Nuclear Energy Australia: A Critical Analysis of GenCost 2024-25

By Nick Hurley, 10th December 2024

The latest edition of CSIRO’s GenCost report, produced annually in collaboration with the Australian Energy Markey Operator (AEMO) was released yesterday. It has attracted an increased degree of scrutiny, particularly due to the ongoing nuclear energy ban in Australia. The article analysis some of the report’s shortcomings.

Key Takeaways

The GenCost 2024-25 report represents Australia’s primary reference for electricity generation technology costs and projections. While it provides valuable insights for policy makers and investors, several significant limitations and gaps affect its utility for long-term planning and decision-making.

This analysis identifies critical shortcomings in methodology, scenario planning, and key assumptions. Of particular concern are the limited treatment of extreme weather events, emerging demand drivers, and development challenges. The study’s handling of integration costs and backup generation requirements raises questions about the projected costs of high renewable scenarios.

Our review suggests that while GenCost offers a solid foundation for technology cost comparison, its findings should be interpreted with caution, particularly regarding long-term system reliability and integration costs. Decision-makers need to consider these limitations when using the report for policy development or investment planning.

1. Methodological Limitations

1.1 Modeling Framework

The study’s heavy reliance on the Global and Local Learning Model (GALLM) represents both a strength and a limitation. While GALLM provides a structured approach to projecting technology costs based on deployment rates and learning curves, the model’s impact on findings lacks sufficient sensitivity analysis. The study’s use of a single 5.99% discount rate across all technologies oversimplifies the varying risk profiles and financing structures of different generation technologies. Being relatively high, it naturally favours the business cases for faster to deploy technologies such as batteries.

The model’s projection of technology learning rates often lacks empirical validation, particularly for emerging technologies like nuclear SMR. The absence of detailed sensitivity analysis around these learning assumptions creates uncertainty about the robustness of cost projections.

1.2 Data Quality Issues

A fundamental challenge facing the study is its reliance on varied data sources of differing quality. For established technologies deployed in Australia, cost estimates benefit from direct market evidence. However, for technologies not yet deployed domestically, the study must extrapolate from international data, introducing significant uncertainty.

The acknowledged uncertainty range of ±30% in capital cost estimates, while transparent, raises questions about the reliability of comparative cost analysis. This becomes particularly problematic when comparing technologies with similar projected costs, as the uncertainty ranges often overlap significantly.

The study’s treatment of first-of-a-kind (FOAK) costs remains problematic. While acknowledging potential cost premiums of up to 100%, the study doesn’t adequately incorporate these premiums into its analysis, potentially understating the real costs of deploying new technologies in Australia.

2. Critical Gaps in Scenario Analysis

2.1 Demand Forecasting

A significant oversight in the study’s scenario planning is the limited consideration of emerging demand drivers. Notable among these is the potential surge in electricity demand from AI data centers, which could significantly impact system requirements and technology choices.

The study’s demand scenarios, while covering different levels of electrification, may underestimate the potential for step changes in demand patterns. This becomes particularly relevant when considering the rapid electrification of transport and industry, combined with emerging technologies.

2.2 Weather and System Reliability

Perhaps the most critical limitation is the study’s treatment of extreme weather events and system reliability. While the analysis includes nine historical weather years, it lacks explicit modeling of extended periods of low renewable generation across large geographic areas – for instance, a week-long period of low wind and solar resources across Australia’s east coast.

The study’s integration cost analysis reveals significant reliance on fossil fuel backup generation, with approximately 45% of dispatchable capacity in worst-case scenarios depending on fossil fuels. This raises questions about the true costs and feasibility of achieving high renewable penetration while maintaining system reliability.

3. Development and Implementation Challenges

3.1 Project Timelines

While the study provides a detailed analysis of nuclear development timeframes, supporting its minimum 15-year estimate with evidence from democratic nations, it falls short in addressing growing challenges in renewable project development. The increasing complexity of planning approvals for wind farms, particularly in relation to community consultation and environmental impact assessments, is not fully captured in the cost and timeline projections.

The study’s assumption that wind projects can recover from current cost pressures by 2035 may be optimistic given mounting development challenges. The compound effect of planning delays, grid connection complications, and community opposition could extend development timeframes beyond current projections.

3.2 Integration Costs

The integration cost analysis is limited to 90% VRE penetration, leaving uncertainty about costs beyond this level. The study’s treatment of transmission infrastructure requirements, while including committed projects, may underestimate the complexity and cost of future network augmentation needs.

System security considerations, particularly the requirement for grid-forming inverters and synchronous condensers, are included but their cost implications over time could be more thoroughly explored. The study’s assumption about the availability of storage duration options may not fully reflect market constraints and cost implications.

4. Economic and Market Considerations

4.1 Cost Assumptions

The study’s handling of first-of-a-kind costs represents a significant limitation. While acknowledging potential cost premiums up to 100%, these are not explicitly incorporated into the analysis. The treatment of land costs, while including an escalation factor, may underestimate the impact of increasing competition for suitable sites.

The financing assumptions, particularly the use of a single discount rate (5.99%), oversimplify the varying risk profiles and financing structures across technologies. This could particularly impact the comparative analysis of technologies with different risk profiles and development timeframes.

4.2 Market Dynamics

The study’s treatment of supply chain constraints and their resolution by 2027-2030 (depending on scenario) may be optimistic. Recent experience with wind turbine manufacturing suggests more persistent challenges. The analysis of competition for sites and resources could be more comprehensive, particularly regarding the cumulative impact of multiple projects seeking similar locations and resources.

5. Policy Implications

5.1 Near-term Policy Decisions

The study’s findings have significant implications for near-term policy decisions, particularly regarding reliability standards and investment signals. The projected cost differentials between technologies influence policy choices, yet the uncertainties in these projections may not be sufficiently emphasised in policy discussions.

5.2 Long-term Planning

The study’s projections to 2055 provide valuable insights for long-term planning but contain increasing uncertainty over time. The assumption of continued technology learning rates may not fully capture potential disruptions or plateaus in cost reductions. The interplay between technology choices and infrastructure requirements could be more thoroughly explored.

6. Recommendations

6.1 Methodology Improvements

We believe that the GenCost study requires several key methodological enhancements to improve its utility for decision-makers:

Sensitivity Analysis

• Implement comprehensive sensitivity analysis of GALLM assumptions, particularly learning rates
• Develop probability distributions for key input variables rather than simple ranges
• Test robustness of findings against multiple demand and technology development scenarios

Cost Assessment Framework

• Develop a more sophisticated framework for assessing first-of-a-kind costs in the Australian context
• Create a structured approach to incorporating planning and development timeframes
• Establish improved methodologies for assessing integration costs beyond 90% VRE

Weather and Reliability Modelling

• Include specific modelling of extended low renewable generation events such as several days of limited sun and wind along the east coast.
• Incorporate correlation analysis of weather patterns across different regions
• Develop more detailed storage requirement analysis for different reliability scenarios

6.2 Additional Research Requirements

Several critical areas require dedicated research to complement the GenCost analysis:

System Reliability Studies

• Detailed modeling of system response to extended low renewable generation periods
• Analysis of storage requirements under various weather scenarios
• Assessment of backup generation requirements and associated costs

Emerging Trends Analysis

• Investigation of AI data center demand implications
• Assessment of electrification trends and their impact on demand patterns
• Analysis of supply chain constraints and their long-term implications

Development Constraints

• Detailed study of planning and approval timeframes for renewable projects
• Assessment of land availability constraints and competing land uses
• Analysis of transmission corridor availability and development challenges
• Investigation of community acceptance issues and their impact on project timelines

Market Structure Analysis

• Evaluation of market design implications for different technology mixes
• Assessment of financing structures and their impact on technology costs
• Analysis of competition effects on resource availability and costs

7. Conclusions

The GenCost 2024-25 study provides valuable insights into technology costs but contains significant limitations that affect its utility for policy and investment decisions. While its methodology is generally sound, the study would benefit from more comprehensive treatment of extreme events, emerging demands, and development challenges.

The study’s findings regarding renewable integration costs and backup requirements deserve particular scrutiny, especially given the limited analysis of extended low renewable generation periods. The growing challenges in project development, particularly for wind farms, warrant more detailed consideration.

Decision-makers should view the study’s projections as indicative rather than definitive, particularly for longer-term planning. Additional research focusing on system reliability, integration costs, and development challenges would complement the study’s technology cost focus and provide a more complete basis for policy and investment decisions.