Investors are closely watching the battle over computer chips as China and the United States vie for global leadership in artificial intelligence — but Wall Street may be looking in the wrong direction. The real winner will depend on a far less dramatic sector: energy. Specifically, nuclear energy. After all, no matter how advanced the artificial intelligence chip is, it will eventually become useless if it cannot be powered. Nuclear energy is quickly becoming one of the core solutions to support the artificial intelligence boom and protect against potential energy shortages.
nuclear energy option
Richard Windsor, founder of research firm Radio Free Mobile, said: "I think there is no doubt that one of the best sources of power for data centers must be nuclear power. If you want to obtain stable and zero-carbon power, nuclear power is almost the only option."
By 2026, the energy consumption of data centers powering artificial intelligence worldwide could exceed the national energy consumption of all of Japan, according to Bank of America data. Research firm Bloomberg New Energy Finance points out that power flow in global power grids is expected to surge by 30% as soon as 2030. Companies such as cloud infrastructure provider CoreWeave Inc. have identified energy as a key bottleneck in the development of artificial intelligence.
Felix Chen, a research strategist at Bank of America, emphasized that nuclear energy is becoming a "necessity" for the development of artificial intelligence in the future due to its advantages such as low life cycle costs, low carbon emissions, and stable baseload power supply. Windsor also highlighted that nuclear energy can provide 24-hour continuous power, which traditional renewable energy sources such as wind and solar cannot guarantee.
Today, most nuclear power plants use nuclear fission technology, which releases heat by splitting heavy atoms such as uranium to generate electricity. China and Russia have put into operation some small modular reactors, which represent the next generation of more advanced miniaturized fission technology. But Bank of America believes the real "holy grail" will be nuclear fusion - which releases huge amounts of energy by fusing light atoms such as hydrogen at extremely high temperatures.
"Build it, build it quickly!"
The U.S. government is acutely aware of opportunities in nuclear energy and has a growing sense of urgency to close a technological development gap with China that is said to be 10 to 15 years, the U.S. Department of Energy said.
Over the past decade, overall energy capacity growth in the United States has stagnated; China has added grid capacity equal to the entire United States' existing electricity capacity every 18 months. Felix Chen pointed out that at the current rate, China's nuclear energy production capacity is expected to surpass that of the United States in 2030.
In May 2024, President Donald Trump signed four executive orders aimed at promoting investment in nuclear energy. The "Trump Artificial Intelligence Action Plan" released in July 2025 emphasizes the need to vigorously develop artificial intelligence infrastructure - especially nuclear power generation capacity, and calls on the country to "Build, baby, build!" (Build, baby, build!). In addition, nuclear energy is also one of the core topics during Trump's visit to the UK this week. He and British Prime Minister Keir Starmer signed a technology cooperation agreement involving investment in artificial intelligence and nuclear energy.
Sean Farney, vice president of data center strategy for Jones Lang LaSalle (JLL) in the Americas, pointed out that digital infrastructure has been deeply bound to economic and military power, which has led to increasingly coordinated initiatives between governments and enterprises in the field of nuclear energy. Large technology companies are continuously investing resources in the research and development of nuclear energy technology: Nvidia, Microsoft, Metaverse Platform, Amazon, and Alphabet have all deployed nuclear energy investments, striving to seize megawatt-level power resources in the race to develop the most advanced artificial intelligence models.
Can nuclear energy be delivered on schedule?
Building nuclear energy infrastructure is no easy task. Although small modular reactors (SMRs) are cheaper and faster to build than traditional power plants, fusion technology is still in its infancy.
Benjamin Lee, a professor of engineering and computer science at the University of Pennsylvania, believes that nuclear energy technology has potential, but also admits that there are practical constraints.
"Nuclear energy is not a short-term solution, and it may not even be a mid-term solution," Li said. "Currently we do not have a mature industrial system that can ensure that nuclear power plants are completed on time and on budget."
In contrast, Felix Chen pointed out that China can build a nuclear power plant in only 3 to 5 years.
Today, strict regulatory policies, high upfront costs, and the loss of technical talent due to long-term stagnation in nuclear power plant construction all pose challenges to the development of nuclear energy in the United States. But Farney said that given that the Trump administration has been pushing for deregulation in the energy sector, the urgency of the AI race may help lower the "extremely high regulatory threshold." He predicts that with more government support, the time to deploy small modular reactors in the United States could be shortened from seven years to three years.
Li also mentioned that the industry is ushering in a new wave of private investment as big technology companies invest part of their huge capital expenditures in nuclear energy, which is expected to accelerate technological development. He said the key is to "obtain enough orders for a new generation of nuclear power generation projects" because only through repeated construction can the United States rebuild its technological strength in the field of nuclear energy.
The stakes are high — the financial future of big tech companies and U.S. national security are tied to it. But Farney is optimistic about a breakthrough in the energy industry.
“There’s huge amounts of money involved,” Farney said, “but also extraordinary innovation.”