An updated study by the Fraunhofer Institute for Solar Energy Systems (Fraunhofer ISE), entitled Levelized Cost of Electricity: Renewable Energy Technologies, is currently attracting considerable attention on social media. Its prominently promoted central claim is that photovoltaic systems in Germany can generate electricity more cheaply than conventional power plants, even when combined with battery storage. At least in the simplified communication on social media, this creates the impression that renewable energy technologies have thereby achieved a cost advantage at the level of the electricity system as a whole.
However, the study itself makes clear, albeit somewhat indirectly in connection with the assumed full-load hours of conventional power plants, that it does not examine system-wide electricity generation costs. Nor does it take into account system-specific cost factors such as the construction of additional backup power plants, increased curtailment or grid expansion associated with variable renewable energy sources. Waste management costs for both renewable energy technologies and nuclear power are also excluded. Consequently, this study, which focuses solely on technology-specific levelized costs of electricity (LCOE), fails to address the fundamental challenge of designing an effective electricity system that combines security of supply, climate and environmental protection, and economic competitiveness.
The limited scope of the study is also reflected in its assumptions regarding the various technologies. This becomes particularly apparent in the treatment of electricity storage. Since the study does not aim to provide a comprehensive system analysis, the fundamental challenge posed by variable electricity generation remains unresolved, and the cost calculations are based on very limited battery storage capacities. For small rooftop PV systems, the assumed storage capacity corresponds to one hour of the PV system’s peak output. For large rooftop installations, the figure is just 30 minutes, while ground-mounted PV systems are assigned 40 minutes, with efficiency losses further reducing the usable capacity. In practical terms, this means that if the batteries are used for peak shaving during periods of maximum generation and to compensate for declining output in the afternoon, they will be completely discharged before nightfall. For wind power, which is also characterized by variable generation, storage is not considered at all, since the study is not intended to provide a system-level analysis.
Other parameters used in the calculations reveal a noticeable bias in favor of the renewable energy technologies preferred by the study’s authors, while nuclear power is made to appear particularly expensive, with estimated generation costs ranging from EUR 130 to EUR 480 per megawatt-hour. For variable solar and wind generation, the study assumes unrealistically high full-load hours. For photovoltaic systems, the assumed range extends from 935 full-load hours in northern Germany to 1,280 hours in the south, although the national average is only around 950 hours. For onshore wind power, the study assumes between 1,800 full-load hours at inland locations and 3,200 hours at favorable sites in northern Germany, many of which are presumably already in use. In reality, the nationwide average is only approximately 1,850 hours.
A similar discrepancy can be observed for offshore wind power. The study assumes between 3,200 and 4,500 full-load hours, whereas actual installations achieve approximately 3,400 hours in practice. The assumptions regarding plant lifetimes are equally striking. Nuclear power plants are assigned a relatively short operating lifetime of 45 years. Lignite-fired power plants are assumed to operate for 40 years, while hard coal-fired power plants are assigned just 30 years. These assumptions also place conventional generation technologies at a disadvantage.
The study also assumes significantly higher interest rates for both debt and equity financing of nuclear power plants than for photovoltaic and wind power projects, including offshore wind. For nuclear projects, an equity return of 12 percent is assumed. This figure even exceeds the implicit return associated with the heavily criticized financing arrangements for Hinkley Point C and reaches a level that effectively discourages investment in nuclear power projects. This is precisely why countries seeking to make the broader economic benefits of nuclear energy available to businesses and consumers are developing financing mechanisms that allow for lower required returns and borrowing costs.
Such considerations, however, appear to be entirely absent from the authors’ approach, as the cost analysis for nuclear power is conducted “within the framework of energy policy planning for the German energy system”. In other words, the hypothetical inclusion of nuclear power in the study is based on the assumption that nuclear power plants would somehow be constructed while an anti-nuclear policy continued unchanged and the objective of achieving an electricity supply based entirely on renewable energy remained in place.
This assumption is reflected in the figures used for nuclear power plants. The study assumes between 4,300 and 6,300 full-load hours in 2024, declining to just 2,000 to 4,000 hours by 2045. These figures have devastating implications for the cost structure of nuclear power plants, which historically achieved approximately 7,500 to 8,000 full-load hours per year in Germany. Ultimately, this serves to create a circular argument within an unquestioned and effectively unquestionable policy paradigm, allowing the authors to demonstrate to an already sympathetic audience that nuclear energy is excessively expensive and entirely pointless.
In conclusion, the latest Fraunhofer ISE study on the levelized cost of electricity may well have merits in its extensive examination of photovoltaics, including from a global perspective, as this is the institute’s primary area of expertise. However, as an instrument for informing energy policy, this study, like any analysis based solely on LCOE, is unsuitable. It cannot, and does not seek to, meet the expectations attributed to it in the communication surrounding the study on social media.