The country-level energy consumption vs per capita numbers look off? China almost certainly does not have more per capita energy usage than the US.
And as for the impact of growing renewables on grid (and off grid) uses, I wrote something almost three years ago now but I think is still very relevant
Yes, I noticed that at least some countries had per capita numbers very closely proportional to their total numbers, despite having very different populations. (The US compared to India and China was one, but I think Canada compared to Russia and South Korea might have been another.)
1. You said: "Footnote: Solar and wind do only convert a portion of input energy into electrical energy, but because this energy would be expended regardless (the sun shines and the wind blows whether solar cells and wind turbines are there or not), their low conversion efficiencies don’t have the same meaning, even aside from questions of carbon emissions."
The easy way to characterize the difference you reference here is that a fuel system converts "natural energy" in a chemical form into "primary energy" in the form of heat. The conversion to electricity and then some final form at somebody's house or whatever has an efficiency. A PV system converts "natural energy" in the form of light into "primary energy" in the form of electricity. So there are, by definition, no efficiency losses at the point of electricity generation for a PV system. It is important to recognize that we need a lot less primary energy for PV and other non-fuel systems than for our conventional fuel-based fleet.
2. You said: " A generator connected to the turbine turns this kinetic energy into electrical energy, which takes the form of electrons moving back and forth in an alternating current. This electrical energy then moves through the transmission and distribution system, its voltage and current being modulated along the way by transformers to minimize distribution losses, until it eventually reaches someone’s home."
Electricity is weirder than this. In most generating equipment, an electron goes from one side of the circuit out to a load (which would be in somebody's home or what-have-you) and transforms its energy there, but then the electron proceeds all the way back to the generating equipment to the other side of the circuit. Now in a grid, each electron won't necessarily go back to the same piece of generating equipment from where it started, so the picture is more complex; but the single electron's journey could indeed be thousands of miles. Electricity doesn't flow *to* your house, it flows *through* your house. This is strange because electrons are changing their form of energy in the *middle* of their motion from one place to another, rather than at the point of interaction, as we see with mechanical energy. Also, AC circuits don't have electrons going "back and forth." Actually, the reservoir of electrons moves back and forth between different sides of the circuit. With a rotating generator, this is done through the rotation of a magnet around fixed conductors.
Electron's don't travel from a power station to someone's home. They're not the unit of energy transfer. The speed of effect is relativistic, but the average velocity of electrons per unit of wire is like inches per minute or something like that.
Yes, it's weird.
There was a whole nerd youtube kerfuffle about this a few years ago.
Electrons actually move at drift velocity which is very slow, of the tune of 0.5 inches per minute. But the overall electric field (signal) moves at nearly the speed of light.
Thanks for doing the extensive intellectual work needed to produce this, perhaps supported by many megajoules of chemical intake.
Some notes:
1. Those growth rate comparisons promote poor intuition. Even if solar continued to grow at a 25% rate of compounding, starting at a 1% fraction implies that it takes more than 20 years to get to parity.
2. Ammonia production, mainly for fertilizer essential to feeding humanity, consumes 6.5% of natural gas production (in 2020, per EIA). This presumably is inside industrial on the Sankey chart.
3. Beyond that, hydrocarbons play an essential role in the production of steel, cement, and plastics, the other three pillars of civilization as discussed in How The World Really Works by Vaclav Smil.
Fertilizer is the reason the no-more-oil mantra is so absurd. Hydrocarbons are both the feedstock and energy source that grow our food. Paul Kingsnorth's calls it "fossilized sunlight", and without it (using only the energy of current sunlight) the Earth caps out at 4-5B people.
that's only because most of the reactive nitrogen feeds animals to be eaten by humans take that huge inefficiency out and feeding 8 billion is much more doable and don't forget about half food produced is wasted in wealthy countries
A few simple energy calculations (average hours of daylight, % of the Earth that's cultivatable, chlorophyll efficiency, etc...) can demonstrate the carrying capacity limits of the Earth. So can history: when I was growing up famine was a common thing; today it isn't thanks to Norman Borlaug and the green revolution.
Even a 100% vegetarian diet requires manure as fertilizer. (Paul Kingsnorth composts his own poop.) The most efficient way to do this is using grassy rangeland (that's not good enough for row crops) as forage for animals whose dung is then gathered as fertilizer for better farmland.
The most important current constraint is that there is no "cost effective storage" that is available in grid scale quantities. Lots of articles saying "batteries are getting cheaper by the day", which is wonderful but completely misses the point. Write down the daily US or EU electricity consumption. Write down the output of the world’s largest battery factory. Compare the two. When someone says "batteries are getting cheaper by the day" and doesn’t compare these two values, it’s because they have massively missed the point.
The US consumes about 11,000 GWh of electricity per day. According to a random website CATL is the largest lithium ion battery manufacturer and produced almost 100 GWh of lithium ion batteries last year. So after 100 years of production they could produce enough batteries for one day of storage for the US grid. And after 5-10 years you would need to replace these batteries.
One day of storage would be a help, many days would be better, but hopefully, given that we have engineering minded people here, you all get the main point.
Four hours of storage is enough to load-shift from peak solar output to peak demand during any given day. Longer-term energy storage makes more sense to handle with other tech, cheaper per unit capacity at the cost of responding more slowly.
> produced almost 100 GWh of lithium ion batteries last year. So after 100 years
Are you familiar with the concept of things scaling up over time? I'm pretty sure CATL wasn't producing a hundred gigwatt-hours of batteries per year ten, or even five, years ago. Following that same curve, they also won't be capped at their current capacity for long, it'll keep increasing - or be surpassed by some rival.
I think what a lot of people don’t quite understand, including Trump’s pick to be Treasury Secretary, is that the Age of Oil is effectively over. We are now in the Age of Natural Gas and the world has more than enough cheap natural gas for the foreseeable future.
So counterintuitively the time to develop renewable energy sources is now when natural gas is cheap and abundant. So we can’t have another energy crisis like we had in 2001-2008 in which CPI was elevated for 4 straight years peaking at 5.5% in July 2008 when oil and North American natural gas hit near record high prices at the same time. Those 4 years of elevated CPI degraded lower income disposable income AND forced jobs to China (that had cheap energy) and culminated in the Global Financial Crisis.
The Gulf nation of Qatar has been awash in natural gas. They built the $15B Pearl GTL (gas to liquid) plant. For 12+ years, it's been converting gas into 250,000 barrels per DAY of various petroleum products.
Expect to see more such animals built where the deranged protests of the enviro anti-industry-industry can be ignored.
Natural gas can be converted into diesel and kerosene at current prices…which is why the price of oil can’t go higher for significant period of time. When the price of oil was around $100/barrel under Obama Lake Charles was going to be the region in which billions of dollars in infrastructure was going to be built to convert natural gas into diesel and kerosene…and then OPEC+ flooded the market and the projects were cancelled. But they couldn’t stop the billions in LNG infrastructure being built in Louisiana. Oh, and Speaker Johnson changed his tune very quickly on aid to Ukraine because the natural gas being exported to Louisiana comes from his district and is piped to Lake Charles to be converted to LNG.
It wasn’t just cheap Chinese labor that forced manufacturing production overseas, if we used that logic then cheap renewables would be our primary sources of energy and the more expensive, dirty and unhealthy burning of fossil fuels would be only used for back up.
And specifically in the case of plastics manufacturing, US producers sold plastic raw materials to China so much cheaper than they did to their US customers that it created the era of finished Chinese products being cheaper in the US than American products before US labor costs. ( PE bags in Texas)
Now its happening with cars, EV designs can be inherently cheaper to be manufactured . China isn’t doing it with just cheaper labor but with more effective and efficient design for manufacturing.
I actually feel sorry for Lamborghini owners because a $10k Chinese EV will perform better than their Lambo in 2030. Btw, the New Orleans terrorist attack was possible because the truck was an EV that performed like a Lambo and it was silent. So the first block when people heard tires screeching and the cop yelling had no fatalities because they got out of the way.
I would just caution readers (and authors) to be aware that there can be confusing ambiguities when we try to add up and compare amounts of energy that are in different forms.
For example, EIA and LLNL recently changed their accounting method for non-combustion renewable electricity (wind, solar, hydro, geothermal), and the flow chart in this article uses the new accounting (though the fine print at the bottom hasn't been updated to reflect this!). The chart shows, for example, that the US consumes more than five times as much nuclear energy as wind energy. This is correct if we measure the nuclear energy by the heat it produces, while measuring the wind energy by the electricity it produces. But we wouldn't need to do it that way. If we measured the nuclear energy by the electricity it produces (to be consistent with wind), the chart would show only about a third as much nuclear energy—so it would be less than double the amount of wind energy. On the other hand, we could measure nuclear energy by the amount of energy that can in principle be extracted from the fuel, and that would be larger than the heat produced, because not all the fuel is consumed in the reactor. We could measure wind energy as the kinetic energy in the wind that passes through the disk swept out by the turbine blades, and that would be more than double the electricity produced. There's no single right way to do it, and different accounting methods give very different results.
A similar ambiguity affects comparisons between energy delivered to consumers as electricity and energy delivered as a chemical fuel. When you say a household uses more energy as gasoline than in the home, are you measuring electrical energy as it enters the home, or are you measuring the primary energy used to produce it?
A related issue is how to interpret claims about the specific percentage of all energy that gets "wasted" or used to produce "useful work". It's merely a convention that the heat going up my chimney is "wasted" while the heat escaping through my uninsulated foundation is "useful". The heat going out my car's tailpipe is said to be "wasted", while the heat created by air drag as my car speeds down the highway is said to be "useful". The technical meanings of these words in energy accounting can be surprisingly different from their meanings in ordinary English.
The actual usefulness of most energy, in the everyday sense of the word, can't even be measured in energy units. Instead it's miles traveled or kilograms of steel produced or bits of information processed. Or as Amory Lovins famously said, hot showers and cold beer.
To the point about oil being surprisingly high energy density, it's worth recognizing that there is a limited sense in which this was at least partly the result of biological evolution. Life is benefited greatly by having a high energy density medium for storing energy that is also low mass density (at least relative to water). Yes, most of what was buried was carbohydrate and not hydrocarbons, and the hydrocarbons are much higher energy density than carbohydrates. But the hydrocarbons wouldn't be there at all if they carbohydrates had never been buried and then undergone the chemical reactions that produced the hydrocarbons we're now consuming. Meanwhile, biological evolution continued and animals with surplus calories make good use of fluffy, slippery, and floatable hydrocarbons for storing energy for later use!
> One problem with using joules is that a joule is a tiny amount of energy, and using it to describe quantities of energy used in everyday life results in huge figures: burning a gallon of gas releases about 121 million joules. For intuition building, it's useful to use a unit that doesn’t have so many trailing digits. One common unit is the kilowatt-hour (kWh)
I don't think that the usage of kWh over Joules has anything to do with the number of trailing digits. Metric prefixes render this a nonissue; no one would spell out "millions of Joules", they would just say Megajoules. Indeed, your example of Kilowatt-hours is doing exactly this, using a prefix to avoid saying 'thousands of watt-hours".
I once assembled a bunch of values that all represented one kilowatt-hour and some of my favorites were the food energy in a stick of butter, the electricity from 666 AA cells, lifting an average human 5000 meters, and a 1000 kg car traveling at 190 miles per hour.
Two comments from a researcher working on the topic. 1) mAh is only a measure of charge, you still need to multiply it by the battery voltage to get Wh. 2)always good to remember that no matter how efficient and how clean the energy we use is (or will be) there are hard coded thermodynamic limits to our growth, lest we want to boil the planet in a few 100s years (https://dothemath.ucsd.edu/2012/04/economist-meets-physicist/) Cheers!
I'm also a researcher in the topic and that guy's textbook didn't sit right with me. I have a physics background so I appreciated the first principles math, but I also thought some of his conclusions had a Paul Ehrlich style Malthusianism that doesn't necessarily map to the data. The true insight that there is an upper bound on the waste heat the planet can actually propagate away is trivial to the energy transition. The global demographic transition and the absolute uncoupling in developing economies seem to me to at least complicate the relationship between waste heat and total utility. If our descendants come up against the finite ceiling of the planet that will have meant we succeeded in avoiding catastrophe
There is no absolute uncoupling. Rich countries use more energy per capita, to do more work per capita. The biggest new source of electricity demand is data centers - they're building entire power stations, one per data center in some cases.
I'll believe in decoupling when rich people all take their holidays by bicycle, not plane.
To quote our conclusion: "Furthermore, when looking beyond 2028, the current surge in data center electricity demand should be put in the context of the much larger electricity demand expected over the next few decades from a combination of electric vehicle adoption, onshoring of manufacturing, hydrogen utilization, and the electrification of industry and buildings."
It is a source of demand growth, it is worth keeping an eye on, but according to our model is it not the largest new source of electricity demand.
What is the expectation for electric vehicle adoption, hydrogen utilization, and electrification of industry and buildings? I know that many people write as if these trends are already established, but I've seen no clear argument that they are technically or economically feasible, nor that anyone is even considering building the industrial base to support such moves. For instance, my back of the envelope calculations would require a tenfold increase in global lithium production to allow the US to replace all new car sales with electric vehicles, leaving aside EVs in Europe, Japan, China, or other parts of the world. Did your research touch on this aspect?
While I do other work (with some of the same people) on critical materials supply, those other topics were out of scope for this report, and the specific demand projections for those sectors isn't really my topic area. So caveat this with I haven't actually done an extensive bottom-up technoeconomic model like I have in data centers.
I'll cite the International Energy Agency's annual energy outlook (https://www.iea.org/reports/world-energy-outlook-2024). Their data center specific energy estimates are consistent with our results, so I'm reasonably comfortable taking their line on this. To quote that report:
"The main drivers of [electricity demand] increase [by 2030] are the electrification of transport, industrial processes and space heating, and rising demand for appliances and cooling...In the base case, data centres account for less than 10% of total electricity demand growth at the global level, which is roughly on a par with demand growth for desalination, and less than a third of the demand growth for both EVs and space cooling in the buildings sector. Therefore, growth of electricity demand for data centres is projected to be rapid, but the level looks set to remain relatively small in the context of overall global demand growth"
It's worth noting that critical mineral supply can be inherently circular. A fossil based energy system depletes, while we can economically harvest large fractions (depending on chemistry and recycling technique up to 95%) of cathode active material in batteries at EoL (https://www.sciencedirect.com/science/article/abs/pii/S0921344923001763). There's a short run question about bringing enough processing capacity online but that's different from "we don't have enough lithium" (https://wires.onlinelibrary.wiley.com/doi/abs/10.1002/wcc.768, p9)
Building electrification is already the most life-cycle cost effective way to construct new buildings (https://rmi.org/economics-of-electrifying-buildings/), so that's not a vaporware trend. Pem hydrogen is an open question, but that's on the demand side not constructing the electrolyzer side.
I don't know if that answered your question, but this is some context
Long before hard limits of waste heat hit planet-boiling levels, we'll have enough transit capacity to export energy-intensive industries to space, at which point radiator area is no longer constrained by the planet's surface geometry, so that thermodynamic model's assumptions break down.
You sound pretty confident about this fantastical scenario. I suspect the more likely reason humans will avoid the hard waste heat limit is that organized industrial society will disintegrate as population collapses along with potential exacerbations from large scale pandemics and nuclear war.
You sound pretty confident in dismissing businesses which already exist as "fantastical." GPS is old news. Starlink is steadily expanding, and profitable enough to prop up other strategic interests while competing with ground-based service providers. Cost per kilogram to LEO goes down, real estate prices go up, soon enough it's cheaper and easier to build the next marginal datacenter out there where solar power isn't intermittent, NEPA's irrelevant, and heat sinks can hit single-digit kelvin without even needing any moving parts.
Then somebody gets annoyed with debris avoidance and Coronal-Mass-Ejection-related downtime, realizes they can solve two stones with one bird by having maintenance bots harvest asteroids to add more spall jackets and radiation shielding... and as long as you're mining anyway, might as well check if any of that space dirt has rare-earth minerals which let you brag about avoiding planetside child labor... then smelters (with huge inflatable mirrors for cheap, super-high-grade process heat) move closer to the ore, machinists follow the smelters (first to save transit costs, which would likely still be on the order of $10 or $100 / kg, then exploring microgravity-exclusive possibilities such as foamed composites or convectionless chemistry), and so on through the tech stack. https://projectrho.com/public_html/rocket/infrastructure.php
> organized industrial society will disintegrate as population collapses
Just how bad are you expecting things to get? Usual estimate I hear is that ten thousand people would be enough to maintain both enough genetic diversity to avoid inbreeding, and all the essential skillsets of an industrial economy. Could squeeze through a population bottleneck of a few hundred with careful planning https://wiki.opensourceecology.org/wiki/Global_Village_Construction_Set and a lot less personal freedom or safety margin against unexpected developments. Ten million or so for a more well-rounded modern tech base, jets and advanced electronics and suchlike - meaning that's the real minimum for a self-sufficient colony on planets without natively breathable atmosphere or edible wild plants.
> along with potential exacerbations from large scale pandemics
Smallpox is extinct in the wild. Polio, guinea worm, and Yersinia pestis, nearly so. Vaccines are already rolling out against malaria and tooth decay... well, that latter is technically more of an engineered symbiote. Techniques developed for Covid could generalize to the common cold. Many species have inconvenienced humanity, few with lasting success. The lucky ones get domesticated.
> and nuclear war
Against who? The Russians? Putin would sooner peel his own face off, assume a new identity, and live out his twilight years in (luxurious) exile, than push the big red button and go down in history as the man who sealed their doom. He didn't get to be the last one standing in all those internal KGB power struggles by being brave and principled.
Chinese "wolf warrior diplomacy" is rooted in getting symbolic revenge for the Century of Humiliation, proving that their core virtues were right all along, that everybody who bullied them was inherently flawed, and that now they know the tricks they can be *even better* at bullying. That's why they over-commit to infrastructure, domestically clamp down on anything even vaguely isomorphic to paying silver for opium, and gleefully inflict such vices on foreigners. Nuclear annihilation isn't a plausible route to the moral high ground, so they'll avoid it, just like they carefully avoid unambiguous acts of war while hunting for Air Bud exceptions in the international definition of "territorial waters."
Trump is too senile to issue a properly-formatted launch order, and too arrogant to delegate it. Even if he did manage to follow official procedures, everyone else in the relevant chain of command is too sane to play along without a good reason, and he won't be able to provide one of those, because he gives more weight to G.I. Joe cartoons and Nazi propaganda than nonfiction sources when it comes to figuring out how a military social context is supposed to work.
No other credible US candidate is in favor of first-use, and no other country has enough warheads and delivery systems for much besides thwarting an otherwise-overwhelming conventional invasion of their own territory - one of many reasons military conquest is rarely even attempted in the developed world.
As is making polysilicon for solar much less building a solar installation and the fact that solar panels can theoretically last 30 years, the practical reality is much less <20 years…solar wouldn’t be possible without mountains of cheap Chinese coal… grow up everyone
I can’t speak to all manufacturers, my SunPower panels have been demonstrated to have a 40 year life. They don’t actually die after that time, they slowly loose efficiency over time.
Nitpick: a turbine does not put (much) energy into the kinetic energy of the machinery's motion. That would involve rotating absurdly fast! Rather, it uses that motion, against the resistance of the load, as a medium to transfer the energy to another form.
The best thing we can do for carbon emissions is displace coal with natural gas and eventually nuclear. Replacing coal with natural gas would take us almost halfway to netzero. Which is lot closer than we will ever get by covering the planet with windmills and panels. And natural gas extraction is much less harmful than coal extraction.
The country-level energy consumption vs per capita numbers look off? China almost certainly does not have more per capita energy usage than the US.
And as for the impact of growing renewables on grid (and off grid) uses, I wrote something almost three years ago now but I think is still very relevant
https://www.tsungxu.com/p/clean-energy-transition-guide
Yep, this is an error, I’m away from my computer atm but will correct it as soon as I’m able.
Thanks!
This has now been fixed.
Yes, I noticed that at least some countries had per capita numbers very closely proportional to their total numbers, despite having very different populations. (The US compared to India and China was one, but I think Canada compared to Russia and South Korea might have been another.)
Nice summary article. Two notes:
1. You said: "Footnote: Solar and wind do only convert a portion of input energy into electrical energy, but because this energy would be expended regardless (the sun shines and the wind blows whether solar cells and wind turbines are there or not), their low conversion efficiencies don’t have the same meaning, even aside from questions of carbon emissions."
The easy way to characterize the difference you reference here is that a fuel system converts "natural energy" in a chemical form into "primary energy" in the form of heat. The conversion to electricity and then some final form at somebody's house or whatever has an efficiency. A PV system converts "natural energy" in the form of light into "primary energy" in the form of electricity. So there are, by definition, no efficiency losses at the point of electricity generation for a PV system. It is important to recognize that we need a lot less primary energy for PV and other non-fuel systems than for our conventional fuel-based fleet.
2. You said: " A generator connected to the turbine turns this kinetic energy into electrical energy, which takes the form of electrons moving back and forth in an alternating current. This electrical energy then moves through the transmission and distribution system, its voltage and current being modulated along the way by transformers to minimize distribution losses, until it eventually reaches someone’s home."
Electricity is weirder than this. In most generating equipment, an electron goes from one side of the circuit out to a load (which would be in somebody's home or what-have-you) and transforms its energy there, but then the electron proceeds all the way back to the generating equipment to the other side of the circuit. Now in a grid, each electron won't necessarily go back to the same piece of generating equipment from where it started, so the picture is more complex; but the single electron's journey could indeed be thousands of miles. Electricity doesn't flow *to* your house, it flows *through* your house. This is strange because electrons are changing their form of energy in the *middle* of their motion from one place to another, rather than at the point of interaction, as we see with mechanical energy. Also, AC circuits don't have electrons going "back and forth." Actually, the reservoir of electrons moves back and forth between different sides of the circuit. With a rotating generator, this is done through the rotation of a magnet around fixed conductors.
Electron's don't travel from a power station to someone's home. They're not the unit of energy transfer. The speed of effect is relativistic, but the average velocity of electrons per unit of wire is like inches per minute or something like that.
Yes, it's weird.
There was a whole nerd youtube kerfuffle about this a few years ago.
https://www.youtube.com/watch?v=bHIhgxav9LY
https://www.youtube.com/watch?v=2Vrhk5OjBP8
Are a couple of videos in that kerfuffle and there's more.
Electrons actually move at drift velocity which is very slow, of the tune of 0.5 inches per minute. But the overall electric field (signal) moves at nearly the speed of light.
Thanks for doing the extensive intellectual work needed to produce this, perhaps supported by many megajoules of chemical intake.
Some notes:
1. Those growth rate comparisons promote poor intuition. Even if solar continued to grow at a 25% rate of compounding, starting at a 1% fraction implies that it takes more than 20 years to get to parity.
2. Ammonia production, mainly for fertilizer essential to feeding humanity, consumes 6.5% of natural gas production (in 2020, per EIA). This presumably is inside industrial on the Sankey chart.
3. Beyond that, hydrocarbons play an essential role in the production of steel, cement, and plastics, the other three pillars of civilization as discussed in How The World Really Works by Vaclav Smil.
Fertilizer is the reason the no-more-oil mantra is so absurd. Hydrocarbons are both the feedstock and energy source that grow our food. Paul Kingsnorth's calls it "fossilized sunlight", and without it (using only the energy of current sunlight) the Earth caps out at 4-5B people.
that's only because most of the reactive nitrogen feeds animals to be eaten by humans take that huge inefficiency out and feeding 8 billion is much more doable and don't forget about half food produced is wasted in wealthy countries
I love a good "If everyone would just ..." argument
A few simple energy calculations (average hours of daylight, % of the Earth that's cultivatable, chlorophyll efficiency, etc...) can demonstrate the carrying capacity limits of the Earth. So can history: when I was growing up famine was a common thing; today it isn't thanks to Norman Borlaug and the green revolution.
Even a 100% vegetarian diet requires manure as fertilizer. (Paul Kingsnorth composts his own poop.) The most efficient way to do this is using grassy rangeland (that's not good enough for row crops) as forage for animals whose dung is then gathered as fertilizer for better farmland.
The most important current constraint is that there is no "cost effective storage" that is available in grid scale quantities. Lots of articles saying "batteries are getting cheaper by the day", which is wonderful but completely misses the point. Write down the daily US or EU electricity consumption. Write down the output of the world’s largest battery factory. Compare the two. When someone says "batteries are getting cheaper by the day" and doesn’t compare these two values, it’s because they have massively missed the point.
The US consumes about 11,000 GWh of electricity per day. According to a random website CATL is the largest lithium ion battery manufacturer and produced almost 100 GWh of lithium ion batteries last year. So after 100 years of production they could produce enough batteries for one day of storage for the US grid. And after 5-10 years you would need to replace these batteries.
One day of storage would be a help, many days would be better, but hopefully, given that we have engineering minded people here, you all get the main point.
Four hours of storage is enough to load-shift from peak solar output to peak demand during any given day. Longer-term energy storage makes more sense to handle with other tech, cheaper per unit capacity at the cost of responding more slowly.
> produced almost 100 GWh of lithium ion batteries last year. So after 100 years
Are you familiar with the concept of things scaling up over time? I'm pretty sure CATL wasn't producing a hundred gigwatt-hours of batteries per year ten, or even five, years ago. Following that same curve, they also won't be capped at their current capacity for long, it'll keep increasing - or be surpassed by some rival.
You might be interested in the following article if you haven’t seen it already: https://www.city-journal.org/article/the-energy-transition-wont-happen
In the definition of the joule, probably you mean acceleration of one meter per second squared; you omitted the “squared”
Thanks, this has been fixed.
Great review. But really sad you didnt cover nuclear energy. Its the most efficient and energy dense power source we have discovered
I think what a lot of people don’t quite understand, including Trump’s pick to be Treasury Secretary, is that the Age of Oil is effectively over. We are now in the Age of Natural Gas and the world has more than enough cheap natural gas for the foreseeable future.
So counterintuitively the time to develop renewable energy sources is now when natural gas is cheap and abundant. So we can’t have another energy crisis like we had in 2001-2008 in which CPI was elevated for 4 straight years peaking at 5.5% in July 2008 when oil and North American natural gas hit near record high prices at the same time. Those 4 years of elevated CPI degraded lower income disposable income AND forced jobs to China (that had cheap energy) and culminated in the Global Financial Crisis.
The Gulf nation of Qatar has been awash in natural gas. They built the $15B Pearl GTL (gas to liquid) plant. For 12+ years, it's been converting gas into 250,000 barrels per DAY of various petroleum products.
Expect to see more such animals built where the deranged protests of the enviro anti-industry-industry can be ignored.
Natural gas can be converted into diesel and kerosene at current prices…which is why the price of oil can’t go higher for significant period of time. When the price of oil was around $100/barrel under Obama Lake Charles was going to be the region in which billions of dollars in infrastructure was going to be built to convert natural gas into diesel and kerosene…and then OPEC+ flooded the market and the projects were cancelled. But they couldn’t stop the billions in LNG infrastructure being built in Louisiana. Oh, and Speaker Johnson changed his tune very quickly on aid to Ukraine because the natural gas being exported to Louisiana comes from his district and is piped to Lake Charles to be converted to LNG.
It wasn’t just cheap Chinese labor that forced manufacturing production overseas, if we used that logic then cheap renewables would be our primary sources of energy and the more expensive, dirty and unhealthy burning of fossil fuels would be only used for back up.
And specifically in the case of plastics manufacturing, US producers sold plastic raw materials to China so much cheaper than they did to their US customers that it created the era of finished Chinese products being cheaper in the US than American products before US labor costs. ( PE bags in Texas)
Now its happening with cars, EV designs can be inherently cheaper to be manufactured . China isn’t doing it with just cheaper labor but with more effective and efficient design for manufacturing.
I actually feel sorry for Lamborghini owners because a $10k Chinese EV will perform better than their Lambo in 2030. Btw, the New Orleans terrorist attack was possible because the truck was an EV that performed like a Lambo and it was silent. So the first block when people heard tires screeching and the cop yelling had no fatalities because they got out of the way.
Lots of good insight here—thanks!
I would just caution readers (and authors) to be aware that there can be confusing ambiguities when we try to add up and compare amounts of energy that are in different forms.
For example, EIA and LLNL recently changed their accounting method for non-combustion renewable electricity (wind, solar, hydro, geothermal), and the flow chart in this article uses the new accounting (though the fine print at the bottom hasn't been updated to reflect this!). The chart shows, for example, that the US consumes more than five times as much nuclear energy as wind energy. This is correct if we measure the nuclear energy by the heat it produces, while measuring the wind energy by the electricity it produces. But we wouldn't need to do it that way. If we measured the nuclear energy by the electricity it produces (to be consistent with wind), the chart would show only about a third as much nuclear energy—so it would be less than double the amount of wind energy. On the other hand, we could measure nuclear energy by the amount of energy that can in principle be extracted from the fuel, and that would be larger than the heat produced, because not all the fuel is consumed in the reactor. We could measure wind energy as the kinetic energy in the wind that passes through the disk swept out by the turbine blades, and that would be more than double the electricity produced. There's no single right way to do it, and different accounting methods give very different results.
A similar ambiguity affects comparisons between energy delivered to consumers as electricity and energy delivered as a chemical fuel. When you say a household uses more energy as gasoline than in the home, are you measuring electrical energy as it enters the home, or are you measuring the primary energy used to produce it?
A related issue is how to interpret claims about the specific percentage of all energy that gets "wasted" or used to produce "useful work". It's merely a convention that the heat going up my chimney is "wasted" while the heat escaping through my uninsulated foundation is "useful". The heat going out my car's tailpipe is said to be "wasted", while the heat created by air drag as my car speeds down the highway is said to be "useful". The technical meanings of these words in energy accounting can be surprisingly different from their meanings in ordinary English.
The actual usefulness of most energy, in the everyday sense of the word, can't even be measured in energy units. Instead it's miles traveled or kilograms of steel produced or bits of information processed. Or as Amory Lovins famously said, hot showers and cold beer.
To the point about oil being surprisingly high energy density, it's worth recognizing that there is a limited sense in which this was at least partly the result of biological evolution. Life is benefited greatly by having a high energy density medium for storing energy that is also low mass density (at least relative to water). Yes, most of what was buried was carbohydrate and not hydrocarbons, and the hydrocarbons are much higher energy density than carbohydrates. But the hydrocarbons wouldn't be there at all if they carbohydrates had never been buried and then undergone the chemical reactions that produced the hydrocarbons we're now consuming. Meanwhile, biological evolution continued and animals with surplus calories make good use of fluffy, slippery, and floatable hydrocarbons for storing energy for later use!
> One problem with using joules is that a joule is a tiny amount of energy, and using it to describe quantities of energy used in everyday life results in huge figures: burning a gallon of gas releases about 121 million joules. For intuition building, it's useful to use a unit that doesn’t have so many trailing digits. One common unit is the kilowatt-hour (kWh)
I don't think that the usage of kWh over Joules has anything to do with the number of trailing digits. Metric prefixes render this a nonissue; no one would spell out "millions of Joules", they would just say Megajoules. Indeed, your example of Kilowatt-hours is doing exactly this, using a prefix to avoid saying 'thousands of watt-hours".
I once assembled a bunch of values that all represented one kilowatt-hour and some of my favorites were the food energy in a stick of butter, the electricity from 666 AA cells, lifting an average human 5000 meters, and a 1000 kg car traveling at 190 miles per hour.
Two comments from a researcher working on the topic. 1) mAh is only a measure of charge, you still need to multiply it by the battery voltage to get Wh. 2)always good to remember that no matter how efficient and how clean the energy we use is (or will be) there are hard coded thermodynamic limits to our growth, lest we want to boil the planet in a few 100s years (https://dothemath.ucsd.edu/2012/04/economist-meets-physicist/) Cheers!
Icymi this is the best write up I’ve come across on this issue…
https://aeon.co/essays/theres-a-deeper-problem-hiding-beneath-global-warming
I'm also a researcher in the topic and that guy's textbook didn't sit right with me. I have a physics background so I appreciated the first principles math, but I also thought some of his conclusions had a Paul Ehrlich style Malthusianism that doesn't necessarily map to the data. The true insight that there is an upper bound on the waste heat the planet can actually propagate away is trivial to the energy transition. The global demographic transition and the absolute uncoupling in developing economies seem to me to at least complicate the relationship between waste heat and total utility. If our descendants come up against the finite ceiling of the planet that will have meant we succeeded in avoiding catastrophe
There is no absolute uncoupling. Rich countries use more energy per capita, to do more work per capita. The biggest new source of electricity demand is data centers - they're building entire power stations, one per data center in some cases.
I'll believe in decoupling when rich people all take their holidays by bicycle, not plane.
Regarding data centers, I'm a coauthor on the recently released report to congress on data center energy use (https://energyanalysis.lbl.gov/publications/2024-lbnl-data-center-energy-usage-report).
To quote our conclusion: "Furthermore, when looking beyond 2028, the current surge in data center electricity demand should be put in the context of the much larger electricity demand expected over the next few decades from a combination of electric vehicle adoption, onshoring of manufacturing, hydrogen utilization, and the electrification of industry and buildings."
It is a source of demand growth, it is worth keeping an eye on, but according to our model is it not the largest new source of electricity demand.
What is the expectation for electric vehicle adoption, hydrogen utilization, and electrification of industry and buildings? I know that many people write as if these trends are already established, but I've seen no clear argument that they are technically or economically feasible, nor that anyone is even considering building the industrial base to support such moves. For instance, my back of the envelope calculations would require a tenfold increase in global lithium production to allow the US to replace all new car sales with electric vehicles, leaving aside EVs in Europe, Japan, China, or other parts of the world. Did your research touch on this aspect?
While I do other work (with some of the same people) on critical materials supply, those other topics were out of scope for this report, and the specific demand projections for those sectors isn't really my topic area. So caveat this with I haven't actually done an extensive bottom-up technoeconomic model like I have in data centers.
I'll cite the International Energy Agency's annual energy outlook (https://www.iea.org/reports/world-energy-outlook-2024). Their data center specific energy estimates are consistent with our results, so I'm reasonably comfortable taking their line on this. To quote that report:
"The main drivers of [electricity demand] increase [by 2030] are the electrification of transport, industrial processes and space heating, and rising demand for appliances and cooling...In the base case, data centres account for less than 10% of total electricity demand growth at the global level, which is roughly on a par with demand growth for desalination, and less than a third of the demand growth for both EVs and space cooling in the buildings sector. Therefore, growth of electricity demand for data centres is projected to be rapid, but the level looks set to remain relatively small in the context of overall global demand growth"
It's worth noting that critical mineral supply can be inherently circular. A fossil based energy system depletes, while we can economically harvest large fractions (depending on chemistry and recycling technique up to 95%) of cathode active material in batteries at EoL (https://www.sciencedirect.com/science/article/abs/pii/S0921344923001763). There's a short run question about bringing enough processing capacity online but that's different from "we don't have enough lithium" (https://wires.onlinelibrary.wiley.com/doi/abs/10.1002/wcc.768, p9)
Building electrification is already the most life-cycle cost effective way to construct new buildings (https://rmi.org/economics-of-electrifying-buildings/), so that's not a vaporware trend. Pem hydrogen is an open question, but that's on the demand side not constructing the electrolyzer side.
I don't know if that answered your question, but this is some context
Long before hard limits of waste heat hit planet-boiling levels, we'll have enough transit capacity to export energy-intensive industries to space, at which point radiator area is no longer constrained by the planet's surface geometry, so that thermodynamic model's assumptions break down.
You sound pretty confident about this fantastical scenario. I suspect the more likely reason humans will avoid the hard waste heat limit is that organized industrial society will disintegrate as population collapses along with potential exacerbations from large scale pandemics and nuclear war.
You sound pretty confident in dismissing businesses which already exist as "fantastical." GPS is old news. Starlink is steadily expanding, and profitable enough to prop up other strategic interests while competing with ground-based service providers. Cost per kilogram to LEO goes down, real estate prices go up, soon enough it's cheaper and easier to build the next marginal datacenter out there where solar power isn't intermittent, NEPA's irrelevant, and heat sinks can hit single-digit kelvin without even needing any moving parts.
Then somebody gets annoyed with debris avoidance and Coronal-Mass-Ejection-related downtime, realizes they can solve two stones with one bird by having maintenance bots harvest asteroids to add more spall jackets and radiation shielding... and as long as you're mining anyway, might as well check if any of that space dirt has rare-earth minerals which let you brag about avoiding planetside child labor... then smelters (with huge inflatable mirrors for cheap, super-high-grade process heat) move closer to the ore, machinists follow the smelters (first to save transit costs, which would likely still be on the order of $10 or $100 / kg, then exploring microgravity-exclusive possibilities such as foamed composites or convectionless chemistry), and so on through the tech stack. https://projectrho.com/public_html/rocket/infrastructure.php
> organized industrial society will disintegrate as population collapses
Just how bad are you expecting things to get? Usual estimate I hear is that ten thousand people would be enough to maintain both enough genetic diversity to avoid inbreeding, and all the essential skillsets of an industrial economy. Could squeeze through a population bottleneck of a few hundred with careful planning https://wiki.opensourceecology.org/wiki/Global_Village_Construction_Set and a lot less personal freedom or safety margin against unexpected developments. Ten million or so for a more well-rounded modern tech base, jets and advanced electronics and suchlike - meaning that's the real minimum for a self-sufficient colony on planets without natively breathable atmosphere or edible wild plants.
> along with potential exacerbations from large scale pandemics
Smallpox is extinct in the wild. Polio, guinea worm, and Yersinia pestis, nearly so. Vaccines are already rolling out against malaria and tooth decay... well, that latter is technically more of an engineered symbiote. Techniques developed for Covid could generalize to the common cold. Many species have inconvenienced humanity, few with lasting success. The lucky ones get domesticated.
> and nuclear war
Against who? The Russians? Putin would sooner peel his own face off, assume a new identity, and live out his twilight years in (luxurious) exile, than push the big red button and go down in history as the man who sealed their doom. He didn't get to be the last one standing in all those internal KGB power struggles by being brave and principled.
Chinese "wolf warrior diplomacy" is rooted in getting symbolic revenge for the Century of Humiliation, proving that their core virtues were right all along, that everybody who bullied them was inherently flawed, and that now they know the tricks they can be *even better* at bullying. That's why they over-commit to infrastructure, domestically clamp down on anything even vaguely isomorphic to paying silver for opium, and gleefully inflict such vices on foreigners. Nuclear annihilation isn't a plausible route to the moral high ground, so they'll avoid it, just like they carefully avoid unambiguous acts of war while hunting for Air Bud exceptions in the international definition of "territorial waters."
Trump is too senile to issue a properly-formatted launch order, and too arrogant to delegate it. Even if he did manage to follow official procedures, everyone else in the relevant chain of command is too sane to play along without a good reason, and he won't be able to provide one of those, because he gives more weight to G.I. Joe cartoons and Nazi propaganda than nonfiction sources when it comes to figuring out how a military social context is supposed to work.
No other credible US candidate is in favor of first-use, and no other country has enough warheads and delivery systems for much besides thwarting an otherwise-overwhelming conventional invasion of their own territory - one of many reasons military conquest is rarely even attempted in the developed world.
Great stuff as usual
I will use this for my students. It is useful to understand the total energy process in order to make informed decisions.
As is making polysilicon for solar much less building a solar installation and the fact that solar panels can theoretically last 30 years, the practical reality is much less <20 years…solar wouldn’t be possible without mountains of cheap Chinese coal… grow up everyone
So nobody has good data…and credulity is a vice
I can’t speak to all manufacturers, my SunPower panels have been demonstrated to have a 40 year life. They don’t actually die after that time, they slowly loose efficiency over time.
Nitpick: a turbine does not put (much) energy into the kinetic energy of the machinery's motion. That would involve rotating absurdly fast! Rather, it uses that motion, against the resistance of the load, as a medium to transfer the energy to another form.
The best thing we can do for carbon emissions is displace coal with natural gas and eventually nuclear. Replacing coal with natural gas would take us almost halfway to netzero. Which is lot closer than we will ever get by covering the planet with windmills and panels. And natural gas extraction is much less harmful than coal extraction.