What each technology is actually good at
Wind and solar supply low marginal cost energy in large volumes, and they do it with long-run cost reductions, though individual project costs vary. What they do not supply is control. Output follows weather, the correlation across a geography like Great Britain is high, and the system has to be able to meet demand during the periods when it is low across the whole area at once.
Batteries supply speed. They are the best available answer to second-by-second frequency management, fast reserve and intraday arbitrage, and they are getting better and cheaper. What they do not supply is duration. A four-hour asset moves energy around within a day. It does not cover a still, cold, cloudy week, and the economics of building enough lithium capacity to do so are not close.
Nuclear supplies firmness and, importantly, physical inertia. It runs at high capacity factor largely independent of weather, it occupies a small land footprint, and as a synchronous machine it contributes to system stability in a way that inverter-connected generation only emulates. What it does not supply is cheap flexibility: the economics depend on running the asset hard, because the costs are overwhelmingly capital.
Those are three different products. A system that buys only one of them pays heavily, whichever one it picks.
The real question is residual demand
The useful analytical frame is not comparing generation costs. It is asking what the residual demand curve looks like after renewables have run, and what it costs to serve it.
As renewable penetration rises, the residual curve becomes spikier and less predictable. Average residual demand falls, but the peaks do not fall proportionately, and the duration of the worst events does not shorten much at all. The cost of serving that curve is what determines total system cost, and it is dominated by assets that run at low utilisation. Interconnectors help by widening the weather geography. Demand flexibility helps by reshaping the curve. Long-duration storage helps if it can be built at the required scale and duration. Firm thermal generation helps if the emissions can be dealt with.
Nuclear’s contribution is to lower the residual curve across the whole period rather than only at the peak. That reduces how much of everything else has to be built, and it does so without consuming land or network capacity at the same rate.
Where it gets commercially awkward
The honest difficulty is that a system with a lot of cheap renewables depresses wholesale prices in exactly the hours a nuclear plant is generating, which is most of them. That is not a reason to avoid nuclear; it is a reason to be clear that nuclear revenue in a high-renewables system cannot come solely from the energy market. Regulated asset base models, contracts for difference with appropriate strike structures, capacity mechanisms and direct corporate offtake all exist to address this, and the choice between them has a larger effect on delivered cost than most technology comparisons do.
There is a parallel difficulty for renewables, which is curtailment and negative pricing during high-output periods, and for storage, which is revenue stacking across markets that are not designed to be stacked. Every technology in a decarbonised system has a market design problem. Pretending that only nuclear has one is not analysis.
Flexibility, carefully
There is growing interest in nuclear plants operating flexibly, and in coupling reactors to hydrogen production, industrial heat or thermal storage so that output can be diverted rather than reduced. Load-following is well established in some reactor fleets; French units have done it for decades. Coupling to other products still requires design-specific assessment. Commercially it is harder, because reducing output on a capital-intensive asset reduces revenue without reducing cost. Coupling to a secondary product is more promising, provided the secondary market is real and the plant’s safety case supports the operating regime.
This is an area where early design decisions matter disproportionately. Building in the capability is relatively cheap; adding it to a licensed operating plant is not.
How to assess the combination
A credible assessment of a future system looks at whole-system cost rather than technology cost, uses weather years that include the difficult ones rather than the average, prices network reinforcement and land take explicitly, and treats inertia and system stability as requirements rather than assumptions. On that basis the answer is consistently a combination, with the proportions varying by geography, demand profile and the credibility of long-duration storage.
The argument worth having is about those proportions, the sequencing, and who carries which risk. The argument about whether nuclear or renewables is better has no answer because it is not a question about the same product.
NCL supports clients with whole-system analysis, investment cases and the regulatory pathway for firm low-carbon generation. Get in touch to discuss your assessment.