Harnessing Nuclear Power for the Development of Artificial Intelligence
By Matthew McIntosh
In a bid to meet the massive, unrelenting electricity needs of artificial intelligence (AI), Microsoft has announced a deal with Constellation Energy, the current owner of Three Mile Island, to restart Unit 1 by 2028. This move aims to power the technology giant's data centers.
However, the history of Three Mile Island is not without its dark chapters. On March 28, 1979, Unit 2 suffered a partial meltdown, an event that shook public confidence in the nuclear power industry. Grainy black-and-white images from cameras sent into the damaged reactor revealed that up to two-thirds of the reactor had been exposed, 45% of the uranium fuel had melted, and 19 tons of it had dropped to the reactor floor.
The 12-year cleanup of Three Mile Island cost $973 million. The accident exposed the faults in Eisenhower's careful distinction between the power plant and the bomb, and the nuclear industry never fully recovered.
Fast forward to 2020, Dr. Stephani Palazzo conducted an interview at the former visitor center of Three Mile Island nuclear power plant for her graduate research. The accident might provide a historical vantage to consider whether to power the future with nuclear energy, given the risks and uncertainties associated with it.
Nuclear power is proposed as a seemingly reliable, carbon-free energy source to support AI's voracious energy demands. Its stable, low-carbon baseload supply makes it an attractive option compared to intermittent renewables like solar or wind. AI's rapid growth in data centers drives unprecedented energy demand, and tech giants like Google, Amazon, Meta, and Microsoft are committing to nuclear power to secure long-term reliable supply.
However, the environmental and human costs of using nuclear power to support AI's soaring energy demands involve significant benefits and serious risks. Nuclear accidents like Three Mile Island underscore inherent risks of nuclear power. While modern reactors and AI-powered simulations aim to enhance safety, the risk of catastrophic failure exists.
Uranium mining and enrichment, necessary for nuclear fuel, releases toxic chemicals and radioactive materials, often impacting Indigenous communities disproportionately. Nuclear waste disposal remains an unresolved challenge, with long-lived radioactive waste requiring secure management to prevent environmental contamination and human health risks over millennia.
The push for nuclear-powered AI infrastructure necessitates balancing technological benefits with rigorous safety, environmental justice, and long-term waste management strategies. The U.S. is in the beginning stages of decommissioning roughly 20 American nuclear plants and disposing of over 90,000 metric tons of nuclear waste from commercial nuclear power reactors.
In conclusion, while nuclear power offers a relatively clean, reliable energy source critical for AI's exponential electricity demand, it carries significant environmental and human costs. The accident at Three Mile Island serves as a symbol of the U.S. nuclear industry's overconfidence in its early expertise and its ultimate failure to deliver on its promises. The revival of Three Mile Island to power AI data centers by 2028 underscores the ongoing debate about the role of nuclear power in our energy future.
[1] Nuclear power and AI: A critical analysis [2] Small Modular Reactors: The Future of Nuclear Power [3] Microsoft's Three Mile Island deal: A step towards carbon-free AI [4] The environmental and human costs of nuclear power [5] AI's Energy Demands and the Role of Nuclear Power
- The revival of Three Mile Island to power AI data centers by 2028 has sparked a critical analysis of the role of nuclear power in supporting AI's energy demands.
- Small Modular Reactors, a potential future of nuclear power, aim to enhance safety and reduce the environmental footprint of nuclear energy.
- The deal between Microsoft and Constellation Energy to restart Three Mile Island's Unit 1 for AI's power needs signifies a step towards carbon-free AI, but it also raises concerns about the associated environmental and human costs.
- AI's rapid growth is driving unprecedented energy demand, and the environmental and human costs of using nuclear power to meet those demands involve significant benefits and serious risks.
