Hinkley Point C was supposed to cost £18 billion and generate power by 2025. In January 2024, EDF revised the estimate to £31-33 billion with first power pushed to 2031. By early 2026, estimates had risen further to approximately £35 billion, with independent analysts projecting figures closer to £40 billion. Meanwhile, Sizewell C received its Final Investment Decision in July 2025 with an initial budget of £20-22 billion for a near-identical EPR reactor design. The UK government has framed Sizewell C as a \"NOAK\" (Nth of a Kind) project that will benefit from lessons learned on Hinkley Point C, the \"FOAK\" (First of a Kind). But the most important lesson from Hinkley Point C is not about concrete or reactor design. It is about what happens when you approve a £20+ billion project without a rigorous quantitative schedule risk analysis.
QSRA for UK nuclear new build is the application of quantitative schedule risk analysis using Monte Carlo simulation to model the full range of schedule outcomes on nuclear construction projects. It replaces single-point completion dates with probability distributions that show the realistic range of when first power might be achieved, accounting for duration uncertainties, discrete risk events, regulatory hold points, and the correlation between activities across a decade-long construction programme.
The nuclear new build context creates schedule risk characteristics that do not exist on conventional power projects: regulatory hold points where the Office for Nuclear Regulation (ONR) must approve each construction phase before work proceeds, first-of-a-kind construction methods with no historical performance data, and a workforce that must be trained and security-cleared before entering the site. These factors make deterministic scheduling not just inadequate but dangerous, because it creates false confidence in a completion date that has no probabilistic basis.
This article examines how QSRA should be applied to Sizewell C using the lessons from Hinkley Point C’s schedule failure, and what a rigorous risk model would have shown before either project was approved.
What Went Wrong with Hinkley Point C’s Schedule?
Hinkley Point C’s schedule failure was predictable. The original 2025 completion date was based on a deterministic schedule that did not account for the systemic risks inherent in first-of-a-kind nuclear construction in the UK. A QSRA run at the Final Investment Decision stage would have shown that the 2025 date sat at approximately P5 on the S-curve, meaning a 5 percent chance of achievement. The P80 date would have been in the 2030-2032 range, which is exactly where the project now sits.
The key schedule risks that materialised on Hinkley Point C were known at FID. They were simply not quantified. For the broader QSRA methodology that would have captured these risks, see Schedule Risk Analysis (QSRA): Guide to Monte Carlo + Examples.
Nuclear-Specific Schedule Risk Drivers
Nuclear new build projects face schedule risk categories that conventional power projects do not encounter. These must be explicitly modelled in the QSRA for results to be meaningful.
Regulatory hold points require ONR approval before construction proceeds to the next phase. Each hold point introduces schedule uncertainty because approval timelines depend on the quality of safety documentation, the regulator’s workload, and the resolution of any technical issues raised during assessment. On Hinkley Point C, regulatory interactions added months to the programme that were not captured in the baseline schedule.
First-of-a-kind (FOAK) construction learning curve affects every activity on the first unit. Workers, supervisors, and engineers are performing tasks they have never done before, in configurations they have never built before. Productivity data from Flamanville (France) and Olkiluoto (Finland), the only other EPR builds, showed that FOAK labour productivity was 40-60 percent lower than planned. This must be modelled as a systematic productivity factor applied across all civil and mechanical activities.
Nuclear quality requirements mean that rework rates are significantly higher than on conventional construction. Welds must meet nuclear-grade standards, concrete pours require extensive quality documentation, and any defect triggers a formal non-conformance process. The schedule impact of rework must be modelled as a discrete risk event applied to quality-sensitive activities.
Workforce security clearance and training delays affect ramp-up. Every worker on a nuclear site must pass security vetting and complete nuclear safety training before accessing the construction area. Delays in the vetting pipeline directly reduce the available labour force during critical ramp-up phases.
FOAK vs NOAK: How Much Schedule Risk Does Sizewell C Actually Save?
The NOAK argument assumes that Sizewell C will benefit from the learning curve established on Hinkley Point C. This is partially true. Civil works productivity should improve because the same contractors will have completed the same concrete structures on HPC first. However, the NOAK benefit does not eliminate schedule risk. It narrows the distribution but does not shift it as dramatically as the government’s cost estimates imply.
| Risk Category | HPC (FOAK) Impact | SZC (NOAK) Expected Impact | NOAK Saving |
|---|---|---|---|
| FOAK learning curve | +24 months | +6 months | 18 months |
| Regulatory hold points | +12 months | +8 months | 4 months |
| Nuclear quality rework | +10 months | +5 months | 5 months |
| Supply chain and logistics | +8 months | +6 months | 2 months |
| Workforce ramp-up | +6 months | +3 months | 3 months |
NOAK Schedule Benefit = FOAK P80 Duration − NOAK P80 Duration
Sizewell C example: HPC P80 at ~15.5 years (FID to first power). SZC NOAK P80 at ~12.5 years. Net NOAK benefit = ~3 years, not the 5+ years implied by the budget comparison.
The critical insight is that NOAK reduces some risks but not all. Regulatory hold points, supply chain constraints, and site-specific ground conditions on the Suffolk coast are new risks that HPC experience does not mitigate. A rigorous QSRA for Sizewell C must model these residual risks independently, not simply apply a blanket NOAK discount. For how confidence levels should drive the FID decision, see P50, P80, P90 Confidence Levels in Risk Analysis.
Best Practices for QSRA on UK Nuclear New Build
First, model regulatory hold points as discrete risk events with variable duration, not as fixed milestones. ONR approval timelines depend on documentation quality and issue resolution. Using a triangular distribution (minimum 3 months, most likely 6 months, maximum 18 months) per major hold point captures this variability.
Second, use HPC actual performance data to calibrate SZC distributions. This is the true value of the FOAK/NOAK relationship: not a blanket discount, but measured data that replaces assumptions. Actual concrete pour rates, weld completion rates, and commissioning durations from HPC should be used to fit lognormal distributions for the corresponding SZC activities.
Third, apply strong positive correlation (0.6 to 0.8) across all nuclear-grade construction activities. If one concrete pour fails quality inspection and requires remediation, the same quality issues likely affect other pours. Without correlation, the model understates the compound effect of systemic quality problems.
Fourth, run the QSRA before the FID decision, not after. The purpose of QSRA is to inform the investment decision with probabilistic data. Running it after FID reduces it to a reporting exercise that cannot influence the decision it was meant to support. For how sensitivity analysis identifies which risks to mitigate first, see Sensitivity Analysis in Schedule Risk: Tornado Charts and Risk Drivers.
First, model regulatory hold points as discrete risk events with variable duration, not as fixed milestones. ONR approval timelines depend on documentation quality and issue resolution. Using a triangular distribution (minimum 3 months, most likely 6 months, maximum 18 months) per major hold point captures this variability.
Second, use HPC actual performance data to calibrate SZC distributions. This is the true value of the FOAK/NOAK relationship: not a blanket discount, but measured data that replaces assumptions. Actual concrete pour rates, weld completion rates, and commissioning durations from HPC should be used to fit lognormal distributions for the corresponding SZC activities.
Third, apply strong positive correlation (0.6 to 0.8) across all nuclear-grade construction activities. If one concrete pour fails quality inspection and requires remediation, the same quality issues likely affect other pours. Without correlation, the model understates the compound effect of systemic quality problems.
Fourth, run the QSRA before the FID decision, not after. The purpose of QSRA is to inform the investment decision with probabilistic data. Running it after FID reduces it to a reporting exercise that cannot influence the decision it was meant to support. For how sensitivity analysis identifies which risks to mitigate first, see Sensitivity Analysis in Schedule Risk: Tornado Charts and Risk Drivers.
Frequently Asked Questions
What would a QSRA have shown for Hinkley Point C at FID?
A rigorous QSRA would have shown the 2025 completion date at approximately P5 on the S-curve, meaning a 5 percent probability of achievement. The P80 date would have been in the 2030-2032 range, which is where the project now sits. This would have forced a realistic budget and timeline before the £18 billion commitment was made.
A rigorous QSRA would have shown the 2025 completion date at approximately P5 on the S-curve, meaning a 5 percent probability of achievement. The P80 date would have been in the 2030-2032 range, which is where the project now sits. This would have forced a realistic budget and timeline before the £18 billion commitment was made.
How does NOAK benefit affect Sizewell C’s schedule risk?
NOAK reduces FOAK learning curve risks significantly (saving approximately 18 months) and moderately reduces rework and workforce ramp-up risks. However, regulatory hold points, supply chain constraints, and site-specific risks on the Suffolk coast remain largely unmitigated by HPC experience. The net NOAK schedule benefit is approximately 3 years, not the 5+ years implied by budget comparisons.
NOAK reduces FOAK learning curve risks significantly (saving approximately 18 months) and moderately reduces rework and workforce ramp-up risks. However, regulatory hold points, supply chain constraints, and site-specific risks on the Suffolk coast remain largely unmitigated by HPC experience. The net NOAK schedule benefit is approximately 3 years, not the 5+ years implied by budget comparisons.
What are the unique schedule risks on UK nuclear new build?
Nuclear-specific risks include ONR regulatory hold points with variable approval timelines, first-of-a-kind construction learning curves, nuclear-grade quality requirements causing higher rework rates, and workforce security clearance and training delays that constrain labour ramp-up during critical construction phases.
Nuclear-specific risks include ONR regulatory hold points with variable approval timelines, first-of-a-kind construction learning curves, nuclear-grade quality requirements causing higher rework rates, and workforce security clearance and training delays that constrain labour ramp-up during critical construction phases.
How should Sizewell C use HPC data in its QSRA?
Use actual HPC performance data (concrete pour rates, weld completion rates, commissioning durations) to fit lognormal distributions for corresponding SZC activities. This replaces assumptions with measured data, which is the true value of the FOAK/NOAK relationship. Apply a NOAK adjustment factor to the fitted distributions rather than a blanket schedule discount.
Use actual HPC performance data (concrete pour rates, weld completion rates, commissioning durations) to fit lognormal distributions for corresponding SZC activities. This replaces assumptions with measured data, which is the true value of the FOAK/NOAK relationship. Apply a NOAK adjustment factor to the fitted distributions rather than a blanket schedule discount.
What is Sizewell C?
Sizewell C is a proposed 3.2 GW nuclear power station in Suffolk, UK, using a near-identical EPR reactor design to Hinkley Point C. It received its Final Investment Decision in July 2025 with an initial budget of £20-22 billion. The UK government frames it as a NOAK project benefiting from HPC lessons learned.
Sizewell C is a proposed 3.2 GW nuclear power station in Suffolk, UK, using a near-identical EPR reactor design to Hinkley Point C. It received its Final Investment Decision in July 2025 with an initial budget of £20-22 billion. The UK government frames it as a NOAK project benefiting from HPC lessons learned.
When should QSRA be run on a nuclear new build programme?
IQRM recommends running QSRA before the Final Investment Decision, not after. It should then be updated at each project gate: post-GDA, pre-construction, during civil works, and pre-commissioning. Each update incorporates actual performance data that improves model accuracy as the project progresses.
IQRM delivers specialist training and consulting in QSRA, Safran Risk, and risk-informed decision making for nuclear, power, and infrastructure mega-projects across the UK. Our QRM Diploma programme equips professionals with the practical skills to build and interpret quantitative risk models on complex programmes like Sizewell C and Hinkley Point C.
Learn more about the QRM Diploma →
Planning a QSRA for a UK nuclear or power generation project? IQRM provides Safran Risk model builds, FOAK/NOAK risk benchmarking, and expert risk analysis consulting for the UK energy sector.
Contact us at info@iqrm.net to request a consultation →
Written by Rami Salem, Quantitative Risk Management specialist, 15+ years in oil & gas, EPC/EPCM, and infrastructure projects.
IQRM recommends running QSRA before the Final Investment Decision, not after. It should then be updated at each project gate: post-GDA, pre-construction, during civil works, and pre-commissioning. Each update incorporates actual performance data that improves model accuracy as the project progresses.
IQRM delivers specialist training and consulting in QSRA, Safran Risk, and risk-informed decision making for nuclear, power, and infrastructure mega-projects across the UK. Our QRM Diploma programme equips professionals with the practical skills to build and interpret quantitative risk models on complex programmes like Sizewell C and Hinkley Point C.
Learn more about the QRM Diploma →
Planning a QSRA for a UK nuclear or power generation project? IQRM provides Safran Risk model builds, FOAK/NOAK risk benchmarking, and expert risk analysis consulting for the UK energy sector.
Contact us at info@iqrm.net to request a consultation →
Written by Rami Salem, Quantitative Risk Management specialist, 15+ years in oil & gas, EPC/EPCM, and infrastructure projects.
Establish Instructor’s Presence
