As the global race to harness artificial intelligence (AI) accelerates, the energy demands of AI systems are becoming increasingly apparent. The rapid expansion of AI technologies, from data centers to machine learning algorithms, is driving significant investments in energy infrastructure. Tech companies are looking to nuclear energy—both fission and the still-developing fusion technologies—to meet these demands, potentially reshaping global energy strategies while impacting net-zero goals.

The energy required to power AI systems is massive. Franklin Servan-Schreiber, CEO of nuclear energy startup Transmutex, highlighted that AI’s energy needs are so extensive that only nuclear energy, particularly nuclear fission, can reliably support the industry. This power source, which generates energy through nuclear reactions, is seen as cleaner than traditional fossil fuels, offering a more stable and efficient solution compared to renewable options like wind or solar power. The sheer scale of energy consumption by AI systems, such as those running data centers for major tech companies, makes nuclear energy a potentially indispensable part of the equation.

Nuclear fusion, an emerging technology that fuses atomic nuclei to release energy, holds even more promise. While still in its infancy, fusion energy could offer greater energy output with fewer greenhouse gas emissions and minimal radioactive waste. Silicon Valley investors are betting on nuclear fusion as the next step in the energy revolution, believing that it will be crucial for sustaining AI’s rapid growth in the long term.

However, the road to reliable nuclear energy for AI infrastructure is not without challenges. Currently, the United States operates only 54 nuclear plants, according to the U.S. Energy Information Administration. Developing a network of nuclear power plants capable of supplying the energy required by the AI revolution would require immense investment and long-term planning. Even with the efforts of companies like Amazon and Google, which have partnered with firms developing smaller, modular reactors, the funding available today is still far from sufficient. The billions of dollars needed to build a global nuclear infrastructure are currently seen as out of reach without strong government backing.

As tech companies race to meet AI’s energy demands, some are turning to natural gas as a short-term solution. Natural gas, while a fossil fuel, provides a more immediate and flexible energy source for AI data centers. Toby Rice, CEO of EQT, noted that tech companies, facing urgent energy needs, are asking how quickly natural gas can be made available to fuel their operations. This pragmatic shift highlights the challenge of balancing AI’s rapid growth with the long-term goal of achieving net-zero emissions.

The increasing energy consumption of AI systems has raised concerns about the tech industry's commitment to clean energy. At the COP29 climate summit in Baku, many big tech companies stayed out of the spotlight, opting not to display in the conference's green zone. This has led some observers to question whether the surge in energy consumption by AI data centers could undermine tech companies' green energy commitments.

According to McKinsey, data centers are expected to rise from 3% to 4% of U.S. energy demand to as much as 11% to 20% by 2030. This shift underscores the importance of finding sustainable energy solutions to power AI's expansion. Industry leaders, however, remain optimistic about the future. Nvidia CEO Jensen Huang expressed hope that the AI revolution will drive an energy revolution, suggesting that using energy to power intelligence could ultimately be one of the most efficient uses of energy available.

BUSINESSINSIDER