The Collapsing Narrative of AI-Driven Nuclear Energy Demand

Key Takeaways

• Efficient AI developments like DeepSeek have cast doubt on the projected surge in energy demand that utilities used to justify new nuclear investments.
• Energy and utility firms are struggling to defend stock valuations tied to the assumption that massive AI data center expansion is inevitable.
• The industry's reliance on 'nuclear renaissance' narratives mirrors historical patterns of overestimation that led to multi-billion dollar failures.

Historical Patterns of Failure

• During the 2000s, utility companies proposed over 30 nuclear reactors based on inflated demand projections; only four were built.
• The V.C. Summer nuclear project in South Carolina was abandoned after $9 billion in spending, fueled by erroneous forecasts of 22-30% energy sales growth that resulted in a 3% decline instead.
• Similar cycles of hyperbole have existed since the 1980s, when experts predicted a 'second nuclear era' that failed to materialize.

Current Market Disconnect

• Despite the cooling AI investment bubble, firms like Santee Cooper, Constellation Energy, and Duke Energy continue to pitch nuclear power as the solution for 'carbon-free' electricity.
• Building a new nuclear reactor requires at least a decade of lead time, making it an ineffective solution for the rapidly shifting needs of tech hyperscalers.
• Industry lobbyists and executives are accused of creating a 'moral panic' regarding energy shortages to suppress opposition and silence debate over the environmental and social costs of AI expansion.