Prioritise safe, viable extensions of Gösgen and Leibstadt. A Swiss study estimates the price needed for 60-to-80-year operation at CHF 46/MWh for Gösgen and CHF 44/MWh for Leibstadt.
These are 2025-money estimates based on operator costs. The planned case assumes existing decommissioning funds are sufficient. Operation still needs safety approval. New reactors and a complete replacement system have different costs.
Extension scenarios and cost scopeExtending existing nuclear
The first decision concerns the reactors Switzerland already has. Extending an operating reactor uses the existing plant. The extension calculation covers further investment, maintenance, fuel and future obligations. The Swiss study asks whether the additional income can cover those costs.
| Study case | Gösgen | Leibstadt |
|---|---|---|
| Planned extension to 80 years | 46 | 44 |
| Operate six months per year | 78 | 72 |
| Close after two extension years | 134 | 104 |
The study uses operator best-estimate investment costs and 7% nominal pretax finance. Nominal includes inflation; it differs from the real rate in the new-plant example.
The planned case assumes sufficient existing prefunding and no additional dismantling costs. Early closure includes continuing costs after shutdown. The estimates cover extensions of the existing reactors. Operation still requires safety approval.
Our assessment: pursue these extensions while safety and owner economics support them. Before retiring useful capacity, establish the cost and readiness of the supply that replaces its service. The study’s prices apply to Gösgen and Leibstadt; Beznau has a separate investment and retirement decision.
A new reactor requires a construction offer and a financing plan. Its cost cannot be inferred from an existing plant’s extension price.
What does “cheaper” actually buy?
A useful solar or wind project can reduce the cost of generating electricity. The next question is what supplies users when its output falls. For nuclear, the equivalent questions include construction finance, fuel, waste, maintenance and unexpected outages. The fair comparison follows each option through to the same delivered service. Generation cost, duty cycle and value · Evidence behind our comparison
LCOE, the levelized cost of electricity, spreads a plant’s stated lifetime costs over its expected output. It answers “what does generation cost?” Dependable delivery needs the surrounding system too.
Generation cost and limitsSolar & wind
- Build, finance and connect
- Maintain, replace and retire
Nuclear
- Build, finance and connect
- Maintain, fuel, manage waste and close
- Winter nights
- Cloudy, calm spells
- Plant outages
- Dry years
Same demand. Same reliability requirement. Same accounting boundary.
Count each cost once. Show owner returns, customer bills and public guarantees separately.
This is a cost map. Box sizes show no prices or cost shares. The required mix and its total price must be calculated.
Hydro, batteries, imports and flexible demand can complement both kinds of generation. Choose their size and operation together. Counting a separate, complete backup system for each generator would misrepresent a shared electricity system.
Single-source and mixed-system comparisonsThe argument that cheaper renewable generation makes nuclear unnecessary skips this comparison. A low price for generation supports building useful generation. Excluding nuclear requires evidence that the remaining supply plan can deliver the same service at a better overall cost.
Argument, evidence and limitsA Swiss price tag: rooftop solar
Start with a cost we can observe. A recent Swiss survey reports what installers offered or charged for rooftop solar. The price varies with size and scope.
CHF per kWp, excluding VAT · scale 0–3,500
A kWp is one kilowatt of rated peak solar power.
Coloured band: middle half of prices. Dark mark: median, the middle price. The band shows the spread of the surveyed prices.
- 2–10 kWp408 records2'497
- 10–30 kWp907 records1'911
- 30–100 kWp141 records1'428
- 100–300 kWp88 records1'102
- 300–1000 kWp47 records800
- >1000 kWp7 records660
The largest class has only seven records. Prices exclude important roof, external engineering and grid work, as well as batteries. Lifetime electricity costs also depend on output, finance, upkeep and replacement.
Price table and survey scope
| Size, kWp | Records | Lower quartile | Median | Upper quartile |
|---|---|---|---|---|
| 2–10 | 408 | 2'096 | 2'497 | 3'114 |
| 10–30 | 907 | 1'659 | 1'911 | 2'237 |
| 30–100 | 141 | 1'125 | 1'428 | 1'751 |
| 100–300 | 88 | 907 | 1'102 | 1'597 |
| 300–1000 | 47 | 705 | 800 | 1'010 |
| >1000 | 7 | 602 | 660 | 772 |
The survey combines 2025 offers and invoices within the rooftop installer's scope. Current procurement, alpine projects and ground-mounted projects require their own offers.
Moving from this price to a lifetime electricity cost needs more inputs: site output, useful life, financing, maintenance and replacement. Moving to reliable supply also needs the rest of the system.
Inputs needed for a generation-cost calculationUseful solar needs a buyer
Cheap roofs can earn a place in the supply mix. The comparison depends on winter output, summer sales and the cost of covering the remaining hours. Compare site costs, summer prices and storage services.
Cost and market value are different
Cheap production is useful. Its income also depends on when electricity reaches the market. An owner needs both a cost estimate and a credible account of what customers will pay.
Cost and avoided valueCHF/MWh · common scale 0–150 · 2026 Q1–Q3
Solar
Hydro group
Wind
Biomass
Q1 is January–March; Q2 is April–June; Q3 is July–September. Coverage is limited to the measured groups. The hydro sample covers selected categories, including run-of-river, diversion and drinking-water plants. These prices describe those quarters; future receipts depend on the market and each owner’s contracts.
What the benchmark includes
BFE weights Swiss day-ahead prices by measured net injections. VAT, plant costs and complete capacity or flexibility value are outside these figures.
The hydro group includes categories such as run-of-river, diversion and drinking-water plants. Reservoir owners can earn different prices through their operating choices and contracts.
A plant’s profit also depends on its contract, own-use savings and costs.
The difference between groups does not establish a battery’s trading margin. Charging and discharge hours, losses and costs must be matched.
We favour solar projects with useful output and a credible buyer. A cheap installation can still earn too little if its electricity arrives when many other plants are producing. Flexible use and well-priced contracts can improve that investment case; adding capacity alone leaves the revenue question open.
Historical trading and commercial limitsFinancing changes the price of new construction
A capital-heavy project spends money before it sells electricity. The financing rate and the time until operation affect the revenue it must later earn. The example below isolates those effects under explicit assumptions.
Assumed new nuclear plant: CHF 10,000/kW before construction finance, 60 operating years and 90% utilization. The calculation uses analyst assumptions to isolate financing effects; a project decision needs a supplier offer.
At 7% real finance: 5 years to build → CHF 138/MWh; 10 years to build → CHF 159/MWh; 15 years to build → CHF 187/MWh.
Partial generation cost, CHF/MWh · common scale 0–350
The rate is above inflation. Subsidies and future cost reductions are set to zero. Grid, backup, taxes, full insurance terms and unresolved dismantling or fund costs are excluded.
All rates, formulas and missing costs
Money is spent evenly at annual midpoints during construction. Each payment carries financing costs until the plant opens. Longer construction also changes the payment schedule here; an idle delay is a different calculation.
Annual cost = financed capital × capital-recovery factor + CHF 120/kW fixed upkeep. Divide by annual output, then add CHF 15/MWh for fuel, variable costs and a waste allowance.
The capital-recovery factor is r ÷ [1 − (1 + r)−n], where r is the annual rate and n the operating years. Annual output is rated power × 8,760 hours × 90%.
Extra site overhead and replacement electricity during delays are excluded. A complete project assessment must add missing items and test outages, reduced use and different lifetimes.
| Real annual rate | 5-year build | 10-year build | 15-year build |
|---|---|---|---|
| 3% | 80 | 84 | 88 |
| 5% | 106 | 117 | 129 |
| 7% | 138 | 159 | 187 |
| 10% | 193 | 243 | 313 |
The 2050 programme applies a broader owner calculation, including operation, refurbishment and funding obligations. It uses its own explicit construction and upkeep allowances.
Programme budget and lifetime owner cashflowsThat calculation identifies waste handling, storage and decommissioning funding, but no separate geological-disposal contribution. Its funding scope needs completion before the required revenue can support a project offer. Achievable sales and received prices remain separate commercial questions.
Disposal funding and revenue boundariesOur choice is to fund project development before committing to construction. Development should produce an offer with a price, schedule, financing plan and named responsibility for delays and overruns. Public guarantees can lower financing costs while exposing taxpayers to risk. We support only an explicit, bounded decision on that exposure.
ETH–PSI construction and financing inputs
A Swiss study makes the financing choice concrete
The ETH–PSI study tests nuclear construction inputs of CHF 5'000, 8'000 and 12'000 per kW, in 2024 money. These overnight costs exclude construction financing. The study compares 8% real financing with a 5% case that assumes lower investment risk, such as through public guarantees. Its scenarios are inputs for testing economics, rather than current Swiss construction offers.
Construction and financing inputs, Table 3 and sections 2.2.2–2.2.3The study minimises costs across the system from a central planner’s perspective. For nuclear and hydro owners, profitability also depends on prices, contracts, operating obligations and who carries the risks.
System economics and the central-planner perspective, page 15Our recommendation: seek financing offers with and without public guarantees. Show the price difference and the obligations transferred to the guarantor. Require a viable owner business case alongside the system comparison. The nuclear chapter shows the study’s conditional result and the contrasting market-driven cases.
Storage cost depends on how often it is used
A battery used most days spreads its fixed costs over more discharged electricity than one kept for rare events. That changes its cost per unit, even if both batteries cost the same to buy.
Cost including charging, CHF/MWh discharged · scale 0–3,000
A full cycle delivers the store's usable energy once. The assumed purchase cost is CHF 1.5 million; life is 15 years; finance is 7% real.
Rarely used backup can still have value. This chart prices discharged energy. Emergency availability has a separate value that needs its own estimate.
Charging, upkeep and use assumptions
Charging electricity costs an assumed CHF 40/MWh. At 85% round-trip efficiency, delivering one MWh requires buying about 1.18 MWh. Charging is counted once.
The annual cost allowance is 4% of purchase cost. It already includes operation and battery augmentation, which replaces lost capacity.
Annual discharged MWh = 4 × full cycles. Cost per discharged MWh = [capital recovery + annual upkeep] ÷ annual discharge + 40 ÷ 0.85.
All four cases fit within a necessary charging/discharging time budget. Achieving those cycles also requires charging electricity and suitable operating opportunities; income depends on the market and contracts. Site costs, taxes and full insurance terms remain incomplete.
We give batteries a short-duration role in the base plan. Frequent useful cycling spreads the fixed cost. Rare emergency service can still be valuable when a readiness contract pays for availability. The example provides no economic basis for assuming a cheap national seasonal battery store. Storage duty cycle and reserve service · Trading and ownership comparison
Who pays for reliable supply?
Hydro owners give up other opportunities when they reserve water. Storage owners buy charging electricity. Investors need to recover construction and renewal costs. Those obligations remain even when a national model can find a useful operating schedule.
Hydro opportunity cost and owner economicsOur preferred payment arrangement is explicit. Customers buying winter delivery pay for that electricity. A reserve contract also pays the owner to remain available. Hydro contracts compensate the owner for water held back and other sales forgone.
For a new reactor, the owner, lenders, customers and supplier need agreed responsibility for construction and operating risks. Any public guarantee should have a stated limit and a named beneficiary. Unresolved insurance and accident-liability exposure remains visible in the wider assessment.
Proposed commercial arrangement Cost, transfers and risk boundariesThe evidence supports keeping nuclear in this comparison. Establishing the cheapest reliable Swiss mix still requires matched project offers, network costs and a complete supply assessment. Our 2035 supply plan identifies the winter energy and peak power that still need to be secured.
Recommendation and remaining evidence needsPublic debate: support payments and profitability
In a July 2025 interview, Annalisa Manera challenged SCNAT’s treatment of nuclear delivery and profitability and compared nuclear income with wind-support tariffs.
SCNAT replied that its timetable reflected sequential project processes and that profitability remained uncertain. These are attributed positions in a public debate, separate from the joint ETH–PSI study.
Manera interview, 13 July 2025 SCNAT's institutional response
For our comparison, a tariff multiplied by output gives gross revenue. Net support needs a market-revenue baseline. An owner must cover construction, operation and financing from all revenues. Equal annual generation still needs a separate comparison of winter delivery and dependable capacity.
LCOE, LFSCOE and value: definitions and comparison limits
What LCOE and LFSCOE can tell us
Levelized cost of electricity, or LCOE, spreads the stated investment and operating costs over the electricity a plant is expected to generate. It accounts for the timing of spending and the assumed return on capital. It helps compare the cost of generation projects under stated assumptions. Market value and reliable delivery need separate analysis.
LCOE definition and limitationsA fair national comparison uses the same demand and weather, then checks outages, networks, trade limits and project delivery. Its costs need a common scope.
For a closer comparison: equations, value and reliability
Levelized full system costs of electricity, or LFSCOE, asks a broader supply question. In Idel’s model, the main experiment requires a selected technology, with storage, to supply a market. It exposes the cost of covering gaps in that configuration. The author also tests wind with solar and a case allowing limited other supply.
Idel: model and alternative casesSwitzerland’s decision includes existing hydro, several generation sources, trade and different forms of flexibility. Pricing that mixed system requires a model with the relevant technologies and reliability requirements.
Single-source and mixed-system costsLCOE divides discounted expenditure by discounted net generation. Discounting converts future amounts to a common valuation date. State the plant life, rate, delivery point and residual value.
LCOS, the levelized cost of storage, uses discharged electricity as its denominator. Check whether charging electricity is included. Duration, cycling, losses and replacement assumptions must match.
The detailed LFSCOE discussion uses the complete February 2021 manuscript; changes in the final 2022 publication remain unverified. Its 95% variant concerns annual energy, not 95% of hours. Hourly reliability and outages still need their own tests.
Its restricted technology set, separate yearly designs, storage power-to-energy ratio and simplified outage/network treatment need checking before applying its results to Switzerland.
IEA’s VALCOE adjusts generation cost for energy, capacity and flexibility value. EIA’s LACE estimates avoided energy and capacity value. These add useful context but depend on their own system assumptions.
A portfolio comparison should price the future assets and operation needed to meet a common demand and reliability standard. Owner returns and household bills remain separate results.
Include emissions, waste, insurance and accident liability in the wider assessment. Identify risks that prices and insurance do not cover, alongside unresolved private contract terms.
LCOE, LCOS and avoided value · LFSCOE equations and limits · Value-adjusted cost · Metric definitions and comparison requirements
References
- Programme assumptions, dispatch choices and cost scope
WattInfo · Archived-input interpretation, 25 September 2026 - WattInfo electricity programme: 66 hourly cases, delivery dates and owner finance
WattInfo · Conditional programme, 19 September 2026 - WattInfo proposed supply path: decisions and supporting calculations
WattInfo · Proposed priorities and fixed-charging supply screen, 18 September 2026 - Generation and dependable supply: cost comparison guide
WattInfo · Qualitative infographic and evidence map, 18 September 2026 - Generation cost, storage cost and avoided value
EIA · April 2026 methodology - Levelized full system costs of electricity: author manuscript
Robert Idel · February 2021 author manuscript - System costs and the value of power sources
Yuhji Matsuo / Energies · Published 2022 - Value-adjusted generation cost: IEA model documentation
IEA · Global Energy and Climate Model, 2025 - Swiss rooftop solar installation prices in 2025
Planair SA for EnergieSchweiz/BFE · 2025 offers and invoices; report dated 3 August 2026 - Swiss reference market prices: first two quarters of 2026
BFE · Publication dated 14 July 2026 - How Swiss reference market prices are calculated
BFE · Printed update: 1 April 2026 - Economics of extending Gösgen and Leibstadt to 80 years
BFE commissioned research · Report dated 12 March 2026; publication date not established - Cost comparisons: inputs, calculations and boundaries
WattInfo · Curated from the cost-metrics research package, 17 September 2026 - Storage owner economics: historical-price benchmark
Swiss Energy research project · Frozen research output - ETH–PSI: Nuclear Power and the Future Swiss Energy System
ETH Zurich and PSI; 19 authors including Manera · 2026-06-29 - ETH–PSI: nuclear white paper modelling documentation
ETH Zurich and PSI · 2026-06-29 - ETH–PSI study: conditions, model inputs and reading notes
WattInfo · Study review, 18 September 2026 - Annalisa Manera: July 2025 interview on nuclear policy and support
Mischa Aebi / Tamedia / Berner Zeitung · 2025-07-13 - SCNAT response to the July 2025 Manera interview
SCNAT official LinkedIn account · Response to the July 2025 interview; exact date unverified - Winter solar: capacity, seasonal costs and paid storage
WattInfo · Conditional component comparison, 22 September 2026 - Reference market prices under EnFV Art15
BFE · 2026-10-13 - Industry fuel totals, generation price references and delivery gates
WattInfo · 2026-10-14
Find definitions, calculations and model inputs in the source register.