4 ms·

Solar is very, very cheap and almost totally worthless without storage. Storage is extremely expensive. Nuke is extremely cheap to generate -once its built. The

Solar is very, very cheap and almost totally worthless without storage. Storage is extremely expensive. Nuke is extremely cheap to generate -once its built. The cost of nuke energy is not because the technology is complex or because resources are scarce. It's because we have very, very burdensome regulations around nuclear reactors (for good reason!) and each nuke plant is a bespoke effort which gets recertified each time. This is enormously expensive. There is reason to believe that small modular nuke plants will vastly reduce this cost. That means we might have a path to cheap nuke, but there is no immediate path to cheap storage barring a technological revolution (not just incremental improvements) in battery tech. In the long run solar power will kill fossil fuels, but we desperately need a bridge to get us there and not destroy the carbon balance in the atmosphere. Nuke is that bridge.

ericd 1y ago
Are you so sure that storage is so expensive? It’s been coming down the cost curve extremely quickly, such that opinions formed even an year ago are severely outdated, and it’s now solar+storage that’s being favorably compared to replacing nat gas plants, not just solar itself.
- johncolanduoni 1y ago
  Storage that is good enough to replace peaker plants, and storage that is good enough to handle seasonal variations in insolation are completely different ballgames. The lithium battery chemistry in your phone will self-discharge on the order of a month - there are alternate chemistries but they have other problems right now.
  - ViewTrick1002 1y ago
    > seasonal variations Or overbuild renewables reducing the seasonal variations. In cost terms when compared to nuclear power those would be insignificant. With fossil based energy systems we didn’t match production capacity to consumption 100% with peakers having low capacity factors. But somehow we can’t overbuild a kWh and need massive seasonal storage when it comes to renewables.
    - idiotsecant 1y ago
      Yes, if you have a magic planet spanning transmission system capable of handling the power flows over building solves the problem. Unfortunately that's orders of magnitude more expensive than storage, which we already can't afford.
      - triceratops 1y ago
        Gonna have to see more numbers for "storage is more expensive than nuclear". And not the unit cost of SMRs with the assumption that mass manufacturing is solved, certified, and permitted. You have to account for those costs too. And time, of course. The climate crisis is here now. We can't wait 10 years for cheap SMRs to be ready (though we'll gladly take them when they are).
        johncolanduoni 1y ago
        I’ve never seen anybody give an estimate for the cost of storage required to fully convert the grid of e.g. the US that wasn’t obviously astronomical and not something the utilities could afford the capital for. If you’ve seen different please share.
        triceratops 1y ago
        I still don't see numbers. What is this "obviously astronomical" estimate? And how does it compare to nuclear, in any form?
        idiotsecant 1y ago
        Battery storage ranges from $150 to $300 / kwH capacity. The entire grid would need something like 5twH of capacity for an 8 hour ride through. I want you to carefully consider those prefixes and the vast, vast Gulf of space between them.
        ViewTrick1002 1y ago
triceratops 1y ago
> Storage is extremely expensive Define "expensive". Over what timescale? Have you seen https://ember-energy.org/latest-insights/solar-electricity-every-hour-of-every-day-is-here-and-it-changes-everything/ https://ember-energy.org/latest-insights/solar-electricity-e... "Achieving 97% of the way to 24/365 solar in very sunny regions is now affordable at as low as $104/MWh, cheaper than coal and nuclear and 22% less than a year earlier." This is right now, July 2025. The costs of batteries continue to fall. How much cheaper will batteries be by the time we start churning out SMRs fast and cheap? By all means keep beavering away at nuclear. Its time will come one day. But I won't hold my breath for it to solve the climate problem in the next 10 years.
- johncolanduoni 1y ago
  “Very sunny” is doing a lot of work there. The storage required goes up dramatically once you run the numbers for somewhere that has seasons. The long-range HVDC lines between hemispheres idea is cute but probably geopolitically impossible; I don’t think the US will let its ability to literally keep the lights on depend on South America. Storage could get there, but I don’t think it’s credible that manufacturing scale alone will solve the problem. We probably need some new, qualitatively different chemistries to become viable for solar to be viable for the whole grid. From a technical perspective the nuclear plants we could build in the 1960s could do it, whether we can still build them (no matter if the barrier is regulatory or practical) is another question.
  - ViewTrick1002 1y ago
    The other side of the question is: How will you get me with rooftop solar and a home battery to buy your extremely expensive nuclear powered electricity when I have my own imperfect solution almost the entire year? Scale this up to a society adding onshore and offshore wind and you quickly realize that the nuclear plant will have a capacity factor at 10% or so. Vogtle with a 20% capacity factor costs somewhere like 85 cents per kWh, or $850 per MWh. Nuclear power due to the massive CAPEX is the worse solution imaginable to fix renewable shortcomings. Take a look at France. They generally export quite large amounts of electricity. But whenever a cold spell hits that export flow is reversed to imports and they have to start up local fossil gas and coal based production. What they have done is that they have outsourced the management of their grid to their neighbors and rely on 35 GW of fossil based electricity production both inside France and their neighbors grids. Because their nuclear power produces too much when no one wants the electricity and too little when it is actually needed. Their neighbors are able to both absorb the cold spell which very likely hits them as well, their own grid as the French exports stops and they start exporting to France.
zekrioca 1y ago
> Storage is extremely expensive. No.
delusional 1y ago
> The cost of nuke energy is not because the technology is complex or because resources are scarce. It's because we have very, very burdensome regulations around nuclear reactors (for good reason!) So its easy, at least if it wasn't for all that burdensome regulation. But also the burdensome regulations is actually good, presumably because it's hard to get right. This sounds like nonsense to me. If the regulation is good, that would usually be because a thing is hard to make work in a liberal society, usually for some misaligned incentive reasons. In that case the regulation isn't "burdensome" but necessary to counteract the failure of the market.
- idiotsecant 1y ago
  You're approaching this with the nuance of a child. Yes, nuke regulation is burdensome, and yes it is necessary because nuke can have quite severe failures when failures occur. The solution is not to dogmatically suppose that one of those two basic facts is false. It's to engineer around the problem by making many exact copies of the same design, reducing the amount of regulation that needs to be applied on a per unit basis. That's what small modular reactors are. Certify once, build many.
  - delusional 1y ago
    > You're approaching this with the nuance of a child. I'm really happy I got to speak to such and adult then. Fuck you too.
magicalhippo 1y ago
> Solar is very, very cheap and almost totally worthless without storage. For say an AI training-oriented data center, you could scale down the power usage when supply is limited. You could change power limits on the CPU/GPUs, put the machines in sleep mode or powered off entirely. So the required storage would just be a slightly bigger UPS. Not sure if the economics works out, but at least technically it's possible as it's more flexible than user-based loads.
- anonymousDan 1y ago
  AI based training is an almost ideal match with this kind of supply. You could even imagine migrating long running training jobs to different parts of the world based on energy availability to optimise costs.
  - idiotsecant 1y ago
    So the model is buy some of the worlds most expensive hardware and let it sit idle for half the time? If I want to save the same throughput I need to buy at least twice the hardware! Load throttling is one of those ideas that seems great as long as someone else is doing it.
    - anonymousDan 1y ago
      Ha, that's a great point. I guess if you have some latency sensitive inference workload the capacity will effectively be dynamic, but that is likely uncorrelated with local energy prices I imagine.
- loeg 1y ago
  You don't want to waste your GPU capex by not running those suckers at 100%. (Other datacenter workloads it makes some sense to demand-regulate, but not AI.)
Kon5ole 1y ago
Solar benefits from storage yes but it's not at all worthless even without it. If your solar panels generate 10 TWh per year, you have 10 TWh unused hydro, gas, even oil and coal that is stored instead of spent. You have saved the planet from megatons of CO2 emissions even if you have no new green storage. Solar is already adding the equivalent of several nuclear power plants worth of new electricity every few months. Getting another month's worth of electricity delivered 10 years from now is not much of a bridge. I think solar and storage just needs every other worse idea to stay out of the way and things will be fine.

ericd 1y ago

Are you so sure that storage is so expensive? It’s been coming down the cost curve extremely quickly, such that opinions formed even an year ago are severely outdated, and it’s now solar+storage that’s being favorably compared to replacing nat gas plants, not just solar itself.

johncolanduoni 1y ago
Storage that is good enough to replace peaker plants, and storage that is good enough to handle seasonal variations in insolation are completely different ballgames. The lithium battery chemistry in your phone will self-discharge on the order of a month - there are alternate chemistries but they have other problems right now.
- ViewTrick1002 1y ago
  > seasonal variations Or overbuild renewables reducing the seasonal variations. In cost terms when compared to nuclear power those would be insignificant. With fossil based energy systems we didn’t match production capacity to consumption 100% with peakers having low capacity factors. But somehow we can’t overbuild a kWh and need massive seasonal storage when it comes to renewables.
  - idiotsecant 1y ago
    Yes, if you have a magic planet spanning transmission system capable of handling the power flows over building solves the problem. Unfortunately that's orders of magnitude more expensive than storage, which we already can't afford.
    - triceratops 1y ago
      Gonna have to see more numbers for "storage is more expensive than nuclear". And not the unit cost of SMRs with the assumption that mass manufacturing is solved, certified, and permitted. You have to account for those costs too. And time, of course. The climate crisis is here now. We can't wait 10 years for cheap SMRs to be ready (though we'll gladly take them when they are).
      - johncolanduoni 1y ago
        I’ve never seen anybody give an estimate for the cost of storage required to fully convert the grid of e.g. the US that wasn’t obviously astronomical and not something the utilities could afford the capital for. If you’ve seen different please share.
        triceratops 1y ago
        I still don't see numbers. What is this "obviously astronomical" estimate? And how does it compare to nuclear, in any form?
        idiotsecant 1y ago
        Battery storage ranges from $150 to $300 / kwH capacity. The entire grid would need something like 5twH of capacity for an 8 hour ride through. I want you to carefully consider those prefixes and the vast, vast Gulf of space between them.
        ViewTrick1002 1y ago

triceratops 1y ago

> Storage is extremely expensive Define "expensive". Over what timescale? Have you seen https://ember-energy.org/latest-insights/solar-electricity-every-hour-of-every-day-is-here-and-it-changes-everything/ https://ember-energy.org/latest-insights/solar-electricity-e... "Achieving 97% of the way to 24/365 solar in very sunny regions is now affordable at as low as $104/MWh, cheaper than coal and nuclear and 22% less than a year earlier." This is right now, July 2025. The costs of batteries continue to fall. How much cheaper will batteries be by the time we start churning out SMRs fast and cheap? By all means keep beavering away at nuclear. Its time will come one day. But I won't hold my breath for it to solve the climate problem in the next 10 years.

johncolanduoni 1y ago
“Very sunny” is doing a lot of work there. The storage required goes up dramatically once you run the numbers for somewhere that has seasons. The long-range HVDC lines between hemispheres idea is cute but probably geopolitically impossible; I don’t think the US will let its ability to literally keep the lights on depend on South America. Storage could get there, but I don’t think it’s credible that manufacturing scale alone will solve the problem. We probably need some new, qualitatively different chemistries to become viable for solar to be viable for the whole grid. From a technical perspective the nuclear plants we could build in the 1960s could do it, whether we can still build them (no matter if the barrier is regulatory or practical) is another question.
- ViewTrick1002 1y ago
  The other side of the question is: How will you get me with rooftop solar and a home battery to buy your extremely expensive nuclear powered electricity when I have my own imperfect solution almost the entire year? Scale this up to a society adding onshore and offshore wind and you quickly realize that the nuclear plant will have a capacity factor at 10% or so. Vogtle with a 20% capacity factor costs somewhere like 85 cents per kWh, or $850 per MWh. Nuclear power due to the massive CAPEX is the worse solution imaginable to fix renewable shortcomings. Take a look at France. They generally export quite large amounts of electricity. But whenever a cold spell hits that export flow is reversed to imports and they have to start up local fossil gas and coal based production. What they have done is that they have outsourced the management of their grid to their neighbors and rely on 35 GW of fossil based electricity production both inside France and their neighbors grids. Because their nuclear power produces too much when no one wants the electricity and too little when it is actually needed. Their neighbors are able to both absorb the cold spell which very likely hits them as well, their own grid as the French exports stops and they start exporting to France.

zekrioca 1y ago

> Storage is extremely expensive. No.

delusional 1y ago

> The cost of nuke energy is not because the technology is complex or because resources are scarce. It's because we have very, very burdensome regulations around nuclear reactors (for good reason!) So its easy, at least if it wasn't for all that burdensome regulation. But also the burdensome regulations is actually good, presumably because it's hard to get right. This sounds like nonsense to me. If the regulation is good, that would usually be because a thing is hard to make work in a liberal society, usually for some misaligned incentive reasons. In that case the regulation isn't "burdensome" but necessary to counteract the failure of the market.

idiotsecant 1y ago
You're approaching this with the nuance of a child. Yes, nuke regulation is burdensome, and yes it is necessary because nuke can have quite severe failures when failures occur. The solution is not to dogmatically suppose that one of those two basic facts is false. It's to engineer around the problem by making many exact copies of the same design, reducing the amount of regulation that needs to be applied on a per unit basis. That's what small modular reactors are. Certify once, build many.
- delusional 1y ago
  > You're approaching this with the nuance of a child. I'm really happy I got to speak to such and adult then. Fuck you too.

magicalhippo 1y ago

> Solar is very, very cheap and almost totally worthless without storage. For say an AI training-oriented data center, you could scale down the power usage when supply is limited. You could change power limits on the CPU/GPUs, put the machines in sleep mode or powered off entirely. So the required storage would just be a slightly bigger UPS. Not sure if the economics works out, but at least technically it's possible as it's more flexible than user-based loads.

anonymousDan 1y ago
AI based training is an almost ideal match with this kind of supply. You could even imagine migrating long running training jobs to different parts of the world based on energy availability to optimise costs.
- idiotsecant 1y ago
  So the model is buy some of the worlds most expensive hardware and let it sit idle for half the time? If I want to save the same throughput I need to buy at least twice the hardware! Load throttling is one of those ideas that seems great as long as someone else is doing it.
  - anonymousDan 1y ago
    Ha, that's a great point. I guess if you have some latency sensitive inference workload the capacity will effectively be dynamic, but that is likely uncorrelated with local energy prices I imagine.
loeg 1y ago
You don't want to waste your GPU capex by not running those suckers at 100%. (Other datacenter workloads it makes some sense to demand-regulate, but not AI.)

Kon5ole 1y ago

Solar benefits from storage yes but it's not at all worthless even without it. If your solar panels generate 10 TWh per year, you have 10 TWh unused hydro, gas, even oil and coal that is stored instead of spent. You have saved the planet from megatons of CO2 emissions even if you have no new green storage. Solar is already adding the equivalent of several nuclear power plants worth of new electricity every few months. Getting another month's worth of electricity delivered 10 years from now is not much of a bridge. I think solar and storage just needs every other worse idea to stay out of the way and things will be fine.