seeing how 2GW of nuclear cost $34B in Georgia, why would Google waste $120B when they can get the same output for at most half the price (and realistically more like 1/10th) using renewables and batteries? and they’d have results in 2 years instead of 2 decades.
edit: to be clear, 1GW of wind or solar is $1B. Build 3GW for overcapacity and you’re still at just 17% of the cost of 1GW of nuclear, and you technically have 3x more capacity. Now figure out how many megapacks you can buy for the $14B/GW you saved https://www.tesla.com/megapack/design (answer: 16GW/68GWh)
> using renewables and batteries? and they’d have results in 2 years instead of 2 decades
We have nothing close to the battery fabrication pipeline to make that timeline true, certainly not at scale. If this move works, Google will have cemented its power needs and economics for decades to come.
Global battery manufacturing capacity was 2,600GWh in 2023 [1], and has probably already exceeded that this year. The IEA projects closer to 4TWh by 2025, and nearly 7TWh by 2030 [2].
You need to pay attention because this is happening fast.
That's a big number. Here's a bigger one: 30,000 TWh, about our current electricity consumption [1]. 7 TWh in 2030 is less than 1/4,000th total electriciy production today. (You obviously don't need 1:1 coverage. But 2 hours in 2030 against a year's demand today is still a nudge.)
Now consider EVs. Then add the tens of TWh of annual power demand AI is expected to add to power demand [2]. (And I'm assuming a free market for battery cells, which obviously isn't where we're heading. So add local production bottlenecks to the mix.)
Battery numbers are going up. But they aren't going up fast enough and never could have, not unless we ditch electrifying transportation. Nukes or gas. Anyone pretending there is a third way is defaulting to one or the other.
Investing in nuclear power today is an insane prospect when the energy market is being reshaped at this speed.
In Europe old paid off nuclear plants are regularly being forced off the markets due to supplying too expensive energy.
This will only worsen the nuclear business case as renewable expansion continues, today being a bonanza fueled by finally finding an energy source cheaper than fossil fuel.
Nuclear power is essentially pissing against the wind hoping the 1960s returns.
> In Europe old paid off nuclear plants are regularly being forced off the markets due to supplying too expensive energy.
This is happening because of subsidies given to renewables (renewable energy certificates, net metering, guaranteed feed in prices, CFD) plus policies at the national and EU level (EU Renewable Energy Directive). Take away these policies and you are left with a low quality (intermittent) energy source that requires far more expensive storage to produce power when it is needed.
A study recently found that a nuclear powered grid to be vastly more expensive than a renewable grid when looking at total system cost.
Nuclear power needs to come down by 85% in cost to be equal to the renewable system.
Every dollar invested in nuclear today prolongs our reliance on fossil fuels. We get enormously more value of the money simply by building renewables.
> The study finds that investments in flexibility in the electricity supply are needed in both systems due to the constant production pattern of nuclear and the variability of renewable energy sources. However, the scenario with high nuclear implementation is 1.2 billion EUR more expensive annually compared to a scenario only based on renewables, with all systems completely balancing supply and demand across all energy sectors in every hour. For nuclear power to be cost competitive with renewables an investment cost of 1.55 MEUR/MW must be achieved, which is substantially below any cost projection for nuclear power.
Which is confirmed by Sweden continuing its renewable buildout with both the cheapest electricity prices in Europe and no subsidies on the books for new renewable production.
> For nuclear power to be cost competitive with renewables an investment cost of 1.55 MEUR/MW must be achieved
Cost assumptions in table 2
Offshore wind 1.9M EUR/MW, 1.67% O&M, 30 year life at 0.51 capacity factor
Onshore wind 1.03M EUR/MW, 2.51% O&M, 30 year life at 0.37 capacity factor
Solar PV 0.6M EUR/MW, 1.50% O&M, 40 year life at 0.14 capacity factor
So they are claiming nuclear (which has a > 0.9 capacity factor in Finland, and 60 year life) needs to have an investment cost between onshore and offshore wind to make sense.
Due to energy system constraints, there might be reasons for down regulating the nuclear power stations, thus as an output the capacity factor might be lower than 90%, but never higher. The study allows for nuclear power to be down regulated to 25% of the maximum load in for instance hours with high wind and solar production.
So the authors decided that the non-dispatchable wind/solar has market priority over nuclear. Hence it is important to pack out the high nuclear scenario with renewables. Also note how the all renewables scenario adds biogas (presumably from all the pig slurry) to firm up demand along with 6GW of inter-connectors to friendly neighbours.
Edit :- Closer analysis of the high nuclear with district heating scenario (figure 4, in the supplementary material) reveals a total electrical demand of just under 10,000MW (unflexible + heating + transport). Note that the authors have chosen to represent nuclear as a continuous 6,686MW of power (rather than the nameplate capacity of 7,400MW).
> Which is confirmed by Sweden continuing its renewable buildout with both the cheapest electricity prices in Europe and no subsidies on the books for new renewable production.
The Swedish government raised its subsidy for solar cell installations from 15% to 20% in January 2023. ... The income tax reduction for households and businesses that micro-produce renewable energy was introduced in 2015
Even without subsidies solar and battery are cheaper than nuclear and are getting cheaper by 15-20% a year. So no nuclear is unlikely to be cost competitive any time soon unless they get some new tech for nuclear
> nuclear power today is an insane prospect when the energy market is being reshaped at this speed
We’re still more than a decade away from having enough batteries to make this shift. Again, excluding EVs and AI. That’s why we’re reänimating coal plants and building new gas turbines.
I’d also love to see the numbers on that simulation going from 98.6% coverage to what we expect from a modern grid. (And if the balance is provided by gas or something else.) It should surprise nobody that going from 1 sigma to 2 can cost as much as 2 to 3, even if the percentage gap is much smaller.
> Europe old paid off nuclear plants are regularly being forced off the markets due to supplying too expensive energy
Europe has invested €1.5tn into new gas infrastructure. That doesn’t go poor without a fight and collateral damage.
> We’re still more than a decade away from having enough batteries to make this shift.
A decade to have significant amounts of battery storage is actually a pretty optimistic timeline compared to nuclear. Nuclear plant construction times are on the order of a decade or (realistically) two decades in the West, if you include planning. In China they're managing 7 years, but their nuclear buildouts, while impressive, aren't trending an upward path when compared to renewables (see chart here [1].) SMRs might change this, but they're years from leaving "research" status and entering the mass-production/learning curves that could make them cost competitive.
This doesn't make me happy. If I thought nuclear was viable on the timelines we have to dampen climate change, I'd be 100% in favor of it. If we could assemble the political will to raise taxes and build nuclear at "wartime" speeds, I'd say go for it. I'm also very much in favor of SMR development, just not willing to bet the house on it.
As it stands, there isn't anywhere near enough nuclear power in the planning pipeline for nuclear to matter much on a 20 year timeline.
In any case, we are not going to a 100% renewable/battery grid in 10 years. The first goal is to get renewables to 90-95% or more of power generation, massively overbuilt with short-term battery storage backed by intermittent fossil fuels for the remaining 5-10%. This will represent a massive reduction in emissions. The last 5-10% will have to be completed over the next couple of decades, and the increasing battery production trend gives hope that it can be.
The worst problem with existing nuclear is that with a 15-20 year planning/construction timeline and the current molasses build rate, new nuclear plants will arrive right at the moment when cheap storage is eating the economic use-cases that make them financially viable.
> * Nuclear plant construction times are on the order of a decade or (realistically) two decades in the West, if you include planning*
Sure. Forecasting twenty years out is tough. But our forecasts out 10 years show the power crunch easing to almost no degree--we'll still likely be making the same tradeoff then as now. (And, I suspect, still filling the gap with gas in teh west.)
You're broadly correct: we need to build faster. There is no reason we can't build a large plant in under a decade and an SMR in a few years. The latter is what Google is experimenting with here. It's a long shot. But so is hoping battery production scales the orders of magnitude necessary for it to become a utlity backbone over the next decades.
> first goal is to get renewables to 90-95% or more of power generation, massively overbuilt with short-term battery storage
We don't have the battery pipeline. What we're repeatedly getting is renewables plus gas generators. There is no world in which you put down trillions of dollars of gas infrastructure and then poof it in a few years because it's no longer needed.
A study recently found that a nuclear powered grid to be vastly more expensive than a renewable grid when looking at total system cost.
Nuclear power needs to come down by 85% in cost to be equal to the renewable system.
Every dollar invested in nuclear today prolongs our reliance on fossil fuels. We get enormously more value of the money simply by building renewables.
The study finds that investments in flexibility in the electricity supply are needed in both systems due to the constant production pattern of nuclear and the variability of renewable energy sources. However, the scenario with high nuclear implementation is 1.2 billion EUR more expensive annually compared to a scenario only based on renewables, with all systems completely balancing supply and demand across all energy sectors in every hour. For nuclear power to be cost competitive with renewables an investment cost of 1.55 MEUR/MW must be achieved, which is substantially below any cost projection for nuclear power.
> a nuclear powered grid to be vastly more expensive than a renewable grid when looking at total system cost
Yes, nuclear is more expensive. SMRs should help with that, but their expense has never been contested.
But marginal economics aren't everything. Renewable and battery production isn't ramping up fast enough to make that margin available at scale. This doesn't seem capital contrained, either--every major economy is throwing gobs of cash at the problem.
> Every dollar invested in nuclear today prolongs our reliance on fossil fuels. We get enormously more value of the money simply by building renewables
False economy. A dollar not invested into nukes doesn't go into renewables--partly because of the aforementioned scaling problem, it tends to wind up in gas.
We’re spending trillions of dollars of new money on gas infrastructure with decades of life and financial liabilities attached to them because we need the power, have maxed out renewables and are left with a choice: gas or nukes. Opposing nukes isn’t playing for renewables, it’s playing for gas.
SMRs can potentially do something that renewables can’t: they could be placed near the loads in places with no space for renewables and without relying on the grid. Think industrial areas or even cities or towns that are surrounded by other developed land. The grid moves slowly, and electricity prices via existing transmission lines are, in many areas, hilariously inflated for a number of reasons. A hypothetical portable, easy-to-acquire SMR producing power at $100/MWh would not be an amazing deal if a large electric utility bought it, but a $100/MWh would be an amazing price in quite a few markets if a small utility could actually buy at that price and deliver via a small last-mile distribution system.
Yes, that shit study which models supplying the entire grid with one energy source and lithium storage through all weather conditions.
I would suggest reading the study I linked so you can see the difference in methodology when credible researches in the field tackle similar questions.
The credible studies are focused on simulating the energy system and market with real world constraints. Which apparently works out way cheaper when not involving nuclear in the picture.
That entire report is an exercise in selectively choosing data to misrepresent renewables and present nuclear power in the best possible light and wishful thinking.
To the degree that the prominent "renewables vs. nuclear" graph they keep repeating on the webpage and figure 6 in the report is straight up misleading.
This is the source:
What is different about different net-zero carbon electricity systems?
Utilizing storage costs from 2018 and then of course making the comparison against the model not incorporating any hydrogen derived zero carbon fuel to solve seasonal problems.
Which is todays suggestion for solving the final 1-2% requiring seasonal storage in the late 2030s.
Something akin to todays peaker plants financed on capacity because they run too little to be economical on their own, but zero carbon.
Would they have chosen the ReBF model the difference between made up optimal nuclear power and 2018 renewables would be: $80-94/MWh and $82-102/MWh.
It is essentially: Nukebros writes reports for nukebros, they confirm their own bias. Simply an attempt to justify another massive round of government subsidies on nuclear power.
lmao, you say shit study but you suggest using green h2 as backup which not only isn't economically feasible (for now at least) but current generators are either using a mix with gas or use pure h2 with huge nox releases due to high temp burning. Not just that, most lcoe costs magically assume that 4h storage is enough. Look at yesterday's Germany generation and tell me how 4h storage will be enough there. Or maybe I should link to amount of subsidies Germany is pouring each year in renewables like https://www.bloomberg.com/news/articles/2024-05-29/germany-s... or like https://www.reuters.com/business/energy/germany-looks-specia...
It's funny that when I ask ren-bros how much subsidies edf in France is getting they are either silent or are linking to price shielding that's totally unrelated and is present in most eu countries after russia's invasion.
Renewable bros as usual are dunking on nuclear and promoting their clean supply like a mecca without facing hard reality - most renewables now are subsidized by fossils and will be in any close future
>you suggest using green h2 as backup which not only isn't economically feasible (for now at least)
That's poor logic, h2 as a last-2%er doesn't need to be feasible until we've gotten to the 98% mark. And honestly, h2 feasibility is a function of cheap energy anyway, which probably means midday solar while solar farms are chasing dusk prices.
not, h2 feasibility in the context of power generation depends on many more factors, including how frequent the plant is used when day hours will be mostly tapped by solar generation and how you'll do price compensation. And in the context of h2 for renewables as a peaker, it'll need to be much more than 2%. And again, the emission problem for h2 generation isn't solved yet beyond fuel cells
Sorry, was writing on a mobile. Here's a more detailed explanation why it's pure BS.
Because it's simply magic thinking. They postulate a "future fully sector-coupled system" and then say that if this somehow magics into existance, then everything's peachy.
Basically, "a sector-coupled system" allows transforming excess power into something useful (district heating, hydrogen, steel, etc.), and shedding the load and/or providing some power back when there's not enough generated power available.
In other words, if you solve the problem of providing 1 month of energy storage for Germany and Denmark, then renewable energy is basically free. Duh.
The problem is that "sector-coupled systems" don't exist, and their creation will result in far, far, far, far more expenses than building fucking PWRs.
Yes, the study incorporates no lithium storage. Including storage we will easily reach far above 90% renewable penetration.
When we get to the final percent in the 2030s we can utilize akin to todays peaker plants financed on capacity markets [1] but zero carbon.
Peaker plants today already run too little to be economical on their own, essentially what in our current grids constitute seasonal storage and emergency reserves.
Simply update the terms for the capacity markets to require the fuel to be zero-carbon. It can be synfuels, biofuels or hydrogen. Whatever comes out the cheapest.
As we electrify transportation we can shift over the massive ethanol blending in gasoline in the US to be our seasonal buffer. [2]
> When we get to the final percent in the 2030s we can utilize akin to todays peaker plants financed on capacity markets [1] but zero carbon.
Capacity markets effectively don't exist in Europe right now. There are plans to create a plan for them by 2027, this is how urgent it is for Europe. But no worries, natural gas is now green, and it's fine to send money to Azerbaijan for it.
There is no pathway for most of Europe to switch to renewables any time soon.
In eu France is the biggest net exporter in the EU while Germany with huge renewable capacity net imported 20+TWh this year. Look how Germany's generation was yesterday to get a sneak peek
This is only because it is profitable for Germany to do so, not because of lack of capacity. Germany imports energy when there is low demand (and price) and exports when there is high demand (and price). Look at this chart: https://energy-charts.info/charts/power_trading/chart.htm?l=...
Again, look at yesterday generation. They were not able to satisfy local demand with renewables and bumped up coal+gas by a lot.
Also, if you look at the numbers - the price difference isn't that huge but trade difference is huge. This year export price is less than 1$ more than import. Problem is Germany net imported 25TWh so they are still in a big trade deficit and it continues to grow considering dunkelflaute is ahead
Yes, Germany is targeting a 80% renewable electricity mix by 2030 and 100% by 2035. They have no illusions about being perfect today. Their current status is 65% renewable for 2024.
Maybe stop looking at instants and start looking at the larger picture: keeping our cumulative emissions as low as possible.
Starting a nuclear construction project which won't deliver any decarbonization for 15-20 years is accepting large cumulative emissions.
they don't target 100% by 2035. They want to close last coal plant by 2038 which is a bit optimistic looking at yesterday's generation. For gas it's even worse - the plan is totally unrealistic and their planned h2 ready plants that'll use gas initially, will probably still use a mix with gas when/if green h2 becomes reality or they'll replace the generators with pure h2(unlikely) which has huge nox emissions due to high burn temperature
I had to check the numbers because it grabbed my attention.
No issue with your quoted figure of 30,000 TWh (annual) global electricity consumption.
But we only need to do 7TWh of battery supply in year 1 (or say only 1-2 of that makes it to grid storage).
30,000/365 is 82 TWh daily. So that’s the number to crack, surely? Because a significant percentage of storage will be to make up for wind and solar, which generally approximately follows some sort of diurnal cycle?
If we will be closing in on a couple TWh annual storage capacity in 6 years (leaving aside any real synchronised attempt to get vehicles to be part of large scale distributed grids) then only a few years on from 2030 we’re going to be able to store a significant percentage of our daily energy demands
"But 2 hours in 2030 against a year's demand today is still a nudge."
How much battery storage do you think we need? Surely not a year's worth.
For solar, we'd likely need 10-16 hours of storage to power stuff overnight. Maybe a little more to cover a few cloudy days. Sounds like we are about 5% of that now?
Generally the worst case is two weeks. In the middle of winter you often get cloudy low wind days for that long. Of course how you handle those worse cases are days need not be how you handle typical. If you can handle 16 hours of no input this will over the typical cases this will be enough to max a massive dent in carbon emissions and we can fall back to existing gas (or even coal) plants for the rest. Plus a lot of power use can turn off when needed - give my company a discount and we can turn the factory off.
10-16 hours is not enough at all. On a cloudy day, solar output will only be 15-20%. On top of that, your panels really only generate for 8 hours on a very good day - the sun is a lot dimmer in the early morning and late evening. Really, you need 2x storage for a good day, if you want to deal with two cloudy days you'd want 50-60 hours of storage.
Could you possibly read the article you're replying to again?
Even skimming through it discusses the coverage of wind and a not 50/50 system particularly to cover winter & night time. There is also discussion of a ~2% from "other" and how much storage capacity is required.
The article even goes into using wind & solar data for the simulation and reducing further the output to be conservative.
I obviously understand it's not a 100% solar system. If it was you would need to be able to deal with at least 2 weeks of bad weather, not two days, and you would have to take into account winter (dropping to about 5 hours instead of 8).
Additionally, mixing solar and wind is not as easy as it seems, because the two are correlated. If you have a major storm that makes wind energy impossible due to wind speeds above ~100km/h, you will also have clouds making solar energy unworkable. I'm not aware of any simulations modelling a 95+% solar/wind grid for storage needs, taking into account extreme weather patterns, grid topology, and equipment damage, but if you do then please link it.
I don't see any article linked in the comment I replied to. Perhaps you're mixing up two comment chains.
> That's a big number. Here's a bigger one: 30,000 TWh, about our current electricity consumption [1]. 7 TWh in 2030 is less than 1/4,000th total electriciy production today.
I don’t think anyone is seriously suggesting powering a portion of the grid with batteries that are cycled once per year. One can optimistically cycle one or even twice a day (if wind peaks when the sun is down). Or you can try to ride through a week of bad weather, but natural gas is not actually a terrible solution for that. And those batteries last for a lot longer than a year.
So I think your 1/4000 should be more like 1/10. Give it a few more years.
Natural gas is a great solution. It's why we're using it. But if your focus is decarbonisation and electrification, nuclear is better. Even if it's pricier.
> your 1/4000 should be more like 1/10. Give it a few more years
The former is calculated from projected 2030 battery production to present energy levels. An essential component of strategy is knowing on whose side time is. Battery production won't reach 1/10 for at least a few decades. That's the point. We need an intermediate solution, and if that's going to be gas, we have to live with the fact that (a) emissions will continue and (b) we perpetuate trillions of dollars of capital infrastructure that will be as difficult to take down in the future as coal has been today.
> Natural gas is a great solution. It's why we're using it. But if your focus is decarbonisation and electrification, nuclear is better. Even if it's pricier.
If you come up with some combination of carbon-free energy sources and storage that covers 90% of grid energy needs, and you need to fill in the gap, and that gap is a whole lot of power but only for a handful of days a year, then I don’t think nuclear is a good option at all to fill in the gap. The capital expense would be absurd.
Decarbonization is great, but in the real world, decarbonization per dollar spent is what matters. Instead of spending a zillion dollars on nuclear peaker plants, spend a lot fewer dollars on gas peaker plants and the the rest for more effective environmental improvements.
> Natural gas is a great solution. It's why we're using it. But if your focus is decarbonisation and electrification, nuclear is better. Even if it's pricier.
There's a crossover point. If you use natural gas to provide <1% of yearly electricity needs, and you save a zillion dollars while doing so, you can find cheaper ways to decarbonize by the same amount.
We'll figure it out. There is too much at stake and there are already a gazillion engineers out there going to bed every night thinking about how to solve this problem.
Innovation is the grim reaper of analyst reports. No one at my company notifies an investment bank when we have a breakthrough internally (lol).
Why are you comparing the rate of change of battery storage capacity, the vast majority of which if grid connected will be used for at most diurnal storage, to yearly energy consumption?
Holy mother of all type errors there.
Multiply it by 365, and it implies that in 2030 alone, we will create enough battery storage to time shift almost 10% of our total electricity use today.
I could just as easily assert the same of nuclear or gas. It doesn't make it true, although there seems to be evidence that nuclear cannot scale as fast as batteries/solar/wind.
Nobody claims renewables + battery doesn't work long term. (And not only work, but do so at rock-bottom costs.)
The problem is the timeline. Time out building that additional infrastructure, including expected demand growth, and you always need more power in the interim. Particularly if you're planning on taking coal offline.
If there is an arugment that we can ramp up battery production even faster than we are, the math changes. But we're already in a Herculean effort to mass produce more batteries faster.
Battery manufacturing capacity is greatly underutilized in China. That was battery cell prices there fell by nearly 1/2 in the last year. There is tremendous room for expansion of production.
The problem with nuclear in Georgia, and in the US, was that no one remembers/ed how to do it, and so all the lessons of yore had to be relearned, and the supply chain had to be stood up.
If you put in an order for several reactors, the very first one (especially of a new model, like Vogtle 3 was) will be expensive AF. The second will be expensive. All models after that will be at a more 'reasonable' cost.
Nuclear reactors are just like any other widget: the cost goes down with economies of scale. If you order 4 or 8 reactors at one sites they'll get progressively get cheaper (there is a floor of course). If you then put in an order at a second site, and move the workforce (or a portion) there, the lower costs will still be present.
If you start and stop construction, or order a whole bunch of different models/types, then there economies of scale goes out the window.
Sort of - nuke plants are fundamentally phenomenally complicated compared to true economies of scale technologies like solar. You won’t reap 100x cost savings in nukes, no matter how many you build
You can't just invent a number because you like it more. Solar and Wind are cheaper than nuclear even if you go beyond LCOE and include system costs. Even the nuclear lobby acknowledges this nowadays and has switched to other arguments.
Not even close. Wind and solar are cheap _only_ if you don't depend on them. In particular, for the wind the adequacy rating is about 10% in most places. It means that you can expect 10% of the nameplate capacity to be available at all times system-wide. So multiply the wind energy costs by 10x, and suddenly they are quite more expensive than nuclear.
It's not even a question for the solar, it simply can't provide power during a day without storage.
> Even the nuclear lobby acknowledges this nowadays and has switched to other arguments.
First, capacity factor is a silly metric to use for this. The industry uses LCOE or system LCOE, because this is about dollars per actual TWh produced, not capacity. In other words: A capacity factor of 10% doesn't matter if building 10x the amount is still cheaper.
With that said, the wind capacity factor in Germany is 20% for onshore and 40% for offshore, so even that was wrong by a factor of 2-4.
Because they need power 24/7 and not only when the weather cooperates.
And new AP1000s in the US would cost significantly less, because there are already experienced workers & supply chains from Vogtle and getting a license requires less work too, because you can copy much of Vogtle.
The median build time for nuclear reactors is 7 years. This is archivable if you continue building and not just build 1 or 2 every few decades.
The scale just isn't there. A single nuclear power plant near me, McGuire Nuclear Power Plant, produced 17,514 GW·h in 2005. The entire potential output of the Tesla (cough Panasonic) Gigafactories in California and China have a combined output of ~50 GWh per year. [0] Nuclear power is amazing at producing a reliable base load of power that massively outstrips our ability to produce and store solar power. Say our load is well aligned with the cycle of solar power and we're ignoring weather so we can derate the amount we want to store to 30% that's 105 years of production out of what I think is the two largest batter plants in existence to store the power produced continuously by a single large nuclear power plant.
Global stationary storage deployed for 2024 will be ~150GWh, and this is accelerating. Batteries are easy, nuclear appears to be impossible (economically speaking).
So 35 years then to store the power generated 24/7 by McGuire at that rate of production which ignores that the huge spike of AI loads will want 24/7 power, if we're looking at that kind of load I'd rate it at 50% for starters (low to be honest because it doesn't account for how solar ramps up during the day) which is around 60 years. Plus that's giving full capacity to those batteries when ideally we'd only use the middle 60% to avoid deep cycling the batteries daily unless they've completely solved that problem.
The nuclear ain't getting built, these are facts. Even if one breaks ground today, you won't push your first kwh to the grid for a decade, at which point another ~10TW of clean energy will have come online globally.
If AI is using too much power in the short term, destroy demand with policy and economics. We are not beholden to the robot trainers, we just don't provide utility access to the load. Unlimited demand of industrial scales of electrical power isn't a right of some sort.
Everyone is a libertarian until it’s their commons experiencing the tragedy. Strange to think that having regulations around large scale electrical loads is a dictatorship. It’s okay for us to collectively say no, depending on the circumstances.
> Everyone is a libertarian until it’s their commons experiencing the tragedy.
Yeah sure and everyone is a socialist utopian until its their own money/liberty on the line.
If we cannot “collectively” reach a consensus what happens?
Im just pointing out that the original suggestion of “ban ML for environmental reasons” is extreme/ham-fisted. This is what dictatorships do in real life all over the globe “for the greater good”.
Should crypto be a government-approved use of energy? What about manufacturing semiconductors? Building data centers? Producing EVs, solar farms or batteries? Are flights for vacation allowed?
You aren’t thinking about the second order impact of having a government that has the ability to gatekeep energy production for specific use cases…
And can be recharged as soon as the wind or the sun comes back.
The sizing of batteries and power sources is highly region specific, and the places where it makes sense today with current manufacturing capacity, don't have to be "everywhere" for it to be fine where it's actually done; and given the roll out rate of renewables, we also don't need to wait until battery output per year can totally displace the existing and currently running gas plants, just back up the newly installed renewables themselves - 4h in this case is how fast the PV farm would recharge those batteries in the best case, the average output of a PV plant is about 10% of the peak, so this is really a 40 hour battery pack not a 4 hour pack.
I mean, look at Germany's grid yesterday& today and tell me, with such overcapacity, how much more overcapacity it would need and how much storage it would need to cover such events with low wind and solar?
I use the approximation of annual capacity factors for PV being 10%, which means 10x. Wind has a higher capacity factor IIRC between 35% and 85% and that heavily depends on location.
A realistic answer would need me to spend at least a month dealing with finding historical satellite cloud cover data, wind records, correlations leading to nationwide dunkelflaute, the planning options for where new stuff can be built, etc.
And even then, that varies depending on international grid connections, and how much storage is on the grid.
cf yearly are good for some purposes but bad for others. Again, look at Germany's coal/gas use yesterday vs today as well as wind/solar generation and imports. If you don't want fossils, how would you cover such events? France was outputting towards Germany equivalent of 3-4 npp and 2 additional from Switzerland, max being about 12+GW from neighbors. How would it be financially viable considering there are many other days when demand will be met for day hours? New solar/wind will not be able to sell energy at negative prices unless they get subsidies. Germany already spends 20bn/yr for price subsidies and their grid is far from overcapacity and that doesn't account for other subsidy types like for transmission for renewables
Today's values, from what I've seen, this country could run on just wind if it had 10x more than now, but it doesn't really need that in isolation, it's just that PV was harder to judge because the graph wasn't even close to a flat line.
> New solar/wind will not be able to sell energy at negative prices unless they get subsidies.
They already do, in good weather.
> 20bn/yr for price subsidies and their grid is far from overcapacity
And how much of that was for a guaranteed price made way back when the stuff was still expensive?
New PV is, by itself, the single cheapest source of electricity; even adding on batteries only takes it up to somewhere between gas and nuclear depending on the specifics.
> and that doesn't account for other subsidy types like for transmission for renewables
How's that a subsidy? I've not seen the breakdown of bill costs here, but back in the UK there was a split between connection cost and use cost.
We haven't reached yet such renewable market penetration to get this problem. It'll happen when a lot of days, 10 day hours will be covered by renewable output.
> And how much of that was for a guaranteed price made way back when the stuff was still expensive?
I have no idea how are these are distributed. Do you have a link for recent vs old projects?
So using your numbers, it is solidly a little less than half the cost, not one tenth (26GWh seems around the necessarily amount for riding out ~14 hours of darkness. I'm assuming your factor of 3 makes up for seasonal variation and cloudy days). The panels take up 9 acres of land area, and need to be kept clean of snow and dust. The battery lifetime is small compared to expected life of a nuclear reactor, but the battery lifecycle is more straightforward. This seems like the territory of having a reasonable tradeoff between the two, not some unequivocal win for an Internet smackdown about how we should avoid one approach.
That is seemingly such an absurdly high number to get a nuclear planet up and running.
Is the majority of that cost dealing with regulatory and legal nonsense that stems from the anti-nuclear hippy groups and laws they got passed in the 60s and 70s?
> Is that majority of that cost dealing with regulatory and legal nonsense that stems from the anti-nuclear hippy groups and laws they got passed in the 60s and 70s?
One part this, two parts the economics of a novel technology platform being deployed in a large size, three parts American labor costs and inexperience with megaprojects.
Similar to why we can't build ships [1]: high input costs, notably materials and labour, and a coddled industry that is internationally uncompetitive. With ships, it's the Jones Act and shipyard protectionism; with civilian nukes, it's misguided greenies. (Would note that we're perfectly capable of nuclear production if it happens under the military.)
Nuclear is still much more expensive than renewables in China, where there aren't too many "misguided greenies" setting policy. Environmentalists were successful in opposing nuclear construction because it was expensive and unprofitable, not the other way around.
The faster people can internalize this lesson, the sooner we'll get to economically-viable nuclear power.
> Environmentalists were successful in opposing nuclear construction because it was expensive and unprofitable
As far as Europe is concerned, there seems to have been various political move and lobbying to affect energy independence (e.g. France): economy is transformed energy, so by nuking (…) energy independence, you're suffocating countries. The military role of nuclear is furthermore crucial; civil & nuclear must be correlated.
That's to say, giving up nuclear is not something a sane, well-driven country should do lightly, regardless of ideologies.
It's a tricky topic; what I regularly hear from economists is that wind & solar are still far from being able to compete with nuclear. And because of the previous two points, people can't but frown upon "green" arguments, even if the underlying intentions are honest and well-intended.
(China may not have misguided greenies, but it has a strong incentive to sell whatever it's offering).
If China had a super cheap nuclear design they would be very happy to export that the same way they export their other technologies like EVs, high speed trains, solar panels, batteries, etc. But it simply does not exist.
The first article refers to 2018 in the future tense, and the second article is three years old without a single announcement of a Belt and Road nuclear plant since then.
china has a super cheap design called hualong and they plan to export it the way russia is exporting their designs. Another plan is finishing local adaptations of ap1000 that can be reselled without licensing problems
> > why China installed 217 GW of solar last year, but only 1.2 GW of nuclear
> > And 114 GW of coal [1]. Don't do nuclear, and that becomes 115 GW of coal. Nuclear and renewables aren't competing for market share.
That is true for China, since their overall energy demand is growing massively. But is that also true for other parts of the world like the US or EU? Because looking at the electricity production [1] this doesn't seem to be the case. So in those markets they would compete for replacing existing fossil power plants. I think we can expect some growth, but not on a level even close to China.
We should see some increase in electricity consumption due to displacement of direct uses of fossil fuels. For example, use of heat pumps in place of natural gas furnaces, electric cars in place of IC engine vehicles. Add to that the ever popular AI and general data center consumption motivating this announcement (but I wonder how much of that is going to move to places with cheaper electricity.)
Yes, these and other innovations will defivinitely increase our overall electricity consumption, but i imagine that it will be a gradual shift as it is aleady happening, since vehicles and heating has long life cycles. It also helps that energy wise these technologies are more efficient, so that offsets some of the increase.
Probably hard to judge right now where AI is heading and if the pace of increased energy consumption remains this high. But i agree that they'll probably end up moving closer to sources of cheap electricity.
The efficiency would decrease primary energy use, if the electricity were being produced from thermal sources, but the amount of electrical energy used would increase.
I'm only slightly exaggerating when I say that the rest of the world is a footnote to China's emissions. Europe's emissions are already dropping fast, though. Presumably if China can decarbonize its economy at the rate it's going, then we presumably the rest of the world (even poorer nations) will be able to fast-follow them due to the learning curve (or else just because China will have so much excess manufacturing capacity that they'll flood the world with cheap renewables.)
The current analysis is that China's emissions peak this year [1,2] and will enter a structural decline. This is because new renewables are being deployed faster than growth in energy demand. The new coal construction is mostly "dispatchable" production that will be used to backstop the fast-growing renewable grid, with payments going to coal plants in exchange for not generating (and built-in expectations that these payments will rise over the next few years as renewables and storage serve more of the demand.)
is it? New plants cost 3-3.5bn for a stable 1gw output. For renewables - much more needs to be built to provide same reliability or compensate with fossils
IMO they only continue to exist because of the Jones Act not the way I think you're implying where Jones Act protectionism prevents them from flourishing. High material and labor alone are enough to explain why people wouldn't build ships in the US. What special capabilities could Us shipbuilders bring that would make the cost of labor here competitive with China or South Korea? Gone are the days when the US dominates on skill or capacity, and that's not because the US has lost something the rest of the world just caught up with us.
Whenever we're looking at the 1900s and wondering why the US used to be so dominant as an industrial power I think it's incredibly important to remember our industry got all the upside (an absolute torrent of money and demand) and none of the downside (bombing) of two world wars. IMO the US industrial base was riding high on that easily into the 80s and people mistake that dominance for skill and prowess rather than the waning boon of WW2's mobilization and destruction of every other extant industrial power.
The point is there are downstream costs to our moribund shipping industry. We have a internally-navigable waterways we barely use, offshore wind power gets stalled due to lack of ships, et cetera.
Post-WWII effects are one component. But another is that we want a protected shipbuilding industry for its own purposes, which is fine, but that curtails a lot of other production.
> What special capabilities could Us shipbuilders bring that would make the cost of labor here competitive with China or South Korea?
Energy. Our energy costs are much lower than theirs.
There is a huge difference between the US accounting for 20% of Global GDP and merely being "in first place" at the end of WWI and the USA having half of global GDP (and 80% of the world's hard currency reserves) at the end of WW2. While also say, having a Navy easily more powerful than the rest of the world combined, and being able to to focus on an upcoming surge in consumer consumption as opposed to desperately struggling to stabilize food production and rebuild cities and industries that had been ravaged by war.
Britain, a victor that had never been occupied, wasn't able to lift many significant food rationing schemes until the 1950s. Bread, which wasn't rationed during the war, had to be rationed from '46 to '48.
There is a meaningful distinction between being the leading industrial power and being the overwhelmingly dominant economic power.
When the German soldiers first encountered the US doughboys, they were struck by their height, their excellent food, and their supplies. That was when Germany knew they had lost WW1.
And this was despite having to ship all that stuff across an ocean.
France, with all their nuclear base, just raised their estimate for new reactors (I'm so shocked!):
> State-owned Electricite de France SA has raised its estimate for the future construction costs of six new atomic reactors in France by 30% to €67.4 billion ($73 billion)
6 reactors, 1650MW each, $7B per 1GW vs Vogtle's $17B. Planned. In 2 decades, after it's finally built, it will have doubled of course lmao.
That right, blame the hippies. Nothing at all to do with nuclear power plants being the one thing that you really do want to be engineered well. But no, regulations are of course to blame!
The anti-nuclear hippy movements of the 60s and 70s are pretty directly responsible for a lot of the slow down in expansion of nuclear power.
>Between 1975 and 1980, a total of 63 nuclear units were canceled in the United States. Anti-nuclear activities were among the reasons, but the primary motivations were the overestimation of future demand for electricity and steadily increasing capital costs, which made the economics of new plants unfavorable.
because 1 - 1gw of solar capacity isn't the same as nuclear, even 3gw of solar isn't the same as 1gw of nuclear (to get a proper perspective, look at germany's grid yesterday& how much overcapacity of solar/wind they have and how much was actually generated/imported). 2 - vogtle unit 4 was 30% cheaper than unit 3, proving positive learning curve, meaning (in theory, according to https://liftoff.energy.gov/advanced-nuclear/ ) new builds should be significantly cheaper
edit: to be clear, 1GW of wind or solar is $1B. Build 3GW for overcapacity and you’re still at just 17% of the cost of 1GW of nuclear, and you technically have 3x more capacity. Now figure out how many megapacks you can buy for the $14B/GW you saved https://www.tesla.com/megapack/design (answer: 16GW/68GWh)