We just need much cheaper catalysts to enable these sorts of industries that could be a very efficient use of ultra cheap excess electricity. Right now catalysts to create hydrocarbon fuels from the atmosphere are so expensive that it isn't worth it to build them, only to run them the ~20% of the time when electricity is so abundant that it can't be stored.
There's a reason people watch movies on planes - it's to distract you from the uncomfortable environment you're in. VR is more immersive than a 2D screen so it's more distracting.
Same - take lightbulbs for example. Thanks to LEDs, the amount of energy you need to generate X amount of light has reduced considerably, but we still have as much if not more lighting than ever. And it's not like our GDP is suffering due to a lack of sufficient lighting, at a certain point there's no gain to productivity gained from having another lightbulb. Same thing can be said about cars, CPUs etc.
It is a curious assertion for sure, but the next wave of AI will be throttled by a few factors: chip fab capacity, water supply to cool data centers, and electricity to power the chips/servers. Look at what is happening with xAI in Memphis where they are illegally running a dozen turbine generators to power their new AI data center since they can't get enough supply off the grid.
Costs are going down a lot due to algorithmic improvements. But maybe that just results in increased usage (Jevons paradox)? When there are conflicting trends, the future is pretty hard to predict.
Are any AI companies making money now? Losses can’t go on forever.
It may be throttled by energy supply and water supply at a desirable site, but not by country-level energy costs and water costs, which is all this blog post is looking at.
> Think it’s meant more as a broad generalization than something that is always true.
I think you're right, but also that the majority of the problems with the worlds economies (in the richer nations) are because of similar generalizations, and as such I think it important to rebuke them.
Having more cheap energy available is good (all else being equal), but optimising for higher energy usage is absurd.
The Tiwai Point aluminum smelter uses 13% of New Zealand's electricity [0]
It's overseas owners are constantly playing hardball with the country over the price they pay. Feels like every year they threaten to shut the smelter down unless they get better electricity rates.
Exporting aluminum is basically exporting electricity, except aluminum is easy to ship, costs very little to store, and has an indefinite shelf life. For places with a lot of natural gas and no pipelines to export it, it’s often easier to export aluminum than liquified gas.
if you're burning it in aluminum-air fuel cells, it can be literally exporting electricity. right now that isn't a commercial-scale activity, but possibly it will become profitable in the coming years for places with a lot of solar power and no hvdc lines to export it
for some economic activities, energy is not a limiting input; you are implicitly referring to economic production enabled by electric lighting, such as office work, and indeed energy has not been a limiting input for that for at least a century. reducing the cost of energy will not result in more gdp in those sectors
for other economic activities, such as solar panel production, aluminum production, and neural network training, energy is a limiting input. reducing the cost of energy will result in more gdp in those sectors
there's usually a long lag between a drop in the price of an input and the eventual impact on the price of the outputs, because part of the effect is mediated by the adoption of innovations that use more of the newly-cheaper inputs and less of the still-expensive inputs
to take one example, the last time we got access to a major new source of energy was something like watt's steam-engine in 01776. one of the effects of this was the widespread replacement of steel cans (which hadn't been invented in 01776) and glass bottles with aluminum cans in the 01970s, 200 years later. another was the replacement of travel by ship with travel by air, also about 200 years later. the delay is because many intermediating innovations were required, for example, in the aluminum-can case:
- the discovery of electrolysis;
- the discovery of aluminum;
- the discovery of canning;
- the hall–héroult process;
- improved aluminum alloys that permitted the use of 100μm-thick cans;
- the invention of deep drawing;
- epoxy liners that made aluminum cans chemically stable to acidic contents such as coca-cola;
- long-distance trucking which increased the cost imposed by heavier glass bottles.
the issue with nuclear power is that the humans don't yet have the technology to exploit it economically; at their current primitive level it's uncompetitive with other sources of energy. like printing 1000 years ago or heron's aeolipile
but 1.1 gigawatts of mainstream solar panels is 0.14 billion usd. $130 per kilowatt of capacity. even at the dismal 10% solar capacity factor achieved in very northerly countries like germany, the reactor is twice the price per average watt, and it needs to be installed far from the point of use—you can't buy a 440-watt nuclear reactor, so you need transmission, distribution, and transformers, all of which incur energy losses, capital investment, and safety hazards you can avoid with photovoltaic
that large grid also needs regulation, billing, and political stability. (a reactor is an appealing target for both russian glide bombs and enron-style scams.) and the reactor is not dispatchable over timescales of less than a day, while you can short out a solar panel in microseconds
fundamentally the reactor can't compete economically because it's shackled to a pricey steam engine. the reactor itself is a triviality, just a pile of fuel larger than the critical mass. some of them formed naturally at oklo billions of years ago. what's hard is integrating that energy release mechanism into a machine, and that's because the humans are still terrible at making machines
> but 1.1 gigawatts of mainstream solar panels is 0.14 billion usd
A solar farm is more than just solar panels. This 3.5GW solar farm cost 2.13B USD, so by your estimates the panels make up just 1/5 of the cost of the farm. I'd expect the load factor of the nuclear power station to offset the solar farm's nameplate capacity advantage, and lead to steadier prices/fewer storage requirements etc etc.
> and it needs to be installed far from the point of use
Note that this is a problem for solar farms in China; they are installed where land is not valuable. Hence all the HVDC transmission records being broken in China. Plus nuclear power stations can be close to populations. For instance https://en.wikipedia.org/wiki/Daya_Bay_Nuclear_Power_Plant is 50km from Hong Kong.
> the reactor is not dispatchable over timescales of less than a day
Modern reactors have load following capabilities, e.g. the AP1000 can ramp up 5% a minute within the 15%-100% band.
Pure PV farms have minimal operational costs, nuclear has huge ongoing costs. For a more realistic comparison the operating costs of nuclear offset the cost of batteries for solar.
So capital costs vs capital costs on a per Wh basis isn’t in favor of Solar it favors nuclear which has less flexibility. IE: 24 GWh per day of battery backed solar can dump half that power over 2 hours @ 6GW. 24GWh of nuclear IE a 1GWh reactor caps out at 1GW. If you want to ramp to 6GW of output nuclear needs several nuclear reactors and all of their associated costs.
> Modern reactors have load following capabilities
Load following isn’t free for nuclear, any time you’re not operating at 100% you’re losing money. Batteries are inherently way more flexible.
It also costs more to build a load following reactor and they have more experience maintenance issue due to thermal stress. Nuclear inherently favors steady state operations due to the Xe pit (https://en.wikipedia.org/wiki/Iodine_pit) but it also requires being taken offline for long periods for maintaining, refurbishing, and or refueling.
As for batteries, I think a few hundred USD/kWh is a reasonable guesstimate of cost (raw LiFePO4 cells are now sub-100 USD/kWh). Backing up each hour of production of a 1GW power station would cost a few hundred million USD, plus the cost of the solar farm to charge the battery up.
> 24 GWh per day of battery backed solar can dump half that power over 2 hours @ 6GW
At which point the transmission becomes the limitation; the grid operator probably wants a fairly stable flow of electricity through the wires to maximise utilisation so the 6GW is not realistic, nor would moving the electricity during the day to load-adjacent storage be efficient.
> Load following isn’t free for nuclear, any time you’re not operating at 100% you’re losing money.
I was responding to the point that solar panels are inherently more flexible because you can turn them off (because ...????). The same reasoning you've made about nuclear load following being uneconomical can be made about pure solar too.
> Nuclear inherently favors steady state operations due to the Xe pit
Operators change the boron concentration to offset the negative change in reactivity due to Xe-135 levels. For PWRs this is not a big problem, you just have to know it is there and do the calculation for I/Xe concentrations given the power levels.
I disagree with your quoted numbers. They aren’t current or inflation adjust to the same year, they also exclude several costs associated with nuclear such as insurance and setting money aside for decommissioning.
Ex: Your quoted fuel costs would be 0.9c/kWh in (2020 publish date) = 1.3c/kWh in 2024. O&M is often quoted as 4x fuel costs so 5.2kWh. “Fuel costs account for about 28% of a nuclear plant's operating expenses.” https://en.wikipedia.org/wiki/Economics_of_nuclear_power_pla...
A battery system which costs 200$/kWh and does 5,000 cycles = 5c/kWh. (Not every kWh from a solar farm needs to be stored, but this is just a ballpark comparison.)
> At which point the transmission becomes the limitation; the grid operator probably wants a fairly stable flow of electricity through the wires to maximise utilisation
You’re missing the forest for the trees here. Utilization follows demand, a state with peak demand of 6GW is going to have transmission lines setup for 6GW. But comparing the options you have nuclear with 4x 1.5GW reactors averaging ~40% utilization or batteries backed by solar. Run the numbers and Solar wins by a landslide.
> They aren’t current or inflation adjust to the same year
Page 41 states
All costs are reported here in 2018 USD terms.
> several costs associated with nuclear such as insurance
Insurance is required for any industrial facility. The IEA report does not mention insurance. https://world-nuclear.org/information-library/safety-and-sec... puts insurance costs at around 1M USD/year (and separate conditional payments if an accident does happen), which divided by 9M MWh/reactor does not work out to much.
> setting money aside for decommissioning
For nuclear between 0.01 and 0.39 USD/MWh, and solar between 0.03 and 0.58 USD/MWh (depending on discounting).
> O&M is often quoted as 4x fuel costs
The data in the IEA report differs; it is somewhere between the fuel costs and twice the fuel costs.
> Not every kWh from a solar farm needs to be stored
Rooftop solar will cannibalise the utility solar's daytime market. The demand for utility solar's energy will for the most part occur when the sun does not shine.
> a state with peak demand of 6GW is going to have transmission lines setup for 6GW.
But this ignores the physicality of the grid; power stations are dispatched based on location as well as availability because transmission is expensive to build and limited in capacity.
> you have nuclear with 4x 1.5GW reactors averaging ~40% utilization
So your demand model is 2GW for 22 hours and 6GW for 2 hours, right? Are there many places which exhibit such wild swings? Dynamic pricing/load shifting, pumped hydro and OCGTs would be the traditional solutions.
> O&M is often quoted as 4x fuel costs
The data in the IEA report differs; it is somewhere between the fuel costs and twice the fuel costs.
Operations & Maintenance must include fuel costs… They are doing the thing where they break actual costs into several buckets to make actual operational costs seem lower. Refurbishment isn’t maintenance yadda yadda.
Same deal is going on with insurance. That 1.1 M / year covers some nuclear accidents, but the self insurance risk is quite significant even if you exclude the risk subsidy assumed by governments. IE: In the event of a large scale disaster insurance doesn’t make the reactor owner whole meaning their out the value of 1 or more nuclear reactors.
So yea 1.1M / year only works out to 0.01c/kWh but that’s an underestimate.
> Rooftop solar will cannibalise the utility solar's daytime market. The demand for utility solar's energy will for the most part occur when the sun does not shine.
Even assuming vastly more rooftop solar… PV panels produce power on a long tail curve not just at peak hours. Rooftop solar however only supplies the grid with power after the houses needs are met which is a significantly narrower area.
Do you have references for how much a solar plant costs to build and maintain? A breakdown of costs would be good.
> Operations & Maintenance must include fuel costs…
I presume this was done to make section 5.4 "Fuel cost sensitivity" easy.
> Rooftop solar however only supplies the grid with power after the houses needs are met which is a significantly narrower area.
What about if people are over-specifying their solar PV system to make use of net-energy-metering (or high feed in tariffs) to reduce their annual bill (for instance in California)?
> What about if people are over-specifying their solar PV system to make use of net-energy-metering (or high feed in tariffs) to reduce their annual bill (for instance in California)?
Don’t just think about what happens when these systems are at 100% output. At 5% output that home is sucking power from the grid while the solar far is providing the grid with power. Which means even if every home and business adds panels a solar farm will still supply some electricity directly.
i agree that retric's reasoning about capital utilization efficiency applies to both nuclear and solar generation; in fact, it might be even more applicable to solar generation, because although the variable costs (cost of sales, you might say—proportional to power usage) for nuclear energy are low, they're virtually zero for pv. what i was saying about dispatchability is that nuclear plants typically can't be turned off; they keep generating power even when grid prices go negative, so the nuclear plant operator is literally paying someone with a giant resistor bank to burn up the energy the reactors are generating. (nowadays, hopefully they're paying someone with a battery bank to charge their batteries instead, but the future is not yet widely distributed.) your earlier comment about the ap1000 having significant dispatchability is welcome news, and https://en.wikipedia.org/wiki/AP1000 says there are six of them already in operation
there are particularly perverse grid incentives in some places which result in pv farms continuing to operate when grid prices go negative, too, but that's just a fake market; nothing about the generation technology requires that. if you close a contactor to short out your solar panel, it stops dumping any energy into the grid in nanoseconds, literally faster than the contact bounce in your contactor, without any damage or risk to the panel, power electronics, or the rest of the plant
with respect to transmission and battery storage, while there is some reason to locate pv farms some distance from the energy consumers—the consumers may be tightly packed and/or in a cloudy area—there is no reason to locate battery storage far away from energy consumers. you want batteries to be as spread-out as possible, as close to the load as possible, for many reasons: to avoid time-of-day congestion of transmission capacity (and even distribution! point-of-use batteries reduce or eliminate the need to overprovision distribution capacity); to prevent fires from spreading from one battery to another; to eliminate power outages caused by problems in transmission and distribution; to eliminate transmission and distribution energy losses for stored energy; and to reduce the opportunities for rent-seeking by transmission and distribution operators. the land use, climate, and safety considerations that sometimes limit the distribution of pv spreading-out don't apply to spreading battery storage out
as for the o&m costs: while the iea does wonderful work, and i appreciate you pointing to this very informative open-access report, this report is from december 02020, and it's largely built on data from previous years, much of it from plants built years earlier. the main topic of the tomas pueyo article we're commenting on here is how lower prices for solar panels are forcing people to design new solar power generation in ways that 'waste' solar panels in order to commensurately reduce the other associated costs, such as the operation & maintenance costs you refer to in table 3.14
with that in mind, looking at https://www.solarserver.de/photovoltaik-preis-pv-modul-preis..., pvxchange's current mainstream panel price index is €0,12 per peak watt, and in september 02017 (probably about the average time the plants profiled in the iea report were being built, and as far back as the data currently on that page goes, though archive.org has older versions) it was €0,42 per peak watt. that is to say, solar pv modules cost 250% more at the time. those solar farms were designed for a very different world than the one we live in today, one that could tolerate much higher o&m costs in order to make better use of the comparatively scarcer solar panels
i agree that the difference in capacity factor is very important, and i should have made that clearer in my comment. nuclear is typically around 85%, solar typically around 20%. solar farms in the california desert are 29%, so this desert plant might have a similarly high capacity factor, but last time i checked, the prc average was more like 10%, and i don't understand why. possibly factors like transmission congestion are to blame and will be at play here too
especially if it's cheaper to put up more solar panels somewhere more overcast than to build hvdc transmission lines from urumqi to shanghai
it turns out that, if you use solar panels the same way you'd use nuclear reactors, by centralizing them hundreds or thousands of kilometers away from where the energy is used (as in this case), or by concreting over prime beachfront property (which nuclear power plants need) to build giant solar farms on, they can cost almost as much as nuclear reactors do, or even more
this is analogous to how factories first used electric motors: they installed a giant electric motor in the factory's powerhouse to drive the line shafts, replacing the steam-engine the powerhouse was built for. consequently electrification famously didn't increase factory productivity for decades
when i said that nuclear power plants 'need to be installed far from the point of use', i didn't mean that they couldn't be tens of kilometers, or even single-digit kilometers, from the point of use. i meant that they can't be single-digit meters from the point of use. solar panels can, and that dramatically drops costs
i appreciate the correction about the ap1000! naval nuclear reactors have been able to rapidly ramp up and down since forever, so it's good to see that capability making it into commercial nuclear power
> i meant that they can't be single-digit meters from the point of use. solar panels can, and that dramatically drops costs
Transmission costs, yes. Plus if the solar is behind-the-meter you might avoid some of the taxes and levies applied to grid electricity.
(Note that I realise the focus of my comment from here on down has changed from China to the UK, but then again I've not helped install a solar installation over there!)
However with UK rooftop solar home-owners do not have much negotiating power as the market supply is restricted by the MCS scheme (Microgeneration Certification Scheme). This may be changing in the future (Flexi-Orb scheme), but until a greater pool of competent installers are in the market the prices will not decrease.
A relative had 6.4kW solar (and 5kW hybrid inverter) installed last summer for around £7,000. I added in some batteries for another few thousand. The panels generated around 5,100 kWh last year, for a capacity factor of around 9%.
yeah, the uk is pretty miserable for solar! and in terms of regulation, it's definitely not the worst place in the world, but it's definitely not going to be leading the transition to renewable energy the way it led the transition to steam either
One larger cost you might think of with solar is land - but even in the U.K. where land isn’t exactly cheap leasing prices are about £1k an acre per year, and an acre will generate about 350MWh a year, so that’s well under 1 cent per kWh, so it’s lost in the noise.
where land cost comes in is that it forces you to put solar generation far away from energy consumption, which incurs transmission and distribution costs which can be several times larger than the cost of the generation itself, as documented elsewhere in this thread for urumqi
as an example, a 100-megawatt electric arc furnace might occupy 1000 square meters, and it's amenable to solar's intermittent energy supply in a way that blast furnaces aren't, but even at the ideal kilowatt per square meter, it needs 100 000 square meters of solar panels to power it, about ten city blocks. more plausibly it needs several times that. you can't physically fit those panels closer than hundreds of meters from the arc furnace, and land costs mean you probably have to put them out in the countryside, likely tens of kilometers away
fortunately grandma is politically irrelevant in china (ayi's are worse than Mao), though the age of its leadership is showing, as they are still obsessed in trying "socialism" and see skyscrapers as the ultimate hallmark of development (they built so many cathedrals in the sky the government had to put moratoriums on financial requirements to build them)
It reduces the number of shares on the public market, not the total number of shares.
This can drive the price up if there is high public demand for the now reduced number of shares available
