The competition to deploy small modular reactors is fierce and it’s global. My new mini-documentary gives a fast-paced tour of the promise and challenges of SMRs.
The real competition in SMRs isn’t about designs. It’s about who can execute predictable manufacturing, secure fuel supply and hit cost targets that make firm power scalable again.
The companies that treat SMRs like an industrial product instead of a bespoke project will define the next phase of nuclear growth. Bryce’s documentary captures exactly why this race matters.
The killer point - AI is changing so quickly that data center based AI will be obsolete in just a few more years. You will run your own instance on a server in your home office five years from now.
Surely that distributes but doesn't eliminate the power requirement. 5 years is unlikely to make a .. pick a number ... 100% - 1000% reduction in power needs per compute unit.
Numbers quoted by the company using and promoting the new paradigm are 1000 to 1M times faster than current tech, and replacing a data center with the equivalent of two Intel NUCs. So pretty radical however you cut it. The down side is that this approach requires customizing parts of the OS for every application (program) at present. Search Fractal Computing - Co. website and Substack blogs.
No. Go read their Substacks. They have just gone back to basics. The computer industry has coasted on the back of Moores Law and Microsoft for three decades. If you optimize the way data is handled again, as was done in the first decade of microcomputer, you eliminate the massive waste of processor cycles spent waiting for data to arrive.
Just had a read. I'm an old mainframe performance and design person and we, perhaps unusually but through economic necessity, were regularly focussed on improving locality of reference to, indeed, reduce cpu wait time/maximise relevance of data read from disk.
Fractal sounds like a great approach. In the 1980s & 90s we didn't have the luxury of computing resources to provide various function relevant views/object but we got as close as we could. :-) All in IBM 370 assembler, too.
Seemingly forgotten is that the U.S. Navy (and others) have been successfully powering ships and submarines with nuclear since the Nautilus was launched in 1954, followed by the launch of the first nuclear powered merchant ship, the N.S. Savannah in 1959.
Seems like their subsequent success to the present day would stifle a lot of the arguments over the "safety" of Nuclear power.
Agreed. It's worth noting that BWX Technologies has been building reactors for the US Navy for decades. That same company is now planning to deploy SMRs for commercial use. It will be a formidable competitor.
Of course that has been thought about long and hard. These folks are professionals working in a hostile regulatory environment. Mainly it's inaccessible. It's so reliable and intrinsically safe you can bury it in the ground for 20-50 years.
And in the course of my preparation for this book, I rather casually reviewed what had happened at Chernobyl, because I regard Chernobyl as the largest manmade disaster that I knew about. What I discovered stunned me. Chernobyl was a tragic event, but nothing remotely close to the global catastrophe that I was imagining. About 50 people had died in Chernobyl, roughly the number of Americans that die every day in traffic accidents. I don’t mean to be gruesome, but it was a setback for me. You can’t write a novel about a global disaster in which only 50 people die.
I was undaunted. I began to research other kinds of disasters that might fulfill my novelistic requirements, and that’s when I began to realize how big our planet really is and how resilient its systems ordinarily seem to be. Even though I wanted to create a fictional catastrophe of global proportions, I found it hard to come up with a credible example. I couldn’t actually come up with anything that I would believe, so in the end, I set the book aside and wrote something else.
But the shock that I had experienced reverberated in me for a while, because what I’d been led to believe about Chernobyl was not merely wrong. It was astonishingly wrong. Let’s review that for a minute.
And at least 30 of those deaths were military personnel who were commanded to fly over the fire in helicopters and dump sand, with not the best protective gear.
Fair enough. I lived in New Zealand before 2003, and don't recall media there offering any detail to speak of. My knowledge came from reading a book by a British journalist. I'm not a typical Kiwi since I've been an enthusiastic nuclear supporter since I reached adulthood.
I would love to see a fusion reactor that did what everyone is hoping for. Not holding my breath though for many reasons, some of which are based on the extreme difficulty and also the fact that there seem to be intractable physical problems (as there are with hydrogen) - basic laws of physics that probably mean it will just not happen. With fusion its that most of the energy out is fast neutrons, best utilized by blanketing the vessel with U238 and reprocessing it after a suitable delay, to extract the Plutonium...
Problems aren't technical, but political or ideological, that said, I think scaling factors make large reactors a better proposition thermodynamically and economically. We have a problem getting things done in the west, and seem myopic about what others have done. The Great Satan (Russia) has their VVER-1200 and the floating SMRs to use in remote places - hat tip where it is due. Russia's modelling of 'global warming' is also the only one close to actual measurements versus the mindless catastrophism in the west. Those that live in physical reality will always have an advantage over policy driven by rainbow coloured unicorns.
If data centres require GWs and GWs of power then wouldn't the big boys be more suitable than 10 300 MWe SMRs for example? Japan built an ABWR (unit 6) in 39 months in the 1990s. OpenAI and NGreedia talk up 10GW for example. The question is: when will the techbros step up and directly fund the ridiculous power requirements by either constructing nat gas/coal/nuclear plants or support the uranium fuel cycle? Someone needs to inform Altman and Huang that there's a slight structural supply deficit in uranium at the minute.
Very exciting video. Great job. With all the development of SMRs where are they going to get steam turbines. You have written a lot about equipment shortages. Thanks
Hi Larry. Yes, all of the hardware -- from breakers to switchgear and transformers -- is in short supply right now. All of these projects -- gas, nuclear, etc. -- will take time to build and get online.
The Westinghouse AP1000 reactors at Vogtle cost about $13/watt. When the first unit was undergoing cold hydraulic tests a vibration in a pipe was found. That could have been corrected with a $30,000 brace, but NRC required a $3 million license review.
China is building upsized CAP1200 using ripped-off intellectual property for $2.69/watt. They have plans for CAP1400 and they have CAP1700 under design. So behemoths aren't intrinsically expensive. It looks like in USA $3/watt is construction cost, and $10/watt is the cost of doing battle with environists (the ones with no mental in the middle) and the NRC. Finland's Olkiluoto 3 was indeed expensive, but when it came online retail prices at the meter decreased 60%.
Terrapower is building a 300 MWe Natrium system at Kemmerer, WY around a GE/Hitachi PRISM — Power Reactor Innovate Small Modular — reactor. It's an upsized version of the liquid sodium-cooled inherently safe EBR-II that GE designed thirty years ago but nobody was allowed to build. NRC was designed to prevent reactors, and it worked fantastically well: Two reactors have been built since NRC was split from AEC. An important component of Natrium (Latin for "salt") is a molten-salt thermal store between the reactor and steam generator. It can provide up to 500 MWe for up to five hours. More importantly, output from it can be changed rapidly to cope with nearby windmills' unreliability. Another important result is that sodium is never near water.
The next important problem to tackle is 5%-used spent fuel, intentionally pejoratively called nuclear waste. 9.26% of fission products need custody for 300 years. Half the rest are innocuous before thirty years and the remainder aren't even radioactive. Unused fuel is dangerous for 300,000 years, but turning it into electricity ends that problem. We know several ways to separate fission products from unused fuel, but Jimmuh Cahtuh insisted that if we didn't do it, other nations wouldn't develop nuclear weapons. India and Pakistan and North Korea didn't get the memo.
The Nuclear Waste Policy Act of 1982 imposed a 0.1¢/kWh fee on nuclear reactors to "do something" about it. The Fund now stands at $43+ billion, and nothing has been done about it. Unfortunately, the Act explicitly forbade the Fund to be used for spent fuel processing. That needs to change. Write to you congresscritters. Oklo is developing a pyroelectric processing pilot at Oak Ridge with their own money.
Read about the Integral Fast Reactor in "Plentiful Energy" by Charles E. Till and Yoon Il Chang. You can get it from Amazon or read it at http://vandyke.mynetgear.com/Plentiful_Energy.pdf where Dr. Chang gave me written permission to post it. More details in my book "Where Will We Get Our Energy? A Comprehensive Quantitative System Engineering Study of the Relationship between Climate, Science, and Technology." Everything quantified. No vague handwaving. 350 bibliographic citations allow readers to verify I didn't simply make up stuff.
"The Westinghouse AP1000 reactors at Vogtle cost about $13/watt. When the first unit was undergoing cold hydraulic tests a vibration in a pipe was found. That could have been corrected with a $30,000 brace, but NRC required a $3 million license review."
Lets hope the Trump administration and Congress are able to turf the antinuclear activists who have been running the NRC and rewrite its mission to use SNT - real science, as the basis for safety regulations. https://jackdevanney.substack.com/p/replacing-lnt-with-snt
The real competition in SMRs isn’t about designs. It’s about who can execute predictable manufacturing, secure fuel supply and hit cost targets that make firm power scalable again.
The companies that treat SMRs like an industrial product instead of a bespoke project will define the next phase of nuclear growth. Bryce’s documentary captures exactly why this race matters.
Make them smaller, cheaper, safer, and more portable.
I want one for my home!
The killer point - AI is changing so quickly that data center based AI will be obsolete in just a few more years. You will run your own instance on a server in your home office five years from now.
Surely that distributes but doesn't eliminate the power requirement. 5 years is unlikely to make a .. pick a number ... 100% - 1000% reduction in power needs per compute unit.
Numbers quoted by the company using and promoting the new paradigm are 1000 to 1M times faster than current tech, and replacing a data center with the equivalent of two Intel NUCs. So pretty radical however you cut it. The down side is that this approach requires customizing parts of the OS for every application (program) at present. Search Fractal Computing - Co. website and Substack blogs.
This is the new "wafer" tech that a start-up is talking about?
No. Go read their Substacks. They have just gone back to basics. The computer industry has coasted on the back of Moores Law and Microsoft for three decades. If you optimize the way data is handled again, as was done in the first decade of microcomputer, you eliminate the massive waste of processor cycles spent waiting for data to arrive.
Just had a read. I'm an old mainframe performance and design person and we, perhaps unusually but through economic necessity, were regularly focussed on improving locality of reference to, indeed, reduce cpu wait time/maximise relevance of data read from disk.
Fractal sounds like a great approach. In the 1980s & 90s we didn't have the luxury of computing resources to provide various function relevant views/object but we got as close as we could. :-) All in IBM 370 assembler, too.
I will, thanks.
Seemingly forgotten is that the U.S. Navy (and others) have been successfully powering ships and submarines with nuclear since the Nautilus was launched in 1954, followed by the launch of the first nuclear powered merchant ship, the N.S. Savannah in 1959.
Seems like their subsequent success to the present day would stifle a lot of the arguments over the "safety" of Nuclear power.
Agreed. It's worth noting that BWX Technologies has been building reactors for the US Navy for decades. That same company is now planning to deploy SMRs for commercial use. It will be a formidable competitor.
What are the security issues in having fissionable materials in many scattered sites vs protecting at one large plant?
Of course that has been thought about long and hard. These folks are professionals working in a hostile regulatory environment. Mainly it's inaccessible. It's so reliable and intrinsically safe you can bury it in the ground for 20-50 years.
That's a good question. And it's one that deserves attention. I don't have a good answer. Like all power plants, they will need security.
Insightful analysis, as always.
Public Precseption
An Evening with Michael Crichton
“States of Fear: Science or Politics?”
November 15, 2005
https://www.independent.org/news/event-transcripts/an-evening-with-michael-crichton/#2
(Snip)
And in the course of my preparation for this book, I rather casually reviewed what had happened at Chernobyl, because I regard Chernobyl as the largest manmade disaster that I knew about. What I discovered stunned me. Chernobyl was a tragic event, but nothing remotely close to the global catastrophe that I was imagining. About 50 people had died in Chernobyl, roughly the number of Americans that die every day in traffic accidents. I don’t mean to be gruesome, but it was a setback for me. You can’t write a novel about a global disaster in which only 50 people die.
I was undaunted. I began to research other kinds of disasters that might fulfill my novelistic requirements, and that’s when I began to realize how big our planet really is and how resilient its systems ordinarily seem to be. Even though I wanted to create a fictional catastrophe of global proportions, I found it hard to come up with a credible example. I couldn’t actually come up with anything that I would believe, so in the end, I set the book aside and wrote something else.
But the shock that I had experienced reverberated in me for a while, because what I’d been led to believe about Chernobyl was not merely wrong. It was astonishingly wrong. Let’s review that for a minute.
(Snip)
And at least 30 of those deaths were military personnel who were commanded to fly over the fire in helicopters and dump sand, with not the best protective gear.
Yup!
His point was Chernobyl was not as it was shown in The Media.
Fair enough. I lived in New Zealand before 2003, and don't recall media there offering any detail to speak of. My knowledge came from reading a book by a British journalist. I'm not a typical Kiwi since I've been an enthusiastic nuclear supporter since I reached adulthood.
For what its worth I did not know that either until I saw his talk.
I'm waiting for fusion reactors. that solves a lot of the problems.
I would love to see a fusion reactor that did what everyone is hoping for. Not holding my breath though for many reasons, some of which are based on the extreme difficulty and also the fact that there seem to be intractable physical problems (as there are with hydrogen) - basic laws of physics that probably mean it will just not happen. With fusion its that most of the energy out is fast neutrons, best utilized by blanketing the vessel with U238 and reprocessing it after a suitable delay, to extract the Plutonium...
Fuel for SMR's. Anyone working a Fission version? Actually Fission Reactors in general.
Do you mean fusion? All these reactors are fission reactors.
I'm going to go stand in the corner now. :-)
Ok-- now you're just fission for sympathy! 🫣
LOL
switch all the wasted funding on W&S to this journey.
Problems aren't technical, but political or ideological, that said, I think scaling factors make large reactors a better proposition thermodynamically and economically. We have a problem getting things done in the west, and seem myopic about what others have done. The Great Satan (Russia) has their VVER-1200 and the floating SMRs to use in remote places - hat tip where it is due. Russia's modelling of 'global warming' is also the only one close to actual measurements versus the mindless catastrophism in the west. Those that live in physical reality will always have an advantage over policy driven by rainbow coloured unicorns.
If data centres require GWs and GWs of power then wouldn't the big boys be more suitable than 10 300 MWe SMRs for example? Japan built an ABWR (unit 6) in 39 months in the 1990s. OpenAI and NGreedia talk up 10GW for example. The question is: when will the techbros step up and directly fund the ridiculous power requirements by either constructing nat gas/coal/nuclear plants or support the uranium fuel cycle? Someone needs to inform Altman and Huang that there's a slight structural supply deficit in uranium at the minute.
Very exciting video. Great job. With all the development of SMRs where are they going to get steam turbines. You have written a lot about equipment shortages. Thanks
Hi Larry. Yes, all of the hardware -- from breakers to switchgear and transformers -- is in short supply right now. All of these projects -- gas, nuclear, etc. -- will take time to build and get online.
Talented family!
Thanks.
The Westinghouse AP1000 reactors at Vogtle cost about $13/watt. When the first unit was undergoing cold hydraulic tests a vibration in a pipe was found. That could have been corrected with a $30,000 brace, but NRC required a $3 million license review.
China is building upsized CAP1200 using ripped-off intellectual property for $2.69/watt. They have plans for CAP1400 and they have CAP1700 under design. So behemoths aren't intrinsically expensive. It looks like in USA $3/watt is construction cost, and $10/watt is the cost of doing battle with environists (the ones with no mental in the middle) and the NRC. Finland's Olkiluoto 3 was indeed expensive, but when it came online retail prices at the meter decreased 60%.
Terrapower is building a 300 MWe Natrium system at Kemmerer, WY around a GE/Hitachi PRISM — Power Reactor Innovate Small Modular — reactor. It's an upsized version of the liquid sodium-cooled inherently safe EBR-II that GE designed thirty years ago but nobody was allowed to build. NRC was designed to prevent reactors, and it worked fantastically well: Two reactors have been built since NRC was split from AEC. An important component of Natrium (Latin for "salt") is a molten-salt thermal store between the reactor and steam generator. It can provide up to 500 MWe for up to five hours. More importantly, output from it can be changed rapidly to cope with nearby windmills' unreliability. Another important result is that sodium is never near water.
The next important problem to tackle is 5%-used spent fuel, intentionally pejoratively called nuclear waste. 9.26% of fission products need custody for 300 years. Half the rest are innocuous before thirty years and the remainder aren't even radioactive. Unused fuel is dangerous for 300,000 years, but turning it into electricity ends that problem. We know several ways to separate fission products from unused fuel, but Jimmuh Cahtuh insisted that if we didn't do it, other nations wouldn't develop nuclear weapons. India and Pakistan and North Korea didn't get the memo.
The Nuclear Waste Policy Act of 1982 imposed a 0.1¢/kWh fee on nuclear reactors to "do something" about it. The Fund now stands at $43+ billion, and nothing has been done about it. Unfortunately, the Act explicitly forbade the Fund to be used for spent fuel processing. That needs to change. Write to you congresscritters. Oklo is developing a pyroelectric processing pilot at Oak Ridge with their own money.
Read about the Integral Fast Reactor in "Plentiful Energy" by Charles E. Till and Yoon Il Chang. You can get it from Amazon or read it at http://vandyke.mynetgear.com/Plentiful_Energy.pdf where Dr. Chang gave me written permission to post it. More details in my book "Where Will We Get Our Energy? A Comprehensive Quantitative System Engineering Study of the Relationship between Climate, Science, and Technology." Everything quantified. No vague handwaving. 350 bibliographic citations allow readers to verify I didn't simply make up stuff.
"The Westinghouse AP1000 reactors at Vogtle cost about $13/watt. When the first unit was undergoing cold hydraulic tests a vibration in a pipe was found. That could have been corrected with a $30,000 brace, but NRC required a $3 million license review."
Ah! The Joys Of Regulation!
Lets hope the Trump administration and Congress are able to turf the antinuclear activists who have been running the NRC and rewrite its mission to use SNT - real science, as the basis for safety regulations. https://jackdevanney.substack.com/p/replacing-lnt-with-snt
Nice work!
Thanks. Mary did a great job.
Kudos to you both. You’re doing rational work in an irrational world.
Brilliant documentary. Congratulations to you and to your daughter.
That’s kind. Thanks.