September 25, 2025
With the S&P 500 recently pulling back from fresh all-time highs, the motivating narrative of the current long-term bull market remains centered around the potential of AI to revolutionize commerce and life to be more efficient. Optimism and pessimism around that narrative along with more mundane concerns like the actual levels of economic growth and inflation have shaped the path of equity prices since the beginning of 2023. The most optimistic narratives have assumed that AI would be integrated into and utilized by a broad mix of businesses to increase efficiency and growth, but adoption and results haven’t been linear and many stocks are struggling to outperform given the enthusiasm for AI pure plays and the projection of systemic need for the technology.
One thing we are noticing is that AI has lifted a few other industries due to their potential utility in its rapidly expanding infrastructure and operating needs. Nuclear power has been a hot theme in this regard along with autonomous technology and electrification. In this week’s thematic column, we’re taking a bit of a deep dive into the relationship between nuclear stocks and AI. What we found was a nascent symbiotic relationship where each technology provides potential solutions for the other as they develop and integrate into the mainstream.
Artificial intelligence (AI) and nuclear power are increasingly viewed as complementary technologies, connected through both practical applications and strategic importance. On one hand, AI is reshaping how nuclear plants are designed, operated, and maintained. On the other, nuclear energy is emerging as a critical solution to meet the surging electricity demand driven by AI and hyperscale computing. Together, they illustrate the intersection of digital and energy transitions in the mid-2020s.
AI is already transforming nuclear power operations. Nuclear reactors produce vast quantities of sensor data, and AI systems are being deployed to analyze these streams in real time. This enables early detection of anomalies, predictive maintenance, and better safety margins. Predictive algorithms can anticipate failures in turbines, pumps, or other critical components, reducing downtime and operational risk. In research settings, AI accelerates complex simulations of nuclear reactions and materials performance, cutting years off traditional design cycles for advanced nuclear systems. A notable example is DeepMind’s work with the Swiss Plasma Center, where reinforcement learning was used to optimize plasma confinement in a tokamak fusion reactor — a task previously thought to require decades of engineering refinements.
The connection also runs in the other direction: nuclear power is increasingly seen as essential to enabling AI itself. Data centers are among the fastest-growing sources of electricity demand. According to the International Energy Agency (IEA), global data centers consumed about 460 terawatt-hours (TWh) in 2022, roughly 2% of world electricity. With the rapid expansion of AI workloads, the IEA projects this figure could double by 2026. Goldman Sachs has estimated that AI alone could account for 8% of U.S. electricity demand by 2030. Meeting this demand sustainably requires carbon-free baseload energy, and nuclear power — unlike intermittent renewables — provides stable 24/7 output.
This reality is prompting partnerships between technology companies and nuclear developers. Microsoft, Amazon, and Google have all explored nuclear energy as part of their long-term decarbonization and energy security strategies. In 2023, Microsoft signed a deal with nuclear developer Helion to purchase fusion power starting in 2028, while Amazon and Google have been in discussions with small modular reactor (SMR) providers to secure future clean energy for their data centers. These deals underscore the role of nuclear power in supporting the infrastructure backbone of the AI revolution.
Stock level performance is confirming investor awareness of these growing co-dependencies. AI hyperscalers and their resource needs are clear growth opportunities in the present cycle. Some key charts are below.
NVDA
GOOG/L
SMR
OKLO
MIR
The connection extends into nuclear research and development. AI is accelerating materials discovery by modeling how radiation affects alloys and composites, which is critical for next-generation fission and fusion reactors. Generative AI tools are being applied to optimize reactor components, improving heat flows, materials efficiency, and safety features. This integration of AI into nuclear innovation reduces costs and time horizons, helping overcome one of the sector’s traditional barriers to growth.
Finally, there is a strategic and policy dimension. Both AI and nuclear are viewed as “strategic technologies” in the U.S., China, and Europe. Policymakers link them as pillars of competitiveness, national security, and decarbonization. AI is seen as a key tool to manage increasingly complex energy systems, while nuclear provides the reliable, carbon-free electricity needed to power digital infrastructure. Together, they embody what analysts sometimes call the “twin transformations” of the 21st century: the digital revolution and the clean energy transition.
Conclusion
In sum, the relationship between AI and nuclear power is mutually reinforcing. AI makes nuclear safer, more efficient, and faster to innovate, while nuclear offers the clean, stable energy foundation required for AI’s exponential growth. Far from being unrelated, these two technologies are becoming increasingly intertwined at the core of global economic and technological strategy. Right now nuclear stocks are still in the Small/Mid Cap. benchmarks and legacy Energy and Utilities still dominate the composition of Large Cap. sectors. Adding nuclear stocks to Energy sleeves has been an excellent way to add some higher beta exposure with positive momentum in 2025. Investors should look for accumulation opportunities on pullbacks from recent overbought conditions in many of these stocks.
Bibliography
-
DeepMind & Swiss Plasma Center. (2022). Magnetic control of tokamak plasmas through deep reinforcement learning. Nature, 602, 414–419. https://doi.org/10.1038/s41586-021-04301-9
-
International Energy Agency (IEA). (2024). Electricity 2024: Analysis and forecast to 2026. Paris: IEA. Retrieved from https://www.iea.org/reports/electricity-2024
-
Goldman Sachs Research. (2023). The AI power demand surge. Cited in Bloomberg and Financial Times coverage.
-
Microsoft. (2023, May). Microsoft signs agreement with Helion Energy to develop world’s first fusion power purchase. Press release. Retrieved from https://blogs.microsoft.com
-
Reuters. (2023, May 10). Microsoft signs deal to buy power from nuclear fusion startup Helion. Retrieved from https://www.reuters.com
-
CNBC. (2023, Dec). Amazon and Google explore nuclear power deals to fuel AI data centers. Retrieved from https://www.cnbc.com
-
OECD Nuclear Energy Agency (NEA). (2022). Nuclear Energy and the Paris Agreement. Paris: OECD. Retrieved from https://www.oecd-nea.org
-
International Atomic Energy Agency (IAEA). (2021). Artificial Intelligence for Nuclear Applications. Vienna: IAEA. Retrieved from https://www.iaea.org
Data sourced from Factset Research Systems Inc.
