copy in Indonesia.
Devanney has written extensively about his thoughts on regulation and has codified them in a proposed law, called the Nuclear Reorganization Act. The underlying driver across all his proposals is changing incentives for fission power plants, especially regarding risk and liability.
High Level Vision
Power plant builders and operators need to own the risk of accidents, but laws and regulations also need to reduce dollar losses.
Operators and insurers can limit dollar losses through clever engineering design that keeps radiation exposure manageable, along with some defined payments for harm.
The low-level work of licensing and supervision will be done by independent (and competing) private regulators. Licensing depends on testing and results rather than following a process.
The Radiation Damage Question
Devanney's idea to reduce dollar losses centers on an arcane radiation health impact model, called Sigmoid No Threshold (SNT). SNT assumes low radiation levels are significantly less harmful to humans than current models suggest due to DNA self-repair.
The current industry model, Linear No Threshold (LNT), assumes lower levels of radiation exposure are more harmful and that lifetime dosage is cumulative. The current policy of minimizing radiation exposure, known as ALARA (As Low As Reasonably Achievable), is a constant target of criticism for Devanney.
Assuming less harm at low levels of exposure allows significantly more design flexibility and opens up new possibilities.
The switch to SNT is the keystone of the argument that unlocks other incentive changes and engineering tradeoffs. I'm not a radiation expert, so I can't endorse its validity, but the book is a great place to start for those curious.
Proposed Changes
These are the most important adjustments to the nuclear regulation paradigm:
-
Facility Operators Must Have Private Insurance for Accident Risk
Repeal any laws like Price-Anderson that involve the government in insurance or liability caps.
-
Use Sigmoid No Threshold (SNT) to Evaluate Radiation Risk
See above.
-
Mandate a Buffer Zone
Radiation dosage decreases significantly with distance. A buffer zone could keep public radiation exposure below very harmful levels even in worst-case accidents. Buffers might be 2 to 5 kilometers for larger power plants and less for small or micro reactors.
-
Fixed Payments for Harm
Have government-mandated payments for exposure above the threshold, based on severity. Set payments ahead of time to limit civil cases and allow private insurers to price risk confidently.
-
Results Over Process
Many highly regulated sectors have the curse where the process of doing something is under scrutiny rather than the outcome. Shifting to functional testing creates clear pathways for new methods or materials, improving costs and quality.
-
Isolated Testing Park
There should be a testing park for exact replicas of planned facilities. Companies can correct failures before deployment at a commercial site. Hanford, WA, is suggested as one possible location for such a park.
-
Personal and Organization Incentive Changes
A tax on every megawatt-hour produced should fund regulatory bodies rather than the current policy of operators paying by the labor hour. Incentives would push regulators to maximize safe hours of electricity production.
Industry executives and shareholders should face more personal financial liability for fines and accidents, as well.
New Paths to Try
The exciting part of Devanney's proposal is the possibility of new technical paths. Maybe there are inexpensive ways to dramatically reduce accident severity (even if the accident rate might increase). Or some combination of plant design, fuel choice, or buffer zones allows the industry to abstract safety, easing the verification problem. Engineers will study the worst-case accident in the same detail as today, but the thousands of sub-cases dealing with individual equipment might not be necessary.
There are similar ideas to abstract safety, other than distance buffers. Deep Fission wants to bury reactors in boreholes. Last Energy proposes using 1000 tons of steel to armor its nuclear island. Others might only use small amounts of fuel.
There could be a breakthrough in the 74th year of effort on sodium-cooled reactor commercialization (or helium, lead, CO2, or molten salt). But it seems more likely that we will make rapid progress by changing paradigms around safety.
Altering the Public Relationship
Implicit in these suggestions is that nuclear safety communication with the public has to change. Devanney is adamant that it is impossible to eliminate accidents because reactors are so complex and so many things can go wrong. The messaging shifts to "Accidents will sometimes happen, but are extremely rare, and this is how we ensure your safety in case of one." The buffer zone is an attempt at simplifying the public safety case. The facility operator only has to show that civilians getting a dangerous dose is nearly impossible.
The public may or may not accept the new framework. At least the message is logical, and safety abstraction is one of the only promising paths to significantly reducing nuclear power costs.
To Second Chances
One of the biggest takeaways is that there is limited flexibility for regulatory frameworks. Any plan has to:
-
Respect the accident potential of fission power plants. An incredible amount of effort from the people who design and operate our existing fission power plants keeps them safe, they are not inherently safe.
-
Accept that nuclear power plants are complex beasts, and it is challenging to prevent all accidents.
-
Create incentives that both encourage safety and promote inexpensive nuclear power without adverse regulator interactions.
-
Ensure paths for successful and safe adoption of new technology.
-
Maintain the trust and buy-in of the public.
There is an opportunity for improvement over the status quo despite these constraints.
A Nuclear Fission Regulatory Blank Slate
2025 October 26 Twitter Substack See all postsWhat if nuclear regulations were reset?
What Would You Do?
It is easy to say, "Nuclear should be regulated more sensibly!" But what if you had a blank slate to redo all nuclear fission regulation? How would you harness the power of fission while keeping the public safe?
There are two paths I find interesting. I'm sure there are many more possibilities, but I chose these because they also help explain a lot of the difficulty (and promise) in regulating fission power plants.
The Difficulties of Nuclear Regulation
The biggest challenge in regulating nuclear is that it's complex, and the potential impact of accidents is high. Outsiders probably underestimate the complexity of running a nuclear power plant by at least two orders of magnitude. A nuclear power plant has 20x more employees than a natural gas power plant of the same size, and it's not only because of the NRC.
Below is an excerpt from a tweet by nuclear entrepreneur Josh Payne that is a response to a hit piece on a different nuclear startup:
Several factors contribute to this complexity.
Managing Reactivity
Splitting uranium atoms requires neutrons to start the reaction. And neutrons interact with all matter, not just uranium atoms.
The actual design of the reactor has to allow neutrons released from one atom splitting to hit the next. One side effect of this is that small reactors typically require more fuel because neutrons are less likely to hit anything useful. The required neutron flux is not steady state, either. It depends on where the reactor is in the fuel cycle and the power plant's current output.
Power plants rely on the natural fission of uranium to initiate reactions, which requires a brake to prevent chain reactions from going too fast. A tiny change can alter the conditions and disrupt the neutron balance. Even a change in the feed water rate can increase the reaction rate by removing more heat. Modern water-cooled power plants control neutron reactivity by mixing boric acid in the reactor coolant water and/or by adjusting neutron-absorbing control rods. Other types of reactors typically use devices similar to control rods.
As you might imagine, feed water, reactor, control rod, and boric acid concentration management subsystems are very complex. They require constant maintenance, vigilance, and care. Some rarely deployed designs try to handle this balance more passively, but it almost always requires more engineered (and expensive) fuel and/or exotic materials. The market has vastly favored organizing systems around cheap fuel, water, and boron.
There is No "Off" Button for Fission Reactors
Fission reactions are messy. A neutron can go any number of directions and hit different types of atoms. When atoms split, they don't always do so the same way; there are many possible products. The reactions also release other types of radiation, such as gamma rays, of varying intensities.
Reactor designs have multiple ways to stop fission if things go wrong. Besides control rods or boric acid, water-cooled reactors stop if all the coolant boils off because there is no liquid water to moderate neutrons.
What is difficult to control are the unstable fission products in the fuel that undergo decay and release heat. Bad things can happen if heat isn't constantly removed, like in the Fukushima accident. Fuel melting, hydrogen explosions, and pressure buildup are some possible issues. Maintaining heat removal under any possible circumstance or disaster takes an incredible amount of effort in design and operations, even if the heat removal method is passive.
Off-Target Neutron Reactions
Materials in the reactor loop must withstand neutrons, avoid hindering the reaction, and avoid nasty side reactions.
Since neutrons can alter materials, many metals can undergo accelerated corrosion. Neutrons and gamma rays sometimes split water molecules into hydrogen and oxygen, gases that also cause corrosion. Many of France's nuclear power plants were offline during the 2022 European Energy Crisis due to poorly managed corrosion. Materials need to be carefully selected and monitored to prevent accidents.
The materials also can't absorb too many neutrons because it hinders the reaction. Part of the reason why the Germans were behind in building atomic bombs in World War II is that they didn't realize natural graphite has trace amounts of boron that absorb too many neutrons. Any materials must be pure enough to avoid these issues.
Neutrons can also turn materials into really nasty isotopes. A classic example is in valves. It is challenging to produce valves that don't leak under high pressure. Valve makers use a miracle metal alloy, called stellite, in valve seats to produce competent valves. Unfortunately, stellite contains cobalt. When cobalt gets hit by neutrons, it forms an extremely dangerous cobalt-60 isotope that produces high-energy gamma radiation. Little pieces of stellite rub off during valve operation, travel through the reactor, absorb neutrons, and migrate to traps where they can harm employees. What is worse, leaky reactor loop valves or random sources of high-energy gamma rays in the workplace? Not an easy question to answer or deal with!
The upshot is that neutrons and radiation add engineering complexity that drastically limits acceptable materials and fluids. Materials going into reactors can cost many times more than commercial grades, and the engineering and testing burden is many times greater.
Radiation Impact Variance
The types and intensity of radiation that a fission power plant can produce are remarkable. That makes protecting employees and the public difficult. Even worse, the method of dosing can increase the effect by orders of magnitude. Ingestion can be particularly harmful. An open Coke can or bag of chips in certain parts of the plant is a crisis and a safety failure.
Another confounding factor is that the impact of a dosage isn't always immediate, nor is it always the same. These factors can add legal and public relations challenges.
Responsibility, Non-linearity, Verification, and Incentives
One thing to lament about fission power plants is that the complexity means there is no magic software module that can handle all safety cases. One small group could put it together once and then copy it over and over. It could undergo much more significant scrutiny than other parts of the design, construction, and operation. Competing technologies, such as solar, wind, or natural gas, can abstract most safety issues to a piece of software in a manufactured piece of equipment, like an inverter or turbine control box.
Many aspects of nuclear safety are non-linear: dosages, accident interactions, and earthquake impacts. Ideally, in safety, you can use a worst-case basic calculation to set the baseline. These "worst-case" numbers are quick and easy to do. A detailed model or study can reduce the safety factor if the simple safety factor calculation is too expensive to implement. This algorithm does not work as well for non-linear effects. A model of a simple 8.0 earthquake on individual pieces of equipment might suggest everything is fine, but a more systemic model that includes equipment interactions can identify issues. We can also see this in trying to judge exposure effects from radiation, showcased in an argument between several experienced nuclear engineers and a nuclear startup CEO:
Three smart individuals get answers that are all orders of magnitude apart. The radiation exposure depends on the time the spent fuel has been out of the reactor, the actual material, and the distance from the body to the source. None of these effects is linear, so answers depend heavily on starting assumptions. These three almost certainly had different starting places. Lengthy, uniform standards for important safety calculations are the norm.
The lack of abstraction, non-linearity, and the nearly infinite possible interactions within fission power plants make verification of safety incredibly challenging, not only for regulators but also for executives and other stakeholders. Verification is extremely costly, but necessary for the confidence of the public, regulators, and executives.
Challenging verification tends to breed risk aversion. Someone must authorize the entire project. Approving something that contributes to a serious accident creates extreme guilt, kills careers, puts the individual through a years-long legal wringer, and never goes away. Engineers must check their calculations repeatedly. Even the slightest concern is enough to delay.
Individual responsibility is necessary to keep skin in the game for the people designing and operating power plants, but absorbing verification efforts, regulatory risk, and accident liability is expensive. The nuclear power industry has no shortage of bankruptcies. A nuclear accident is one of the most costly mistakes a capitalist can make. It ultimately proves challenging to maintain the correct incentives.
Paths Forward
Given these difficulties, what regulatory frameworks might one try? We must be clear-eyed about risk and incentives to improve the design and engineering of nuclear power plants.
Industry-led Standards
The first approach is more conservative and applies best practices that have reduced risk over the decades of operating nuclear power plants.
The basic structure would be:
The logic that underpins this path is:
Nuclear Power Plant Operations are Complex but Mostly Understood
Power plant operation and understanding of accident cases have improved since the 1970s. Experience and probabilistic risk analysis are behind the improvement.
Remove Complex and Expensive Interaction with Regulators
The gains in operational efficiency and safety are still time-consuming and hard-won, and the complexity does not interact well with a regulator like the NRC, which lacks the desire or manpower to keep up. Wild goose chases and delays are costly for the nuclear industry.
Private Regulation Can Improve Safety
The nuclear industry already has a private regulator in the Institute for Nuclear Power Operations (INPO). The NRC even delegates operator training standards to INPO, the most critical part of safety. Many NRC statutes contain little detail, and it is industry groups, like INPO, that determine the actual processes. It is also common for INPO and the industry to have standards that are more strict than the NRC's. Individual companies are also likely to propose fixes that go beyond standards when reporting issues to the NRC.
The incentives here are clear. Interacting with the NRC has massive costs. It is often cheaper to be stricter because the government is less likely to get involved. The operator can choose their preferred solutions.
A private regulator without the NRC would be more efficient by avoiding expensive requirements that may not add much safety value. It would have the same incentive to favor safety because whatever reappears after an accident will likely be much worse than the NRC ever was. Nuclear plants are so valuable and long-lived that they must be protected from intervention to maximize shareholder value.
The operators would have to pay for this regulator, and INPO's yearly budget is just over $100 million.
Shared Accident Pain
One unfortunate feature of the nuclear power industry is that an accident anywhere can trigger increased regulatory scrutiny, shutdown of viable facilities, and cost increases. Having a private regulator with agreed-upon rules, information sharing, and oversight for all parties reduces the risk of a poorly run plant ruining it.
Maintains Public Outreach Success
Nuclear power has significantly improved its image in the last few decades by reducing accidents and near misses while providing extensive analysis and transparency about safety measures to the public. None of these attributes would change.
The net result of the changes should be lower costs and better safety. Standards will be stricter, but operators are free to use the most cost-effective measures. Rules for licensing new facilities and operating existing ones will be easier to follow.
Maintaining Innovation?
New designs, especially those that aren't water-cooled, could be challenging to test or license if the private regulator and industry have to agree.
If the market truly favors water-cooled reactors, then there are limited opportunity costs. Performance-based regulation can still unlock many incremental improvements in water-cooled reactors through methods like qualifying cheaper materials or simpler safety systems.
There should be test parks that improve understanding of reactors with other types of coolant, fuels, or configurations, in case water-cooled reactors are not the final form.
Preserve and Expand
The primary benefits of this route are making it cheaper and safer to operate existing power plants, reducing the difficulty in building new water-cooled reactors, and pushing incremental improvements. Maintaining some semblance of entry for technologies like sodium, lead, salt, or helium-cooled reactors requires thought. These reactors have almost all been built at commercial scale and failed to beat water-cooled reactors. Many want to try again, and they should be allowed. An industry-led, incremental approach is very compatible with the history of nuclear power policy and deployment. The trick is to find balance.
Maximizing Innovation and Altering Public Nuclear Views
One of the only coherent criticisms of the current system comes from Jack Devanney, founder of the nuclear startup Thorcon, and writer of "Why Nuclear Power Has Been a Flop." Devanney designed Navy warships and ran a successful oil tanker business before entering the nuclear industry. He also wrote a fascinating book about oil tanker safety. Devanney played a part in devising a method to prevent interior corrosion on double-hulled ships and advocated for having two independent engines since most ship disasters start with the loss of engine power.
The thesis behind Thorcon is to build nuclear power plants in shipyards to reduce costs. Their proposed reactor is molten salt-cooled with clever passive safety features. They are currently applying to site the first copy in Indonesia.
Devanney has written extensively about his thoughts on regulation and has codified them in a proposed law, called the Nuclear Reorganization Act. The underlying driver across all his proposals is changing incentives for fission power plants, especially regarding risk and liability.
High Level Vision
Power plant builders and operators need to own the risk of accidents, but laws and regulations also need to reduce dollar losses.
Operators and insurers can limit dollar losses through clever engineering design that keeps radiation exposure manageable, along with some defined payments for harm.
The low-level work of licensing and supervision will be done by independent (and competing) private regulators. Licensing depends on testing and results rather than following a process.
The Radiation Damage Question
Devanney's idea to reduce dollar losses centers on an arcane radiation health impact model, called Sigmoid No Threshold (SNT). SNT assumes low radiation levels are significantly less harmful to humans than current models suggest due to DNA self-repair.
The current industry model, Linear No Threshold (LNT), assumes lower levels of radiation exposure are more harmful and that lifetime dosage is cumulative. The current policy of minimizing radiation exposure, known as ALARA (As Low As Reasonably Achievable), is a constant target of criticism for Devanney.
Assuming less harm at low levels of exposure allows significantly more design flexibility and opens up new possibilities.
The switch to SNT is the keystone of the argument that unlocks other incentive changes and engineering tradeoffs. I'm not a radiation expert, so I can't endorse its validity, but the book is a great place to start for those curious.
Proposed Changes
These are the most important adjustments to the nuclear regulation paradigm:
Facility Operators Must Have Private Insurance for Accident Risk
Repeal any laws like Price-Anderson that involve the government in insurance or liability caps.
Use Sigmoid No Threshold (SNT) to Evaluate Radiation Risk
See above.
Mandate a Buffer Zone
Radiation dosage decreases significantly with distance. A buffer zone could keep public radiation exposure below very harmful levels even in worst-case accidents. Buffers might be 2 to 5 kilometers for larger power plants and less for small or micro reactors.
Fixed Payments for Harm
Have government-mandated payments for exposure above the threshold, based on severity. Set payments ahead of time to limit civil cases and allow private insurers to price risk confidently.
Results Over Process
Many highly regulated sectors have the curse where the process of doing something is under scrutiny rather than the outcome. Shifting to functional testing creates clear pathways for new methods or materials, improving costs and quality.
Isolated Testing Park
There should be a testing park for exact replicas of planned facilities. Companies can correct failures before deployment at a commercial site. Hanford, WA, is suggested as one possible location for such a park.
Personal and Organization Incentive Changes
A tax on every megawatt-hour produced should fund regulatory bodies rather than the current policy of operators paying by the labor hour. Incentives would push regulators to maximize safe hours of electricity production.
Industry executives and shareholders should face more personal financial liability for fines and accidents, as well.
New Paths to Try
The exciting part of Devanney's proposal is the possibility of new technical paths. Maybe there are inexpensive ways to dramatically reduce accident severity (even if the accident rate might increase). Or some combination of plant design, fuel choice, or buffer zones allows the industry to abstract safety, easing the verification problem. Engineers will study the worst-case accident in the same detail as today, but the thousands of sub-cases dealing with individual equipment might not be necessary.
There are similar ideas to abstract safety, other than distance buffers. Deep Fission wants to bury reactors in boreholes. Last Energy proposes using 1000 tons of steel to armor its nuclear island. Others might only use small amounts of fuel.
There could be a breakthrough in the 74th year of effort on sodium-cooled reactor commercialization (or helium, lead, CO2, or molten salt). But it seems more likely that we will make rapid progress by changing paradigms around safety.
Altering the Public Relationship
Implicit in these suggestions is that nuclear safety communication with the public has to change. Devanney is adamant that it is impossible to eliminate accidents because reactors are so complex and so many things can go wrong. The messaging shifts to "Accidents will sometimes happen, but are extremely rare, and this is how we ensure your safety in case of one." The buffer zone is an attempt at simplifying the public safety case. The facility operator only has to show that civilians getting a dangerous dose is nearly impossible.
The public may or may not accept the new framework. At least the message is logical, and safety abstraction is one of the only promising paths to significantly reducing nuclear power costs.
To Second Chances
One of the biggest takeaways is that there is limited flexibility for regulatory frameworks. Any plan has to:
Respect the accident potential of fission power plants. An incredible amount of effort from the people who design and operate our existing fission power plants keeps them safe, they are not inherently safe.
Accept that nuclear power plants are complex beasts, and it is challenging to prevent all accidents.
Create incentives that both encourage safety and promote inexpensive nuclear power without adverse regulator interactions.
Ensure paths for successful and safe adoption of new technology.
Maintain the trust and buy-in of the public.
There is an opportunity for improvement over the status quo despite these constraints.