Nuclear in Norway – necessary but in need of political vision and a fact-based debate

Key takeaways from this short review:

1.?????? Nuclear energy is necessary due to the scale of the energy transition and also cost efficient for Norway, and indeed in general, if projects are reasonably well executed.

2.?????? The debate on nuclear energy is also about energy security and -resilience.

3.?????? Nuclear is the most suitable energy source going forward concerning all sustainability targets.

4.?????? Many Small Modular Reactors (SMR) designs will surely fail, but some will provide the basis for an industrialized rollout of nuclear power that will both be large in scale and competitive.

?John Kerry stated it well at COP 28 in Dubai: “One cannot achieve the goal of net zero greenhouse gas emissions by 2050 without nuclear power. This has nothing to do with politics and ideology. It is pure science, mathematics and physics”.

The debate about the future of nuclear energy in Norway has been going on for some years, and recently Rystad Energy released a report that basically said that Norway should wait until 2050 before using nuclear power. This report is one of the most comprehensive and thorough reports on the subject published. It therefore deserves a critical review.

The Rystad Energy report did offer valid perspectives and some valid content, but it failed at capturing the greater energy picture for Norway. Furthermore, its conclusion is inconsistent with the rest of the content. Indeed, the report appears at times, as it is written by different people who do not quite agree.

1 Lack of understanding of scale

The Rystad Energy report takes as given the report of the Norwegian Energy Commission the11th of February 2022 that estimates Norway will need an additional 40 TWh/year electricity supply in addition to the roughly 150 TWh produced annually today. Taking this assessment is obviously convenient, but ultimately problematic because it does not make sense given the climate goals Norway has committed to, its population growth and the fact that new industry must be established when the oil- and gas industry scales down.

For starters, just the Norwegian domestic shipping fleet will require 50 TWh/year for decarbonization using today’s energy policy means of green ammonia or similar. Add 20 TWh/year for the population growth and then aviation, trucking, industrial heat as well as replacing oil and gas. If we factor in these variables, we talk about a doubling of the power system capacity over the next generation.

This lack of understanding of scale is a major political problem in Norway, because many essentially argue that a few wind turbines here and there, some solar panels, a little bit of energy savings and other limited endeavors will bring us there… Such approaches will bring us nowhere.

Norway has over the last 20 years managed to increase its wind production to roughly 10 TWh/year… peanuts for the task at hand. Critically, similarities in the weather systems over the larger Nort Sea area give over production in one hour or lack of production in another hour, which means that a large share of balancing/back-up power is necessary in all the countries around the North Sea at any given time, as shown next. The Omega high pressures make the situation all the worse because they tend to last weeks and cover large areas.

The result can be readily seen in Figure 1 where we see a German Dunkelflaute in December 2022 lasting about 16 days. Keep in mind that Germany has more than 120 GW solar- and wind power, and with the poor results shown in Figure 1 we understand the futility. With an hourly shortfall between demand and all the renewable energy sources in Germany of about 30 GW at any given time, the aggregated shortfall is about 12 TWh. It would take more than 3800 batteries identical to today’s largest grid battery of 3000 MWh (Moss Landing Energy Storage Facility in California) – to handle this Dunkelflaute. Using the Lithium-ion battery price worldwide from 2013 to 2023 we see that the price may converge towards a 100 USD/kWh. The cost of handling the Dunkelflaute in Figure 1 will therefore become approximately 1200 bn dollars, or enough to construct 40 Vogtle nuclear power plants with a nameplate capacity of more than 2200 MW and an annual production of 17 TWh. Worse, while Vogtle has a design life-span of 65 years, such batteries have 15 years at best. Also keep in mind, that in Germany only 25% is non-fossil energy sources, which means that further electrification without using nuclear power will double or triple the challenge just discussed. Basically, using batteries is utterly unrealistic unless we are talking about several orders of magnitude improvement over today lithium-ion batteries.

Many believe that Norwegian hydropower can do the job – also wrong. Norwegian hydropower was designed primarily for baseload operations and relatively slow variations with intermittent changes handled by either the grid reserve or import from Sweden and Danmark (before the cables to Germany and the UK). The NordLink to Germany and the North Sea Link (NSL) to England both came into partial service in 2021 with NSL coming into full service of 1.4 GW first towards the end of Q1 2022. To follow the erratic production of VRE (Variable Renewable Energy) sources, in particular wind, will bring a lot of maintenance and downtime for hydropower, which is not accounted for in most reports including the one from Rystad.

2 Why wind power cannot power the energy transition

In Figure 2, we see the wind power production in Norway in 2020 – ca 10 TWh produced. There are several observations which prove that the current energy policy is based on flawed assumptions. First, there is a weak positive correlation (ca 30%) between wind and hydropower in contrast to the claim that hydropower act as backup power to wind power plants. The weak correlation is in fact due to the seasonal variations whereby we have most wind in the winter and also the highest usage of hydropower due to lower temperatures and higher demand. The fact that we consume most energy during winter is to some extent an argument for the wind power, but not when we examine the details.

Second, the wind power production is highly variable – almost random. The vertical axis in Figure 2 shows the number of MWhs produced per hour during the year in question. The horizontal axis covers the entire year starting with 00:00 on 1st of January to the left, ending at 24:00 the 31st of December for both years. The data behind Figure 2, reveal that in 2020 the hourly production varied from merely 8.3 MWh per hour to 3,275.6 MWh per hour.

By sorting the data from the highest hourly production to the left and the lowest to the right, the graphs in Figure 3 are obtained. The upper figure shows clearly that when we organize the graph according to the descending demand data (in dark blue/black), it is hydropower that fulfills the demand. The demand data are removed in the lower figure, and the graph is organized according to descending hydropower. Then, we see that net import is also highly driven by the hydropower because when Sweden, Denmark, the Netherland, Germany and/or the UK have surplus power due to high wind production, and therefore lower prices than Norway, then Norway uses its interconnectors to import this low priced or even negatively priced electric power. To accommodate imports when imports are not needed to balance supply and demand, Norway reduces the production of dispatchable hydropower. Norwegian wind power is more or less ‘random noise’ at the bottom of both graphs.

The wind power production data are analyzed in greater details in Figure 4. As we see, the increase in capacity from 2019 to 2020 has some effect on the maximum production to the left, but the problem with the low hourly production (within the dotted box) to the right is more or less the same. The reason for this is the weather covariation/correlation over large areas.

When there is wind, most wind power plants produce well, but when the wind is lacking, most plants stand still. For the NO2 price area in the south of Norway, the average wind power correlation is 66%. In the slightly more geographically dispersed price area of NO3 in the middle of Norway, the average wind power correlation is 52%. In the geographically most diverse area, NO4 being Northern Norway, the average wind power correlation is 28%. Clearly, a large and long geographical area helps, but none of them are remotely large enough to overcome the wind power covariation/correlation issue.

To illustrate, review Figure 5 where all the wind power production in European Economic Area is assembled, ignoring bottlenecks and other practicalities. Despite 26% increase in capacity over the time period, there is always a major problem with lack of wind power production. Even with completely uncorrelated weather system, which is the ideal situation, at least four wind power plants are required to guarantee the output of one. This is the best case, and the higher the correlation the worse it gets.

Indeed, in 26 OECD countries the most commonly used balancing/backup power for VRE is gas power, which increases the climate gas emissions from VREs to the extent that they can never secure a low-carbon grid. Thus, the VREs fail at the very core of their purpose – to cut greenhouse gas emissions to a sustainable level. VREs only appear useful when considered in isolation at asset-level because system issues are ignored.

From this simple analysis, it becomes evident that two large power cables to the UK and Germany would essentially make the southern Norway (NO2) a part of the European grid for good and bad. The impact on prices should also be no surprise. Earlier studies of Europe are very clear; “…1% increase in the exporter’s electricity price reduces energy trade by an average of 0.7%, whereas a similar increase in the importer’s electricity price increases energy trade by an average of almost 1%”.

Given this short background on wind power in Norway, Rystad should have pointed out a number of facts that should have questioned the very starting point of their analysis. At least they should have reached a conclusion that nuclear must become part of the power mix in Norway, if we are going to have any chance of reaching the double-transition aimed at decarbonization and building new industry to replace the oil- and gas industry while on top handling the population growth. All referenced analyses here show that doubling the energy system in any realistic way without nuclear power is basically unattainable. The Rystad report should have pointed this fact out.

3 Why hydrogen heroics and batteries will not suffice

The standard answer to the discussion above, in some reports, is to produce hydrogen from excess VRE power, and then use the hydrogen either as chemical battery or use it in the transportation sector, replacing fossil fuels with green ammonia and the like. Luckily, the Rystad report did not venture into this derailed discussion.

Let us start with the simplest case – batteries. Batteries can surely improve grid performance through peak-shaving, but it stops there as shown earlier concerning Germany and Norway, except for minor consumers, such as passenger cars. The large ships in the worlds shipping fleet, consume about 3000 MWh daily, which is the same size as the current largest grid battery in the world. Therefore, the number of batteries required to have a meaningful impact is simply too large for practical purposes. Moreover, the entire material resource base is too small for a green transition as many envision, largely based on batteries, wind- and solar power. IEA argues we must increase mining by multiples of up to 20-40 times to meet the climate objectives! At the same time, we have major under-investments in mining and gas (as discussed later) despite that fact that both are necessary for VREs to exist as intended due to their balancing requirements.

The case for hydrogen production on a large scale using VREs is equally futile. Sure, hydrogen can be used for smaller applications, but not on a scale that matters for the big picture. The reason is the energy and efficiency losses in the process. The production of hydrogen requires about 50 MWh/tonne during electrolysis, which means that the Norwegian production of hydrogen from gas – 225,000 tonnes per year – will alone require 11 TWh/year. Global consumption of hydrogen was 94 million tonnes in 2021, which would translate into 4700 TWh/yr, or 1.5 times the EU power production in 2022, based on hydrogen as a precursor, which is the key energy yielding component for ammonia. Then, we have merely replaced the existing grey hydrogen with green hydrogen and not added a kilogram of new, fossil free hydrogen. Also, we have not discussed the fact that process industry needs 100% reliability, which effectively excludes VREs as a power source for hydrogen production on their own.

If we venture into new usage of hydrogen, the numbers are even more depressing. Consider shipping as an example of the challenge. Maritime industry constitutes only 3% of global emissions, but it consumes about 300 million tonnes heavy fuel oil (HFO) per year (note that the numbers available are not consistent among different sources). To abate this sector’s hydrocarbon consumption using green ammonia, which is often suggested, will require 2.7 times the entire EU power production in 2022. The reason is that HFO has an intrinsic thermal energy of 11 MWh/tonne, while ammonia has only 5 MWh/tonne and it requires about 12 MWh/tonne to be produced via electrolysis. Then add the hydrogen needed for decarbonizing aviation, trucking and all the high temperature processes in industry. Decarbonizing the world economy will basically never take place with the current energy policies.

Add to this the practicalities of using hydrogen. The tiny molecule leads to hydrogen induced stress in metals frequently causing catastrophic failure over time, which in combination with its high explosive potential – 1 kg under pressure is roughly equivalent to 1 kg TNT – makes it a very challenging gas to handle on grand scale unless it is transformed immediately to whatever end product we need such as synthetic fuels. The Space Shuttle Challenger accident in 1983 gives us a reminder of hydrogens explosive nature, and the same can be said about the second and largest explosion in Chernobyl as well as the one in Fukushima. Hydrogen is hard to handle!

4 Without nuclear there is no solution – also for Norway

Thus, we are left with the point that the question is not whether or not we want nuclear power, but only how much and by when. Otherwise, anything that looks like an energy transition will not materialize if we include all sustainability criteria. To obtain an idea about the amount, we have to assess the current energy usage and the forecasted population increase. The reason is that the correlation between energy and population growth globally between 1820 to 2020 is 99.6%! This is not likely to change much.

For Norway, the primary energy consumed is shown in Figure 6 excluding offshore installations. In 2022 the total primary energy consumption calculated using the substitution method was 527 TWh out of which 377 TWh was renewable energy sources, mostly hydro. As noted earlier in Section 1, the increase in population in Norway will increase the annual electric primary energy consumption by 20 TWh by 2050, while the total annual primary energy usage will increase by 28 TWh given the same mix as today.

Globally, the numbers are far worse as the global share of electricity is 15% of final energy consumed. Thus, for most countries the energy transition itself will be more demanding than in Norway, even though Norway awaits a double transition. The Norwegian economy is also facing a slow decline of the oil- and gas output, which means that additional industry must be established in addition to the energy transition just mentioned. For Norway a need to double the power system should therefore be a reasonable estimate, as mentioned earlier. For many other countries, we are talking about much more.

A doubling of the Norwegian system implies about 150 TWh/year of additional power, or the equivalent of 14 South-Korean APR1400 reactors as built in UAE. Using the cost in the UAE, this is approximately a 100 bn dollar project. Ironically, the Rystad report documents the falling overnight costs of the four reactors built in the UAE but fails to take that into account. We can only speculate why.

This leads us to another big flaw in the Rystad report. It assumes that we have to do everything in Norway. This assumption is contrary to all historical evidence. We did not do it when we discovered oil and gas around 1970. We do not do it in the marine industry. The most relevant historical evidence that follows the argument along the lines of Rystad, is the failed transition of the sail ship period to the steam ship period where Norwegian shipping companies suffered huge losses, and many went bankrupt due to strong opposition to the inevitable.

5 Why wait to do what is “right”?

Within this context, the advice from Rystad to wait until 2035 before starting possible planning for nuclear energy in Norway, and subsequently for possible implementation towards 2050 and beyond, becomes even stranger. The sheer scale of the transition is so large that we need the time, and if it makes business sense in 2050; why wait? Here, it seems likely that Rystad falls prey to another narrative in media and politics – the 2030/2050 targets which holds that by reaching certain goals on these years, we can stave off climate change. However, what is important is to make the right choice. Add to this the fact that to execute a legal licensing process in Norway concerning any power plant takes at least seven years. Hence, 2030 is illusory by any standards, and Rystad should know this. An ironic element to these dates and benchmarks is the recent statements by the US special envoy John Kerry in front of heads of state at the Dubay COP28 climate conference referred to at the opening of this review.

If nuclear suddenly becomes right in 2050, as Rystad acknowledges, then the logical conclusion would be that it is better to start early than late and make any amendments to smooth the transition. After all, one of their referenced reports highlights the importance of “as much as possible, as quickly as possible”. Since, Norway has failed to foresee the situation we are in now – energy shortages in few years and very high energy costs – the rational action is not to proceed, but to revise, regroup and find a new approach that works. For example, gas power plants would make an excellent transition power, as we witness many places in the world.

However, here, Rystad once again falls prey to the politics of Norway, but politics should be left to the politicians and analytical work should maintain its integrity based on facts. Rystad must know that the usage of gas power in the transition will make sense. So, is it then better to make irreparable damages to nature by expanding wind power and hydropower to gain a few years of transition time without making any significant steps towards the transition? Obviously not. The energy transition is not only about climate gases, but of equal importance also about biological diversity, nature and more.

With these issues concerning the background of the report, it is difficult to escape the conclusion that the report is of meager quality and coming from Rystad – it is even more difficult to escape the sentiment that the answers were given before the analysis was done. However, there are a significant number of issues on the technical levels of the report as discussed in the remainder of this note.

6 Failing to understand the nature of Small Modular Reactors (SMR).

Many opponents to nuclear power fail to understand what SMRs are about. Ironically, Rystad cites a number of reports and they have also calculated the fall in overnight costs of the first APR 1400 in UAE to the last. They estimate 2.5 MUSD/MW for the fourth APR1400 unit. Yet, they fail to translate this insight into the realm of the SMRs. For example, some SMRs have up to 90% less system complexity than the pressurized water reactors of today. Do they seriously believe this will have no cost impact?

The large nuclear power plants can be described as bespoke, advanced handcraft leading to major efforts in engineering, licensing, key construction processes such as cutting and welding due to size, on-site construction, access to people and much more. SMRs, however, represent a totally different path through industrialization of essentially a product. This will also cut costs on engineering, licensing and essentially the entire supply chain.

I will use a case from my former career as General Manager of a manufacturer of advanced pressure vessels in Norway. A robotic cutting center that can take just about all kinds of exotic steel can cut in 120 minutes all the nozzles of a pressure vessel body that we had estimated to take 3 weeks, fulltime, for two operators. Not only that, the quality is better which makes the subsequent welding also easier. The effects of an industrial production versus handcraft are many more than those just mentioned where many will also have direct, time-saving effects.

The primary challenge in an industrial setting is not the unit cost alone, but also the cost of idle capacity. Hence, keeping an industrial plant occupied to achieve the intended unit costs requires volume. Rightsizing the supply chain for the SMRs will therefore be key.

Let us remind ourselves that in 1908 there were more than 250 car manufacturers in the US alone. 40 years later, it was five. It is likely that a similar shakeout will also take place in the SMR domain. In 1908, it was at best a niche industry until Henry Ford took it the next step through industrialization. Ford Model T cost initially 850 USD in 1908 but through industrialization it fell to 350 USD in 1916! It was said that Ford extracted the iron ore on Monday and delivered the car on Friday! This is the power of industrialization, and it illustrates that we must move nuclear industry from handcraft today to the first Model T nuclear power plant design over the next decade.

To ignore these effects for the SMRs is a major flaw of the Rystad report, and at best we can say that the report is suitable only for first-of-a-kind (FOAK) estimates for Generation III SMRs, but largely useless for the longer-term competitiveness of SMRs. It is the latter that counts for policy. One possible explanation to why Rystad used FOAK cost estimates may be that the consulting firm is basically unaware of the landscape of SMRs since they collectively discuss SMRs as if it is a homogenous whole.

I am heading perhaps the leading nuclear reactor project in Norway for the time being, and we have reviewed all the 90+ SMR concepts available as of year-end 2022. The top vendors we have selected are not even mentioned by Rystad. Two of them will most likely have commercial versions ready before 2030. Thus, most of what the Rystad report writes about SMRs is misleading, particularly concerning Generation IV which they portray as far into the future. Is Rystad unaware of the fact that China has already started a Gen IV for commercial operations?

Rystad should also expect that most of these concepts will not survive according to the history of innovation, as exemplified above by the US automotive industry. If just a handful of them survive, the industry will become a major force for change. Waiting until 2050 will then become a major strategic blunder possibly robbing Norwegian industry the chance of become a part of this industrial landscape unless Norwegian industry relocate such activities to outside Norway.

7 Real options are ignored

Another major flaw is that VREs are also consistently underestimated in terms of costs because the costs of balancing/providing backup to ensure delivery to customers is never included, while the upsides of nuclear are not even mentioned. Essentially, the options implicitly assumed in the Rystad report fail to capture the picture to sufficiently make an informed choice.

Interestingly, Rystad uses outdated numbers of wind power costs. It may be that they think that today’s turmoil in the wind turbine industry is just temporary, but when the entire ESG sector has double digit losses it is time to rethink. Reflecting this trouble, the capital value of almost all ESG projects within wind, solar, carbon capture and hydrogen have collectively within the last three years lost about two thirds of their value measured by western stock markets. The only ones that have made money are project initiators, backed by systemic subsidies, and certainly not the bulk of ordinary investors. It is time to rethink. This should be even more imperative given the failed wind power auctions in UK and Germany, the recent history of cancelled projects and many bankruptcies of wind power plants in Sweden.

The same is true concerning the material resource situation, as noted earlier. The mining industry faces serious underinvestments. The underinvestments within oil- and gas industry were 56% in 2021 compared to 2014. Then, how likely is it that energy costs will be reduced enough to help the turbine manufacturers regain competitiveness by cost improvements? One should notice that most of the input energy is fossil when making VRE technology since the process industry behind this technology requires 100% reliability. Hence, the likelihood of falling costs is close to zero. Thus, the solution is either price increases and/or relocation to cut costs structurally.

Indeed, the VRE industry is increasingly dominated by China by a 85-90% market share for a number of critical metals and minerals. This introduces a supply chain risk, and the risk for further prices increases is very large. Why should not the Chinese use their market power to improve their margins? Thus, the real options that Rystad draws up in their report is simply put like this; overexaggerated issues and hurdles concerning nuclear risks set up against underdeveloped risks of VREs.

Rystad also demonstrates a significant lack of knowledge about the supply of uranium. Nuclear is largely concrete and steel, but the fuel is an issue today since Russian is a major supplier. This is relatively easy to rectify, and it is being done so internationally and several countries have stated that securing the fuel supply is a priority. However, overall, the material requirements for nuclear power are very limited compared to the other energy sources as shown in Figure 7. The ‘other’ category is the critical- minerals and metals as mentioned before.

However, the ocean itself is the biggest deposit of uranium with more uranium constantly entering from a deep geological reactor in the core of Earth. There is enough uranium in the oceans to power all of humanity for billions of years using advanced reactors. Extracting the uranium from the ocean might sound like another fairytale, but technology is being developed at reasonable costs. It needs commercialization and perhaps also less available uranium on land to improve its competitive position. The access to fuels is therefore only a temporary issue, and with a competitive solution for extraction from the seawater in the oceans nuclear fuel will be the most available fuel on Earth. However, uranium is fairly well spread out on land today yielding less supply risks than a many of the other critical minerals and metals needed today for VRE.

Sure, some of the legacy costs of nuclear experiments and military activities in the past are large because decommissioning was not factored into the planning of the facilities, the documentation has been incomplete or even deteriorated to the extent it cannot be used, but this is not illustrative of modern, commercial nuclear power. It can best be seen as an expensive lesson from the past of what poor planning and poor availability of documentation can lead to.

We also have to look at the other side of the equation. Nuclear has given nuclear medicine, and much more. For example, Norway benefitted hugely from nuclear research as it gave critical support to the development of the multi-phase flow technology used in the oil- and gas industry, dubbed the most important innovation in Norwegian history by Aftenposten. Thus, it is easy to point fingers at some of the legacy issues of nuclear, but then we must also take the entire picture into account. Nobody, however, takes the entire picture into account when it comes to renewables – it is just assumed it will be fixed. Well, the whole VRE industry is met with nature costs and enhanced EU policies to protect nature. This has been a preeminent factor for onshore wind projects, and we are presently confronted with major environmental concerns for offshore wind as well. Alternative costs and environmental risks are discussed later in Sections 11 and 12, but they must also be seen in the context of real options.

When it comes to the upsides of nuclear power, it is not even contemplated apart from producing electricity. The fact that nuclear energy produces a lot of thermal rest energy that can be used, is ignored. For example, my latest analysis of the Melk?ya LNG terminal shows that nuclear can do the job at roughly half the cost compared to a wind power alternative. Then we have production of freshwater and district heating rendering up to 80% of the total energy output usable. If nuclear propulsion for merchant shipping becomes a reality, the business potential is enormous possibly dwarfing today’s nuclear power industry.

Rystad and their likes frequently makes other basic mistakes, as discussed next.

8 Selective overnight costs and project execution

Another issue is the overnight costs (total project costs except financial costs) of energy projects. Here, Rystad has done a monumental mistake by using old numbers for wind power – thus escaping the heavy inflation recently as noted earlier – but using prototype numbers from nuclear power. They also fail to include the costs of providing reliable output from wind power, i.e., the balancing/backup costs. In other words, Rystad’s comparison is as tilted in the disfavor of nuclear as it gets.

Also, Rystad still plays the learning curve model on wind, as if it is an immature technology, factoring in gains of more than 30% in the coming years. The only immature in the wind power industry is floating offshore wind, but the numbers there are so bad that no serious analyst can even suggest that they will approach anything that looks like competitive prices. That is also before the long-term economic issues on wind rotor and turbine performance is taken into account. The Siemens quality disaster unveiled the last year should spring to mind. This quality issue comes from lack of respecting basic mechanical engineering or plain technology optimism.

A wind power column is long and strong, but it is never completely stiff or still resulting in some motion. When you have a wind column in motion and a huge rotor on top giving a very large rotational torque onto the same column – with gears etc assuming all the forces – it is just a matter of time before gear failures will take place. The larger the rotor the higher the failure rate. Floating offshore wind will become much, much worse because then the motion of the sea will add to the challenge. This is perhaps possible to minimize but not without major increases in costs.

The overnight costs used for nuclear is even stranger. Typically, we see that reports use the European and American construction experiences, which are bad by all relevant standards of today. However, all of them are prototypes with the usual mumbo-jumbo prototype projects experience in any industry. On top, lack of experience and politics have exacerbated the issues. Interestingly, Rystad documents almost a 50% drop in overnight costs from the first APR1400 in UAE to the last, down to 2.5 MUSD/MW, but this is ignored in the report. The fact that UAE executed this project together with KEPCO with 12 years construction time is also ignored. That is, the report has listed the facts, but they seem to assume that Norway cannot learn from that.

Sure, the workforce onsite in Norway will be more expensive, but site works and civil works in total account for only 20% of capital costs and even a doubling will therefore change the 2,5 MUSD/MW to perhaps 3.0 USD/MW. So, why is Rystad so hesitant to using Korean numbers on nuclear in their report? Rystad probably use Korean numbers on deep-sea shipping and other markets, so why is nuclear so different? Nobody uses numbers from the US or many European countries in deep-sea shipping to estimate costs because they are not competitive. A Norwegian shipyard can deliver an advanced stern trawler at half the cost of US shipyards.

The time aspect is equally selective of the Rystad report. The fact is that nuclear outperform VREs also at scaling as shown in Figure 8. Nobody has decarbonized faster than Sweden. With a core team of just six people, Asea Atom built 12 reactors in 15 years. The French have also done very well and have the record among the larger countries. Note that the solar- and wind power numbers in Figure 8 do not include the balancing/back-up power. Figure 8 is therefore unfair towards nuclear despite showing overall very positive numbers for nuclear.

With this kind of selective application of numbers, anything goes. In fairness, an analysis of a technology should take into account reasonably well execution using proven technology. Otherwise, it is incomparable to mature technologies. The economic analysis of the report is therefore flawed.

9 Discounting in reverse

There are several financial and cost issues in the Rystad report that raise eyebrows, as they do in many other reports. Let us keep to a couple of issues, starting with discounting.

The Rystad report suggests that the discounting factor for wind should be 4% and 7% for nuclear power. Given that most reports in the field give energy sources across the board the same discounting factor, this report stands out on a substantive and important parameter without any serious disclaimer or relevant comment. We can reason around discounting factors in several ways.

Discounting factors are supposed to take into account the cost of capital, and one of the most commonly used is the Weighted Average Cost of Capital (WACC), which is essentially a weighted average between all sources of debt and equity. With the highly politicized atmosphere around energy projects, there is obviously great uncertainties. However, some basic facts cannot be escaped.

First, the time horizon of energy sources is critical because it attracts different investor classes. VREs may last for 25 years whereas nuclear easily last for 65 years and some have received license for 80 years in the US. The result is that those investors that invest in the companies making these technologies are different (could be Venture Capital, Private Equity and others) than those that invest in the energy projects itself (a large host of investor classes because often are the assets sold and resold). Nuclear, however, attracts different investors than VREs. Infrastructure investors and pension funds are those likely to invest in nuclear whereas VRE projects are investors with shorter time horizon. Critically, the longer the horizon the lower the return on equity (ROE) requirements.

This brings us to the discussion of what is a realistic long-term ROE. For example, Estrada finds that “…average across the 19 countries in the sample, stocks provided investors with an annualized real return of 4.7%, 3.8 percentage points higher than that of bonds (0.9%)”. This finding was based on the very large Dimson-Marsh-Staunton dataset, which covers 19 countries over 110 years. US stocks from 1802 to 2002 had a total annualized return of 7.9% whereas a third large data set across 17 countries from 1900 to 2005 averaged approximately 5%. In short, investors cannot expect the same ROE for such investments than for medium-term investments, which makes sense as a result of the powerful time diversification Bernstein wrote about.

Assuming 7.9% as ROE, which is historically high, 4% interest rate on debt, 23% corporate tax rate (OECD average in 2020) and the typical 30%/70% equity/debt ratio of nuclear power plant projects, the WACC becomes 4.5%, and this is important because the capital costs of nuclear power plants account for about 60% of the LCOE. A WACC of 4.5% is more aligned with the social discount rate used in many countries. The 7% used by many today therefore comes across as wishful thinking, ignorance or something else.

Applying the same approach to more medium-term investments, such as VREs, the economic parameters are more market driven by the alternatives existing at the time of investment. A ROE of 10% is often voiced and with the same debt and equity structure as well as interest on debt as for nuclear, the WACC becomes more in the 6-8% range. Thus, reading the report of Rystad is like they have just accidentally switched the discounting numbers around. That can be an honest mistake, but a major mistake concerning their conclusion.

10 Realistic LCOE calculations require RAM understanding and a system perspective

The Levelized Cost of Energy (LCOE) is perhaps one of the most abused metrics in energy today. Most seem to believe it is just a formula calculating the weighted average cost of producing energy throughout the life-cycle discounted to the present time. Both overnight costs, discounting and much more are therefore critical inputs. Unfortunately, much of the open domain information here is wrong. Indeed, by using audited information from Special Purpose Vehicle companies – which many wind power plants are for the purpose of managing risks – a study finds an accurate way to calculate the LCOE for given years. The study reveals that new wind power plants are achieving a LCOE of around 100 GBP/MWh which is a considerably higher level than implied by the most recent CfD bids of 57.50 GBP/MWh. Have Rystad fallen into the convenient trap of using open domain data?

Nevertheless, Rystad makes the traditional mistake we can witness just about everywhere, also in peer reviewed journals, they handle electricity as a conventional commodity. They fail to realize that electricity comes with many qualities that must be accounted for in a proper cost analysis. The reason for this erroneous bias cannot be anything else that the people using the deceptively simple LCOE metric are basically unaware of the greater context required to ensure apples are compared to apples.

In 1994, National Renewable Energy Laboratory (NREL) in the USA published an important report that explicates how we are to model the resource attributes to generate comparable numbers when planning energy systems. It is well worth repeating it, see Table 1.

Table 1 – Resource attributes of energy sources. Source: Logan et al. 1994.

Out of these, three are of great importance dubbed RAM (reliability, availability and maintainability) and therefore embedded in the Life-Cycle Costing standard of NATO. The short story of this is that many LCOE calculations may be mathematically right, but contextually flawed and highly misleading. The Rystad report is no different and in some respects even worse due to the flawed WACC and overnight costs. Therefore, discussing the numbers is waste of time because the entire premise around them is wrong, but one thing is obvious – to compare VREs directly to dispatchable energy sources is wrong unless the analysis is lifted to the system level to ensure a specified RAM. Bank of America has attempted to do just that, and the numbers speak for themselves, see Figure 9. The challenge with the System LCOE calculations is the complexity of the calculations and it becomes context specific, which is obviously mimicking reality.

I have done a similar life-cycle emission analysis for the Irish grid, and the results show that wind power can never enable us to reach the climate goals as long as wind power is balanced by fossil energy. Cost analysis of solar PV in CASIO shows the same. My analysis of the Norwegian grid where batteries are to balance wind is extremely depressing. The results shown in Figure 9 is therefore not unrealistic. Thus, adding RAM requirements to VREs increase their costs many times.

Then, we have some pundits that argue that the grid and the market will connect the VREs together and overcome the RAM issues discussed. First of all, that sentiment is mixing costs with expenses because costs are the resources required to do a job while expenses are the capacity provided to do the job. Expenses is what we find in the general ledger whereas costs can only be estimated unless they are direct. In our context, that means that the costs assumed by the grid is not visible unless the VREs require capacity increases, which is what we see. For example, in Norway wind power has existed within the grid reserve and therefore triggered very small amount of system expenses although the costs are there. Therefore, as VRE penetration increases and reaches high penetration levels, capacity investments will be required increasing system expenses. Furthermore, dispatchable energy sources are also subsidized to stay operational due to the low reliability of VREs. It is exactly like Vaclav Smil wrote concerning Energiewende in Germany; “In 2000, Germany had an installed capacity of 121 gigawatts and it generated 577 terawatt-hours, which is 54 percent as much as it theoretically could have done (that is, 54 percent was its capacity factor). In 2019, the country produced just 5 percent more (607 TWh), but its installed capacity was 80 percent higher (218.1 GW) because it now had two generating systems”. Obviously, keeping a reserve system in operation due to the poor reliability of the new system is extremely costly.

Second, using the grid to overcome the shortcomings of VREs implies that others have to pick up the costs. Thus, in many countries we end up in the situation that not only do the VREs have guaranteed income through CfDs, but the customers receive subsidies to soften the costs transferred to them by the same policies. What kind of market offer financial guarantees to both customers and suppliers? In between, we find a parody of a market.

Third, if wind power producers would have to pick up the tab of up to 100 euro per tonne EU-ETS allowance costs by the natural gas-powered utilities that must balance the wind power, the total energy cost would be even worse. Today, these emissions are directly transferred to customers, where the wind industry is largely exempted for this cost.

Fourth, nothing is being said about decommissioning- and waste handling costs in any LCOE analysis I have seen concerning VREs. Once more, it is just assumed that it will be handled, but the numbers are depressing as discussed in Section 11. In Norway, many offshore companies own the wind power plants, and I find it more than na?ve to believe that such companies will pay for the decommissioning and waste handling.

11 Let us not forget the alternative costs of unreliable energy supply

When people reject nuclear, do they really understand what they reject? The moving baseline theory tells us that a comparison towards zero is invalid, we must compare to real options. When these are ignored or mistreated, we have a problem as discussed earlier. However, the much larger issue is that the large alternative/opportunity costs are not discussed at all.

It is likely that if Germany had kept its nuclear power plants it would not have needed any Russian gas, and the brunt of current energy crisis would have been avoided. Deutsche Bank estimates that the energy crisis has cost 1.5 trillion euros to Germany alone. This is the cost of unreliable energy supply.

In fact, Germany could have built 15 Hinkley Point C plants (about 33 bn pound per plant) for the money they have spent on Energiewende (about 500 bn euro), and largely had a fossil-free power supply. Instead, Germany has missed their climate goals and ended up with an energy cost among the highest in the world. In fact, the German Federal Auditors described in 2021 the whole situation as a threat to German industry and population. They were right about that: giant BASF is leaving Europe with its need for reliable competitive power deliveries due to the high energy costs. 16% of German industry has made plans to leave Germany while another 30% is assessing the situation, including the wind power industry itself!

Texas offers another case. If we compare Texas to another grid that suffered the same winter storms, but did not fail, we see the issues. In Figure 10, we see Texas and what happened during the blackout in February 2021 when 246 people lost their lives and accruing at least 195 bn USD in damages. Pundits claim that it was the gas generators that failed, but anybody trained in problem-solving knows that we must always ask why? 5 times to identify the root causes. The reason for their failure was that they were not winterized, and neither was their supply of gas. Many gas pipeline compressors had electric motor drives with electricity supplied by the grid, and when these compressor stations were subjected to rolling black outs due to the power crisis the electricity and gas supply problem multiplied due to the resulting aggravated weather-related fuel limitations for gas fueled CCGTs. Why did the Texans not see all these short comings of their electricity supply system?

Well, the winter is normally a good season for wind power and why should gas generators take the cost of winterization when they will not run, and they have not been compensated to stay ready in case of failing wind power?

Of course, the lack of interconnections to the rest of the US made everything worse, but it is also for that reason the true costs of the system are easier to grasp. The problem in Europe, which is highly interconnected, is that everybody believes there is somebody else to pick up the cost of supplying the last MWh. Thus, a number of countries fail to take energy security seriously, and politicians are mostly looking out for problems tied to their own election cycles. Energy security requires a long-term approach. This failure of assuming long-term risk responsibility will certainly have costs. It is only a matter of time.

The causes of the collapse are shown in Figure 11. Weather was the dominant cause of the collapse. Indeed, if we go north to Colorado, we see in Figure 12 that the fossil fueled power plants functioned just fine in the same storm. In their summary of their report, American Society of Civil Engineers (ASCE) writes: ?ASCE Texas Section identified two primary and related problems: 1) a failure to support reliable dispatchable power generation, and 2) the negative impact from sources of intermittent electric power generation?.

The negative impact of VREs, such as wind power, is also manifest by the extreme volatility of the power prices in grids with high penetration of such energy sources. This also has alternative costs usually ignored.

However, such alternative costs cannot be ignored when reports spend considerable space at discussing all kinds of issues that may arise from nuclear power. The fact is that nuclear power has a very good track record. Three Mile Island, Chernobyl and Fukushima could all have been prevented if management on site had done their job, and if some relatively inexpensive suggested structural modifications had been carried out. An example is that the Chernobyl reactor had known design weaknesses that were handled by operating procedures that the local management willfully put aside triggering the disaster.

The worst case in the western world concerning legacy costs of nuclear is the Sellafield facility. This is also a case with special historical context that will not repeat itself, but it is frequently used by pundits as an example of how expensive nuclear power is. The Guardian recently published an article on Sellafield whereby the legacy cost was estimated to reach perhaps 260 billion GBP. This is, by any standards a huge cost, but it is in this context comparable to the cost of one blackout on a reasonably large grid. The major difference is that with increased VRE in the grid, the grid stability becomes more difficult to ensure and subsequently the risk of blackouts increases whereas Sellafield represents a diminishing problem.

Perhaps the worst of all is the environmental risks of VREs, largely kept out of the debate completely.

12 The environmental risks of VREs are always ignored

Bashing nuclear power for all sorts of ills is commonplace, but in all fairness the Rystad report does not go into the mythical landscape of hysteria like many other reports we find in the public space. However, the risks of renewables should have been communicated because they are in the long term a significant risk for society socially, environmentally and certainly financially.

For example, 20% of endangered? animal species are threatened by climate change whereas more than 80% is threatened by destruction of habitat and the likes where the renewable industry is a significant part of the problem. This only worsens the already large and unresolved conflict surrounding the usage of area for VREs. Furthermore, the puncturing of peatland, witnessed in Norway and many other countries, effectively leads to wind power plants being climate negative even on asset level. Insects, birds, noise and visual factors also add to the picture.

The destruction of the environment through the processing of the critical metals and minerals is also a major issue, as BBC points out. Indeed, for every tonne of rare earth elements produced, 2000 tonnes of toxic waste is produced and released. Slavery and child labor is also well documented, such as in Congo.

While nuclear waste is often portrayed as a huge problem, most reports ignore the sheer amount of waste by VREs. By 2050, 43 million tonnes of rotor blade waste is expected globally, and recycling is minimal. Indeed, another 78 million tonnes of solar panels is expected globally by 2050. The recycling is at the level of ordinary e-waste, i.e., about 20% or 40% in the EU. Large parts of the rest waste are shipped to Afrika once more violating the social dimension of sustainability. A final issue is the spread of Bisfenol A and PFAS into the environment by the erosion of wind turbine rotor blades. Radiation, in contrast, is well understood, has half-life and we know how to manage it.

When the cost of all this is to be accounted for in 2050, what will it be? The fact is that the nuclear industry has tangible costs because it is understood, whereas the mentioned issues are often externalities in the hope that some solution in the future will solve the matter.

13 Some final thoughts

A recent study finds the collapse of the human population inevitable on the account of resource consumption, but the study fails to connect resource consumption to energy usage. Humanity has always progressed from lower power density energy sources to higher, allowing increasing level of complexity and sophistication, as Vaclav Smil notes. This progression has fueled economic growth, brought people out of poverty, secured a growing population, fueled innovation and more.

Both the Romans and Ancient China had quite sophisticated technologies, but both civilizations were largely based on human- and animal labor. The industrial revolution was not a revolution in technology alone, but of equal importance were the availability of energy and capital that resulted in scale – scale that transitioned society.

The world today is predominantly fossil fueled based and after 10-15 years of renewable energy investments, there has been very little change (only 4 percentage points measured at primary energy supply excluding hydro power and 7 percentage points including hydro power). The fossil energy sources continue to rise despite all the rhetoric and political statements. Even coal is expected to grow through 2050!

So far, the global investments in renewable energy including hydro power amounts to 4.1 trillion USD (2022), according to numbers from the IEA. The return is far from satisfactory neither economically nor environmentally. A new energy system with such poor reliability as the VREs offer will never displace the old energy system. The marginal return for the total global energy system will therefore always be negative from the simple fact that two energy systems will always be more expensive than one. The problem is that societies collapse when the marginal returns on investments over time no longer support the complexities of the societies. Therefore, the energy transition is all about one primary question – how can we safely exploit the energy source with the highest power density to support population growth and displace the old energy system at the same time?

The energy transition we are facing now is not so fundamentally different than the industrial revolution except that we can no longer assume that Nature can absorb all our wastes indefinitely. We therefore need clean and abundant energy that both minimizes the negative impact on nature and that can be built at industrial scale for the next centuries. There is only one answer to this call – industrialization of nuclear power in all its aspects.

14 Epilogue

This short review of the Rystad report and similar reports, of which there are many, is far from exhaustive and complete. However, this review has been written based on research reports of a quality that should make it reliable and sufficient to demonstrate and understand that as a basis for decision-making the Rystad report has fundamental flaws rendering it inappropriate. I have intentionally used the typical narrative of VREs as a contrast to help the readers understand what we are discussing (the real options) and the distances between options. This is not to say that nuclear will solve everything. It will not.

The simple fact is that to fill in as much as possible rocks, gravel and sand in a jar, we must always start with the largest rocks first for then gradually to fill in the voids between these with a mixture of gravel and sand. This analogy is highly relevant here too. The large rock is the energy source with the highest energy density (nuclear) and not necessarily where we have invested so far.

There will also be substantial demand for other sources of energy, including VREs in the right context, but what we really must achieve is a technology neutral policy where apples are compared to apples and funded equally. For example, to calculate cost of nuclear that includes everything including waste handling and decommissioning, but with the life-span cut short to match the life-span of VREs, for then to compare the resulting numbers to VREs calculated narrowly on asset level with the wrong discounting, wrong investment costs, ignoring system costs and without decommissioning and waste handling is basically not only highly unfair, but plain wrong. This approach will lead to sub optimal, if not wrong, policies, power supply selection and investment decisions.

Coming from a reputed organization like Rystad, it is difficult to ascribe all the mistakes briefly discussed here to pure incompetence, bad luck and/or wrong contextualization. A memorable statement from Mr. Goldfinger to Mr. Bond comes to mind; ‘they have a saying in Chicago: "Once is happenstance. Twice is coincidence. The third time its enemy action”’. Time will tell!

[i] Data obtained 2021-03-01 from https://www.nve.no/energiforsyning/vindkraft/utbygde-vindkraftverk/

[ii] Data obtained 2021-03-01 from https://www.statnett.no/for-aktorer-i-kraftbransjen/tall-og-data-fra-kraftsystemet/last-ned-grunndata/