The Nuclear Mirage: Why Germany’s Atomic U-Turn Doesn’t Add Up

There is a particular kind of political argument that survives entirely on vagueness. It doesn’t need to be right. It just needs to sound reasonable long enough to avoid being checked against the numbers.

Germany’s nuclear comeback debate is that argument.

The premise, advanced recently by prominent voices in German conservative politics — among them Kristina Schröder, former Federal Minister for Family Affairs and a rising figure in the CDU’s economic wing — is seductively simple: the Atomausstieg, Germany’s legislated nuclear phase-out, is not irreversible. Technically, they are correct. The question was never whether a return to nuclear is physically possible. The question is whether anyone with access to a spreadsheet would recommend it.

Let’s open the spreadsheet.

The €30 Billion Question

A modern Generation III+ reactor — the European Pressurized Reactor (EPR), to be specific — costs between €20 and €30 billion per unit. In return, you get approximately 1.6 GW of installed capacity, around 13–16 TWh of annual output at 90% availability, and a first kilowatt-hour delivered somewhere between 2040 and 2045, depending on how optimistic your planning assumptions are.

Take the same €30 billion and build a hybrid portfolio of utility-scale solar, onshore wind, battery storage, and grid modernisation. The result: roughly 25 GW of installed capacity, somewhere between 28 and 35 TWh of annual energy yield — and your first megawatt-hour arrives in year three, not year eighteen.

That is not an ideological statement. That is an opportunity cost calculation. And it does not favour the reactor.

The Interest Trap

The economic case against nuclear isn’t only about construction costs. It’s about what time costs in capital terms.

Even at a favourable 2% government-backed discount rate, a €30 billion project with a 15-year construction period accumulates roughly €10 billion in capitalised interest before generating a single cent of revenue. The hybrid model, meanwhile, begins producing cashflow in year three. That revenue can be reinvested — in electrolysers, in long-duration storage, in grid digitalisation — compounding forward while the reactor is still being poured in concrete.

In finance, this is called time-to-market risk. In nuclear new-build, it is simply called normal.

The Baseload Myth

Proponents of nuclear power often invoke the concept of baseload — the idea that a stable grid requires a constant, dispatchable floor of generation that renewables, with their intermittency, cannot reliably provide. The argument has intuitive appeal. It also describes a grid architecture that is becoming structurally obsolete.

In a modern inverter-based grid increasingly dominated by wind and solar — both with near-zero marginal costs — a nuclear plant running at constant full capacity becomes a liability rather than an asset. On a high-generation afternoon, when solar is flooding the grid and wholesale prices go negative, the reactor keeps running. It has no practical alternative: the financial model of nuclear power depends on near-continuous full utilisation. Curtailment doesn’t merely reduce revenue. It unravels the entire investment thesis.

A hybrid system of wind, solar, and battery storage offers what a steam turbine cannot: flexibility, fast frequency response, and the ability to modulate output with market conditions rather than against them. Baseload was an answer to a question the 2040s grid will not be asking.

The Ordoliberal Paradox

This is where the argument becomes particularly uncomfortable for a specific type of nuclear advocate — one who pairs enthusiasm for atomic power with a stated commitment to free markets and economic liberalism. It is a combination that requires some explaining.

Nuclear power in its current form is not a market product. It is a state-sponsored construction programme dressed in the language of industrial policy. Three structural realities underpin this assessment.

First, there is no private insurance market anywhere in the world willing to cover the full liability of a catastrophic nuclear failure. The state acts as insurer of last resort, providing an implicit subsidy that appears in no LCOE calculation and on no corporate balance sheet. Second, every major nuclear new-build in Europe — Hinkley Point C being the most transparent example — operates under state-guaranteed offtake prices fixed for 35 years at rates well above market. Without those guarantees, private capital does not participate. Third, the costs of waste disposal are systematically deferred to future generations while profits flow in the present — a straightforward violation of the polluter-pays principle that sits at the foundation of any coherent market framework.

This is socialised risk with privatised returns. It is not the free market. And anyone arguing simultaneously for economic liberalism and nuclear new-build owes their audience a reconciliation of those two positions.

The Finnish Lesson

Proponents of nuclear new-build have one genuinely strong card to play: Finland. Olkiluoto 3 is real, it is operational, and Onkalo — the world’s first deep geological repository for high-level nuclear waste — is a legitimate engineering milestone that deserves to be acknowledged as such.

So let’s take it seriously.

Olkiluoto 3 broke ground in 2005. It was projected to take four years and cost €3 billion. It took eighteen years and cost approximately €11 billion. Shortly after going online, it had to curtail output because renewable-driven market prices had fallen to the point where running the reactor at full capacity was a net financial loss.

And Onkalo? Its operating licence covers the waste inventory of Finland’s existing five reactors — precisely those five, and no others. Any new reactor would require a new repository: a process that has taken Finland the better part of four decades, required extraordinary geological conditions, and is not a procedure that scales on demand.

Finland is not a proof of concept for nuclear expansion. It is a demonstration of what is achievable under near-ideal conditions — and even then, at nearly four times the projected cost.

Conclusion: The Arithmetic of Irrelevance

Nuclear power is not impossible. In certain national contexts, with existing fleets and established supply chains, it continues to play a legitimate role in low-carbon electricity systems. The argument here is not categorical.

The argument is specific: for Germany in 2025, nuclear new-build is not a solution. It is a very expensive way to arrive late.

A technology that requires 15 to 20 years to deliver, depends on state liability guarantees no private insurer will underwrite, produces electricity at three to four times the cost of wind and solar, and reaches the grid precisely as cheaper alternatives have already done the work — that technology does not deserve to be called a climate strategy. It deserves to be called what it is: a capital allocation decision with a 20-year payoff horizon in a problem that requires action in the next five.

The spreadsheet has a clear opinion on the matter. It simply doesn’t get invited to the press conferences.