The UK government’s decision to extend the operational life of Sizewell B by 20 years from 2035, contracting its 1.2 GW output under a contract for difference at £70.50 per MWh in 2025 prices, provides the clearest public benchmark yet for what nuclear lifetime extension actually costs in the British market.
The figure is simultaneously useful as a policy instrument, instructive as a market reference point, and uncomfortable for an industry that has long suggested existing plants can continue running at marginal cost. Indexed to inflation and applicable to production beginning several years hence, the price lands very close to the most recent offshore wind CfD allocations in real terms, a comparison with consequences that extend well beyond Sizewell B itself.
The most recent round of UK offshore wind CfDs allocated prices around £90 per MWh in 2024 prices for plants likely to be operational around 2030. Adjusting Sizewell B’s £70.50 per MWh in 2025 prices forward to the point where output actually begins under the new contract produces a figure in broadly comparable territory, given the indexation mechanism. The implication is that nuclear lifetime extension, which involves known and manageable refurbishment costs rather than the procurement, construction, and commissioning risks of new build, is pricing at rough parity with new offshore wind. For advocates of new-build nuclear, that is an awkward reference. If extending a fully amortised plant with existing grid connections and established safety case requires a strike price competitive with new offshore wind, the economics of genuinely new nuclear plant construction, which carries far higher construction risk and longer lead times, face a harder justification.
The CfD Structure and Why It Suits This Asset
The contract for difference mechanism is the appropriate structure for a capital-intensive baseload plant with well-understood operating costs. Sizewell B’s cost profile over the extension period is dominated by the decennial heavy maintenance cycle and ongoing operations and maintenance expenditure rather than fuel price volatility, which is the primary uncertainty facing gas generators. A fixed strike price transfers volume and price risk to the counterparty rather than the operator, while leaving the operator exposed to cost overrun risk on the maintenance and refurbishment programme it controls.
That cost allocation is one of the deal’s genuine strengths. EDF, with Centrica holding a 20% stake, must deliver power at £70.50 per MWh regardless of what the plant costs to keep running. If maintenance costs exceed assumptions, the operator absorbs the difference. If the plant runs more efficiently than forecast, the operator retains the margin. This is a materially better incentive structure than regulated cost-plus arrangements, which allow operators to pass cost overruns through to consumers and reduce the pressure to manage expenditure efficiently. The Office for Nuclear Regulation has confirmed that existing safety requirements remain fully in force and are not superseded by the CfD, addressing the risk that commercial contracts create pressure for safety standard relaxation.
The revenue certainty provided by a 20-year CfD also has balance sheet implications. Where a merchant plant must borrow against uncertain future power prices, a contracted revenue stream at a fixed strike price reduces the risk premium lenders demand. Whether the CfD enables limited recourse financing structures or simply improves the terms of corporate funding, the financing cost reduction it supports could be meaningful at the scale of a 1.2 GW plant over two decades.
The Pricing Question and Its Limits
The critical question about any bilaterally negotiated CfD for an asset without competitive alternatives is whether the agreed price reflects realistic cost assumptions or reflects successful extraction of rent by the operator. There is no market competition for a Sizewell B lifetime extension CfD because there is only one Sizewell B. The price was presumably negotiated based on detailed cost submissions from EDF verified by the regulator and independent advisors, but the information asymmetry between the plant operator and the government counterparty is substantial. EDF has access to decades of operational data and engineering knowledge about the plant’s condition that no independent reviewer can fully replicate.
The Hinkley Point C CfD, agreed a decade ago at £89.50 per MWh for 35 years indexed from 2012, currently stands at approximately £131 per MWh in nominal terms, a level that reflects the significant cost overruns EDF has experienced in construction. EDF has not been satisfied with the Hinkley terms as actual costs have escalated beyond the assumptions underpinning the original strike price. That precedent works in both directions in the Sizewell B negotiation: the government knows what happens when a nuclear CfD is set against optimistic cost projections, and EDF knows that a strike price implying near-parity with offshore wind sets a visible ceiling on what it can claim for future projects.
The nuclear industry faces a specific reputational constraint on the pricing of lifetime extensions. If an existing plant with no construction risk and fully depreciated capital costs requires a strike price that merely matches new renewable capacity, the implied levelised cost of new-build nuclear becomes very difficult to defend publicly. The industry therefore has an incentive to demonstrate that lifetime extensions are cost-competitive, even if the underlying economics could support a higher ask. Whether that constraint produced a genuinely competitive price or a price that is low enough to avoid adverse comparison while still capturing substantial margin is impossible to assess without access to the undisclosed cost submissions.
System Cost Implications the CfD Does Not Address
The strike price of £70.50 per MWh represents the contracted cost of the electricity. It does not represent the full system cost of Sizewell B’s continued operation, a distinction that matters increasingly as the UK’s generation mix shifts toward high shares of variable renewables.
A 1.2 GW plant operating as a must-run baseload asset requires dedicated backup capacity for scenarios where the plant or its associated grid infrastructure is taken out of service. The N-1 contingency planning requirement for a plant of this size necessitates spinning reserve or standby capacity that is rarely utilised but must be available, a cost that sits in the wider system rather than in the plant’s operating budget. More acutely, a plant that must run continuously at full output regardless of system conditions will increasingly come into conflict with solar generation as PV capacity continues to grow. During periods of surplus, nuclear plants as inflexible generators impose curtailment costs on other parties rather than reducing their own output, a system externality that grows with solar penetration and is not captured in the £70.50 per MWh strike price.
The CfD structure could partially address this through calculating the differential payment against an average price over a defined period rather than the instantaneous market price. This would ensure that when power prices are zero or negative, the CfD payment reflects the actual market value of nuclear output rather than guaranteeing full subsidy during hours when the grid has more power than it can use. Many existing plants across wind, solar, and other generation categories operate under CfD or support contract terms that do not impose this structure, which creates inconsistencies in how system costs are distributed and makes the true cost of different generation technologies harder to compare. Applying a consistent capture price mechanism across all supported generation, including Sizewell B, would improve both market signals and transparency. Whether the specific terms of the Sizewell B CfD include such provisions has not been publicly disclosed in detail, but it is a design choice with material consequences for how the plant interacts with an increasingly solar-heavy grid through the 2030s and 2040s.
