In 1989, pro-nuclear lobbyists claimed that wind power couldn’t even provide 1% of Germany’s electricity. A few years later, pro-nuclear lobbyists ran ads in German newspapers, claiming that renewables wouldn’t be able to meet 4% of German electricity demand.
After the renewable energy revolution took off, in 2015, the pro-nuclear power “Breakthrough Institute” published an article claiming solar would be limited to 10–20% and wind to 25–35% of a power system’s electricity.
In 2017, German (pro-nuclear power) economist Hans-Werner Sinn tweeted that more than 50% wind and solar would hardly be possible. And in 2018, Carnegie Science reported a study claiming that “wind and solar could meet most but not all U.S. electricity needs.” According to one of the authors, their research indicates that “huge amounts of storage” or natural gas would need to supplement solar and wind power.
From a pro-renewable perspective, this is encouraging. The claims about the limits of renewable energy have moved from “not even 1% of electricity” to “most but not all of the electricity.” And yet, the anti-renewables message has always been the same: renewables will lead to a dead end.
In order to underscore their point, anti-renewable energy propagandists now publish incorrect cost figures that claim a fully renewable electric grid would be unaffordable or way more expensive than other options, such as, you guessed it, nuclear power.
MIT Technology Review writes about the “scary price tag” that such a purely renewable grid would come with, calculating $2.5 trillion as a price tag for storage requirements alone — 12 hours of storage. Wood McKenzie also talks about $2.5 trillion, albeit for 24 hours of storage. The “Clean Air task force” puts the cost for a 100% renewable grid in California at an annual $350 billion.
Anti-renewable propagandists need to talk about imaginary high costs of renewables, especially because one of their preferred ways of generating electricity — nuclear power — turns out to be incredibly expensive.
Renewable energy gets cheaper each year, nuclear power gets more expensive each year — how come they still adamantly claim that renewables are not a cost-effective way of decarbonizing?
The answer, of course, is that the studies are flawed. Taking a look at these studies shows that several patterns can be observed in many of these studies. Among these flaws are ridiculous overestimates of storage requirements, overestimates of grid expansion needs, and the insistence on uneconomical strategies of storing electricity, such as insisting on batteries to store several weeks worth of grid electricity consumption.
In order to understand how these studies are flawed, it’s essential to understand how a renewable energy grid actually works, how energy storage works, and what costs you can expect. After that, I will describe the flaws in some of these studies and recalculate a more realistic scenario, especially more realistic cost projections.
How a renewable grid works
A few facts are important to know:
Storage will not be necessary for a long time.
The sun doesn’t always shine, the wind doesn’t always blow — yet most of the time, there is either sun or wind available. For now, storage will not play a role for a long time. Solar and wind power will increase their shares of electricity consumption, and until they reach 80% of electricity consumption, grid expansion, moderate curtailment, and gas-fired backup power plants are the only tools necessary to reach such a high share of renewables.
Backup power plants are cheap.
So, if 80% of the electricity is generated using solar and wind power, the remaining 20% has to be created from backup power plants. According to grid operator PJM’s data, backup power plants cost up to $120,200 per megawatt per year. We can calculate the cost for a worst case scenario: To cover the 769 gigawatts of US peak load, backup power plants would cost $92.5 billion per year. Divided by the 4.18 trillion kilowatt-hours that were consumed in the USA in 2018, that amounts to 2.2 cents per kilowatt-hour.
Nuclear power is expensive and gets more expensive over time.
The newest Lazard figures put nuclear power at 15 cents per kilowatt-hour. In addition, that’s more than the cost figures of the previous years.
Even for 80 percent solar and wind, grid investment costs are moderate.
The NREL estimates that, even if you get 77% of electricity from solar and wind power, the grid will have to be expanded from around 85,000 gigawatt-miles to around 116,000 gigawatt-miles. That’s not even a 50% increase.
Getting more solar and wind power will require overbuilding and curtailment.
One study that is often cited as “proof” of the limits to renewables finds that, actually, even without any storage, overbuilding solar and wind to 1.5 times US consumption could get you 93% solar and wind power in the grid. This is still without any storage at all. To put this into perspective, if you overbuild solar and wind power 1.5 times, and you have an LCOE of 3 cents per kWh (according to BNEF, this is possible for solar and wind by 2030), that gives you a total LCOE of 4.5 cents per kWh (ignoring minor system costs for curtailment), which is still very cheap, and far below the 15 cents per kWh figure for nuclear power.
The remaining 7% could be provided, for example, by burning synthetic methane that’s made from hydrogen and carbon dioxide.
You can make a synthetic gas that’s 100% compatible with the existing gas infrastructure. The process is known as power-to-gas. Electrolysis uses solar and wind electricity to split water into hydrogen and oxygen. In a second step, carbon dioxide, which can be captured from the air (direct air capture) is mixed with the hydrogen. This results in methane, which is 100% compatible with the existing gas grid and the gas-fired power plants. Once this methane is burned, it emits only as much carbon dioxide as was previously captured from the air. The cost for this methane is currently estimated at 20 euro-cents per kWh, but costs have come down in the past and will continue to come down. In Germany, there is already a facility that generates renewable methane and injects it into the gas grid.
There might be other storage options as well in the future.
To store the entire grid for many hours or even days, batteries are too expensive. Yet there are other options under investigation. Siemens is testing a simple concept of first converting the electricity into heat, storing the heat, and later using that heat to drive a steam turbine. Highview Power uses cold air to store electricity and use the expanding, reheating air to drive a turbine. Both companies already built a pilot storage plant.
Considering these facts, it is possible to make a calculation about how much a purely renewable grid would likely cost, using today’s technology and today’s prices. Whenever anyone claims way higher costs, we should grow suspicious immediately.
Calculating the cost for a purely renewable grid.
Assuming we used today’s technology, we can compare solar and wind power to nuclear power. According to Lazard, nuclear power costs 15 cent per kWh. Generating all of US electricity from nuclear power, therefore, would cost $615 billion per year. So, how much would a completely renewable grid cost — per year and per kilowatt-hour?
One way a renewable grid would work would include the following technologies
Expanding solar and wind power to reach 93 percent wind/solar.
Using the study “geophysical constraints on the reliability of wind and solar power,” getting to 93 percent solar and wind power would require generating 1.5 times US power demand. This means that you overbuild wind and solar and curtail some of the electricity to increase the amount of solar/wind power that can be used directly. You would have to generate 6300 TWh of renewable electricity, which at current costs (according to Lazard) would cost $271 billion per year.
Paying for backup power plants.
Backup power plants that could provide the entire grid with electricity would cost $92.5 billion per year, according to PJM data.
Expand the grid
NREL data suggests that you need +30 TW-miles to go to 80 percent renewables. Extrapolating, you would need +37.5 TW-miles for 100 percent. That’s around 60 TW-km — thus, around $60 billion grid investment. Calculating the grid investment cost per year, it would cost around $10 billion a year (WACC 10%, 10 year payment). This shows that grid expenditure is negligible.
Burning renewable methane in these backup power plants to reach 100 percent renewable electricity
Using the latest study by Ludwig Bölkow Systemtechnik, generating synthetic natural gas from hydrogen, using direct air capture for the carbon dioxide, 1 kWh of synthetic methane costs around 20 euro-cents per kWh when produced in Europe. In a 60 percent efficient CCGT power plant, 1 kWh would cost 33.33 euro-cents (37.15 US cents). Generating 7 percent of US electricity from renewable synthetic methane costs $110 billion.
Total cost, therefore, would amount to $483.5 billion per year. Divided by electricity consumption of 4100 TWh, the total cost would be 11.8 cents per kilowatt hour. This is already cheaper than Lazard’s estimate for nuclear power, which is currently at 15 cents per kilowatt-hour.
Let’s also stress that this will change. In 2030, according to BNEF, wind and solar power will already be below $30 per MWh. Synthetic methane will cost around 15 euro-cents per kWh, according to LBST. As such, you would annually spend $189 billion on wind/solar electricity, plus 27,86*294= $82bn on synthetic methane, $92.5 bn on CCGT power plants, and $10 bn on grid expansion leading to a total $9.1 cents per kWh. That’s way cheaper than nuclear power.
So, how come we keep reading that a fully renewable electricity grid would be astronomically expensive, especially from pro-nuclear lobbyists? If a quick and dirty calculation already shows that renewable electricity is already cheaper than nuclear power, how come numerous studies point to 100 percent renewable electricity being unaffordable?
Once you understand how a renewable grid works and how much it will likely cost, we can look at the strategies used to discredit renewable energy.
Let’s look at the studies.
One of the studies frequently quoted by MIT Technology Review is the study “Geophysical constraints on the reliability of solar and wind power in the United States.” It’s available on the Internet for free, and a seemingly serious attempt to calculate scenarios of reaching 100 percent renewable electricity. Using 36 years of weather data and comparing it to US electricity demand, the study finds that:
- 80 percent of the US electricity could be provided by wind and solar power if either
- 12 hours of storage were installed or
- there were a continental-scale transmission grid.
To achieve 100 percent solar/wind power, either “several weeks worth of electricity storage” and/or “the installation of much more capacity of solar and wind power than is routinely necessary to meet peak demand” would be required. The availability of “relatively low cost, dispatchable, low CO2 emission power” would obviate the need for extra solar/wind and/or energy storage.
So far, that’s nothing new.
This study, however, goes on by calculating the cost of various scenarios of going to 100 percent renewable energy. However, none of the scenarios considered is even remotely as economical and/or realistic as a solar/wind/backup power plants/power-to-gas scenario. Instead, the study only considers 3 options, which are:
- overbuilding (no storage)
- pumped hydro storage
- battery storage.
There is no precise data on the annual costs for these options, yet it is mentioned that the costs would be $2.7 trillion, the assumed battery life would be 10 years, and the assumed discount rate would be 10 percent — which implies annual costs of $440 billion.
No reason is given why power-to-gas would be completely ignored, at a time in which it was already considered a required future technology to reach 100 percent renewables in Germany. Even precise cost data was already published in Germany (Potenzialatlas Power to Gas). Compared to today, power-to-gas was significantly more expensive at the time the study was published (and so were backup power plants to burn that gas), yet the total costs of storing electricity would have been significantly cheaper.
To get 93 percent solar/wind without storage, generating 1.5 times demand (6000 TWh) would be necessary — at that time around $270 billion. Power-to-gas (synthetic methane) to cover for the remaining 7% would have cost $185 billion, gas-fired power plants would have cost $150 billion. Total cost would have been around $600 billion. This is roughly on par with what nuclear power costs today.
To use batteries, $430 billion would have been necessary for storage alone, in addition, you would have had to generate 8000 TWhs of electricity, leading to a cost of $790 billion. This is equivalent to almost 20 cents per kilowatt-hour in cost.
Therefore, that study calculates a scenario which generates around $190 billion a year in unnecessary costs. In addition, that scenario today is outdated. As already calculated, today’s technology would lead to an annual cost of $483.5 billion. The Caldeira study, therefore, calculates a scenario that is $300 billion per year too expensive. The study is outdated, assumes the use of inadequate technology and therefore shouldn’t be of any relevance any more.
The Clean Air Task Force Study for California
In case you thought a study like the Caldeira study was highly misleading, you haven’t seen the CATF study for California. As expected, this study was reported on by MIT Technology Review as well.
The study assumes that for 100 percent renewable electricity, California alone would have to pay an annual $350 billion for storage alone. This is akin to $1.6 per kilowatt hour. As expected, that study is complete nonsense, but how on earth are such insane figures even calculated and argued for?
The most likely explanation is that this study completely ignores the possibility of overbuilding and curtailment. This is especially problematic in California, because both wind and solar power plants produce less electricity in winter. The most obvious approach to address that problem would be to build enough wind and solar power plants to provide enough electricity in winter. In summer, excess electricity generation would have to be curtailed.
Instead of this obvious approach, it appears that the Clean Air Task Force assumes that California will build giant batteries that can store all excess electricity in summer to save it for winter. Such an approach is completely absurd, as is demonstrated by the price tag of $350 billion for California alone.
Using up-to-date figures we can estimate the actual cost for California. To reach 100 percent renewable energy using solar, wind, and power-to-gas, we can estimate a total cost of $42.4 billion a year. This is akin to an LCOE of 18.4 cents, using current technology. This is still rather expensive, but not much more expensive than nuclear power. Considering the rapid cost declines for solar and wind power, it can be assumed that solar, wind, and power-to-gas will turn out to be the more economical solution for California as well.
The Hans-Werner-Sinn study for Germany
A similar study was already published in Germany, again assuming one scenario in which curtailment was not allowed. So, again, you had to store huge amounts of electricity in summer to save it for winter — 16 TWh of storage altogether to reach 89 percent solar and wind power. The second scenario didn’t allow for storage at all, which made a massive overcapacity necessary. Therefore, 61 percent of wind and solar power would have to be curtailed to reach 89 percent solar and wind power. There already is a rebuttal to that study, published by Zerrahn, Schill, and Kemfert, that showed how a compromise (allowing for 22 percent curtailment) would reduce storage needs to 1 TWh, whereas allowing for 32 percent curtailment would furthermore reduce storage needs to 432 GWh.
The Wood MacKenzie Study
Wood MacKenzie published a white paper, Deep Decarbonization requires deep pockets, estimating capital investment costs of $4.5 trillion for decarbonization using wind, solar, and batteries alone.
The Wood MacKenzie assumptions are the following:
- 1,600 gigawatts of generation (wind and solar)
- 24 hours of lithium-ion battery storage
- 200,000 miles of new high-voltage transmission at overall $700 billion in cost.
Wood MacKenzie’s assumptions are partly in contradiction to the “geophysical constraints” study. It suggests increasing solar and wind power roughly 12.3 fold, which means that there would be no overbuilding at all.
There is little indication that this would suffice to get 100 percent of solar and wind, even if you had 24 hours of battery storage (unlike 12 hours as suggested by the Caldeira study). In fact, the supplementary data provided by Caldeira shows that increasing storage capacity from 12 hours to 24 hours would have little effect on the necessity to overbuild solar and wind power plants. Since battery storage is incredibly expensive, Wood MacKenzie suggests using:
- an inadequate storage strategy
- unnecessarily much storage
- likely too little solar and wind power to actually achieve 100 percent solar and wind.
Even less justifiable is the assumption that $700 billion would have to be invested in grid expansion. Based on NREL data, it’s likely that less than one tenth of that sum needs to be invested. Even the Caldeira study “only” talks about $410 billion of grid investment.
The Jenkins–Thernstrom commentary
Jenkins, former Director for Energy and Climate Policy in the Breakthrough Institute, published one study and one commentary in Joule Magazine, which of course found that a purely wind-solar-storage solution is not a good idea. Jenkins co-authored one study and one commentary on the future of electric grid decarbonization. The study was published in November 2018, the commentary in December 2018.
The commentary points out challenges on the path to a zero emissions grid. It correctly finds that the challenges increase as renewable penetration increases. It also correctly finds that grid expansion cost are negligible compared to other costs and that greening the electricity sector is vital to green the economy.
It correctly finds that there is a necessity to overbuild. However, it finds that between 40 and 50 percent of generated electricity would have to be curtailed and finds that this would almost double the costs of the entire electricity system. This is, of course, completely outdated, since electricity from solar and wind power have fallen drastically in costs.
The study specifically mentions a possible electricity consumption increase for electricity “and fuels produced from electricity, e.g. hydrogen,” to more than 50 percent of final energy demand.
However, oddly, the study completely ignores the possibility of using exactly these fuels to green the electric grid. Producing electrolytic hydrogen and converting it to methane is not considered, arguing that “considerable uncertainty remains about the real-world cost, timing, and scalability of these storage options.” This technology (power-to-gas), which significantly reduces the costs of greening the electric grid, is completely dismissed.
There is no clear definition of “considerable uncertainty,” and Jenkins, Luke, and Thernstrom don’t mention any specifics or any studies that point to that. In fact, in 2018, various German studies (such as the DENA e-fuels study) already were very specific about the cost (and also predicted a significant cost reduction). No reason is given why that data would be completely ignored.
The commentary goes on arguing that several technologies (grid expansion, flexible demand, seasonal storage, and very-low-cost wind and solar) must all become reality, whereas other technologies such as nuclear power, CCS and enhanced geothermal energy could all fill the firm role in a low-cost, low carbon portfolio. Therefore, the commentary argues, the chances of wind, solar and storage providing 100 percent of electricity consumption are lower than the chances of wind, solar plus nuclear, CCS, or geothermal energy.
This logic has a severe flaw. First of all, very-low-cost wind and solar are very likely to become reality and partly already are reality. Just because several conditions have to be met in one scenario doesn’t mean that this scenario is less likely to work out. Jenkins writes about nuclear power, CCS, bioenergy, and enhanced geothermal energy: “Assume that each resource has only a 50 percent probability of becoming affordable and scalable within the next two decades. If all four options are pursued, however, the odds that at least one succeeds would be 94 percent.”
But you cannot do that. You cannot simply assume a certain chance. Jenkins says that these examples are “purely illustrative,” but still goes on arguing that we shouldn’t eschew the development of firm low-carbon technologies because they face challenges today.
But that’s not how it works.
To make wind and solar power cheap, to make batteries cheap, hundreds of billions of dollars had to be invested. We don’t have an infinite amount of money and an infinite amount of time. Should we invest hundreds of billions of dollars in nuclear power, CCS, and geothermal each? This is money that we couldn’t use for making wind and solar power and energy storage — all of which are proven and highly developed technologies —even cheaper. The more time and money we waste on technologies that face severe problems and are expensive, the less time and money we can use for solar, wind, and energy storage — technologies that actually work.
The Jenkins–Sepulveda–Sisternes–Lester study
Again, this study points out a barely new “finding” that a grid that merely consists of batteries, solar, and wind power is likely going to cost more than other alternatives. This is well known. This is exactly why there is investment in power-to-gas and other long-term storage technologies — for example, thermal energy storage.
Of course, again, power-to-gas is ignored entirely, therefore leaving wind-solar and storage with the only storage option of lithium-ion batteries.
What’s more worrisome about this study is the fact that the authors “propose a new taxonomy that divides low-carbon electricity technologies into three different sub-categories: ‘Fuel-saving’ variable renewables (such as solar and wind), ‘Fast burst’ balancing renewables (such as lithium-ion batteries), and ‘firm’ low carbon resources such as nuclear power plants and carbon capture and storage (CCS) power plants.”
This is a very dangerous taxonomy. If we start using it, we implicitly rule out that solar, wind, and some sort of energy storage can power the grid alone. Solar and wind power will always merely be considered an add-on to a grid that is essentially powered by some other resource.
Of course, power-to-gas could be considered a “firm” energy source. However, there is a significant difference between carbon capture and storage (CSS) and nuclear power: capital costs. Equipping a gas-fired power plant with carbon capture features would double the capital costs, which reduces its economical prospects if it isn’t used frequently. Nuclear power is even more capital-intensive and would have to be used frequently as well.
This is also confirmed by what the authors envision: What’s officially named “mid-range scenario” (presumably the most likely outcome, according to the authors) not only indicates that nuclear power will be the most important electricity source — providing around 50 percent of all electricity in the “Southern System” and around 80 percent electricity in the “Northern System.” Jenkins basically did it again: Limit wind and solar power to a maximum of around 50 percent and declare that the most important electricity source in the future will be — you guessed it — nuclear power.
However, looking at the study, you will immediately find significant flaws.
The first obvious flaw, of course, is that power-to-gas is completely ignored. This was expected.
A little less expected are the assumptions for technology costs.
For example, the mid-range costs for solar power are considered to be $900 per kilowatt. This is based on the NREL data for 2017, applying 50 percent cost reduction. In the “Very Low” scenario, solar is assumed to cost $670 per kilowatt — based on the NREL’s estimates for 2047 (Utility PV — Low).
As for wind, mid-range costs are considered 25 percent under the NREL’s “low” assumption for 2017 wind power. “Very low” wind power costs are assumed to be $927 per kilowatt — based on NREL’s estimates for 2047 wind power — (Land Base Wind, TRG 1 — Low).
At the same time, the “Conservative” assumption for nuclear power is $7,000 per kilowatt, based on Georgia Public Service Commission (PSA).
$670 per kW for solar in 2047 are likely way too pessimistic. DNV-GL, for example, now estimates that solar PV would be at 42–58 US cents per watt in 2050. The most optimistic “very low” scenario for solar, therefore, should be at $420/kW, not $670/kW. Wind energy forecasts are more conservative. Thus, wind energy projections made by Jenkins might be correct.
But taking a look at Jenkins’ envisioned grid supply, in most high-renewable scenarios, the largest part of the renewable electricity is provided by solar power anyway. Thus, underestimating the reduction of solar energy costs means to decisively overestimate total costs of a renewable energy grid.
As for nuclear power, Jenkins’ most pessimistic assumption is that nuclear power costs $7,000 per kilowatt. That is actually overly optimistic. Lazard currently estimates that nuclear power costs between $6,500 and $12,250 per kilowatt. In 2016, estimates were at $5,400–8,200 for nuclear ($8,650 for new US nuclear). This means that nuclear power actually got more expensive. Jenkins doesn’t merely assume that nuclear will reverse this trend someday, but even in his most pessimistic scenario have capital costs that would be considered at the low end of the spectrum today.
To sum it up, Jenkins makes overly optimistic cost assumptions even for his “conservative” scenario regarding nuclear. And he makes overly pessimistic assumptions even for his “low” scenario regarding solar power. So he basically compares an optimistic projection of nuclear power costs to a pessimistic projection of solar power costs and finds that nuclear power is cheaper.
Now that we have looked into some anti-renewable energy propaganda studies, we can spot a set of strategies that is used by anti-renewable propagandists to discredit renewable energy.
Ignore power-to-gas
Even pro-nuclear propagandists are very well aware of the ability to store large amounts of electricity using power-to-gas — they simply ignore it. You find Jenkins, Thernstrom, and Sepulveda mentioning that technology, but then simply go on by only calculating the costs of other, less optimal storage technologies. Sepulveda doesn’t give a reason at all for ignoring power-to-gas, Jenkins and Thernstrom dismiss scenarios that rely on power-to-gas, arguing that it “remains unproven at such large scales,” without explaining why power-to-gas, even though it is proven to work, all of a sudden would stop working if a large number of power-to-gas facilities were built.
Insisting on an inadequate storage strategy to store large amounts of energy, such as insisting on lithium-ion batteries for that task is one way to artificially inflate the costs of going renewable.
Overestimate storage needs
The study Geophysical limits talks about 12 hours of lithium-ion or pumped hydro storage needs for the USA. Wood McKenzie all of a sudden estimates 24 hours of lithium-ion storage needs for the USA, Hans-Werner Sinn estimates 16 TWh of pumped-hydro storage (more than 10 days worth of storage) for Germany, and the Clean Air Task force estimates 36.3 TWh of lithium-ion battery storage needs for California, around 46 days worth of energy storage. While the storage estimates for 12 hours of lithium-ion battery storage are already hard to justify (as there is power-to-gas as an alternative), it is quite obvious that arguing that pumped hydro or lithium-ion storage must store more than a week’s worth of electricity consumption is nonsense and designed to artificially inflate the cost estimates of a 100 percent renewable grid. This works by using the next strategy:
Ignore curtailment
The Clean Air Task Force and Hans-Werner Sinn used the strategy of simply not allowing any curtailment of renewable energy at all. This, of course, inflates the cost of storage enormously. If you do allow curtailment, you can build more wind and solar power plants than usually needed — so you have enough solar and wind power even in times of less wind or sunshine, therefore reducing storage needs. For example, to get to 90 percent solar/wind power in Germany without curtailment, you would need more than 16 TWh of storage. If you accept around 22 percent curtailment, storage needs are reduced from more than 16 TWh to 1.1 TWh.
Overestimate grid expansion needs
Another way of artificially inflating cost estimates for renewable energy is to vastly overestimate the needs of grid expansion. The NREL’s estimate is a grid expansion from 85,000 gigawatt-miles to around 116,000 gigawatt-miles for 77 percent solar and wind power. So even if we calculate that for 100 percent solar and wind power, a further expansion to 125,000 gigawatt-miles might be necessary, the costs remain moderate. 1 mile is roughly 1.61 kilometers. At $1 million per gigawatt-kilometer, therefore, it would cost around $65 billion to expand the grid to 125,000 gigawatt-miles. This puts into perspective the vastly overblown grid expansion estimates by Sepulveda (252,000 gigawatt-miles or 408,000 gigawatt-kilometers at a cost of around $410 billion) and Wood McKenzie (200,000 miles of new HVT at a cost of around $700 billion).
Ignore or underestimate progress
A review of “recent literature” by Jenkins and Thernstrom in 2017 found that getting to near-zero emissions would cost significantly more than including technologies like nuclear power and CCS. One of the studies cited by Jenkins and Thernstrom is a study by Brick and Thernstrom from 2015. This study claims to “test the outer bounds of” future scenarios, assuming rapid and significant cost declines for wind and solar: Capital costs of $1000 per kilowatt and increased costs for nuclear ($6500 per kilowatt).
“In November 2018, however, Lazard considered $6500 per kilowatt the lowest end of the price spectrum for nuclear power, whereas the highest end of the spectrum was $12,250 per kilowatt. At the same time, wind and solar were estimated to cost between $950 and $1250 per kW (solar) and between $1150 and $1550 per kW (wind). Thus, what was considered “rapid and significant cost declines” in 2015, in 2018 was already within reach.
Source: Georg Nitsche /CleanTechnica