Sun Xuan from University of Science and Technology of China: AI may "consume" one-fifth of the world's electricity

"I believe many people have heard this saying: The endpoint of AI is energy, and the endpoint of energy is fusion."

On September 11, at the 2025 Inclusion·Bund Conference, Sun Xuan, a professor at the School of Nuclear Science and Technology of the University of Science and Technology of China and founder and chairman of Xingneng Xuanguang, reiterated this important consensus in the industry during his speech—nuclear fusion is the key technology to unlock the next generation of civilization.

Sun Xuan pointed out that the rise of AI is exponentially increasing global energy consumption, and the answer to this ultimate demand is nuclear fusion. Once nuclear fusion is realized, it will not only bring about an energy revolution and trigger an industrial revolution but also be a crucial step towards utilizing the most abundant energy in the universe and advancing to a higher civilization.

Currently, the core driving force for moving fusion energy from the laboratory to industry comes from its largest demand side—AI. Sun Xuan noted: "AI currently accounts for 1.5% of the Earth's electricity consumption. If we liken AI to the 'Earth's brain,' then the human brain consumes 20% of the body's energy. Therefore, some predict that AI's electricity consumption will also account for over 20% of the Earth's total."

This means that just in the field of AI alone, there will be a huge energy gap.

Nuclear fusion is precisely the solution to meet the energy supply needed for the future development of AI technology. Fusion refers to the process of two light nuclei combining to form a heavy nucleus, during which there is a mass loss that can release enormous energy. Sun Xuan pointed out that nuclear fusion has an extremely high energy density: 1 gram of fusion fuel releases energy equivalent to 8 tons of oil.

In fact, capital with a keen sense for cutting-edge technology has already taken action, actively laying out this ultimate energy track. Sun Xuan noted that since 2020, investment in nuclear fusion has significantly increased, with leading international tech companies such as NVIDIA, Google, and OpenAI entering the nuclear fusion field, betting on this ultimate energy becoming an industry consensus.

However, ultimate energy comes with ultimate challenges, and there are still technical difficulties on the road to the "artificial sun."

Sun Xuan explained that the core scientific difficulty in achieving nuclear fusion lies in confining plasma at temperatures exceeding hundreds of millions of degrees, "it's a bit like trying to cage a temperamental wild beast with a cage, which is very difficult." Currently, the mainstream technologies pursuing controllable nuclear fusion are divided into two major directions: laser inertial confinement and magnetic confinement, both of which have extremely high engineering construction requirements. Whether it is developing precision giant lasers or constructing ITER-level behemoths, there are issues of high costs and long construction periods.

To address this, scientists have proposed a hybrid path called "magnetic inertial confinement," which can significantly reduce costs and construction time while improving iteration efficiency. However, this new fusion approach presents a "challenge to human intelligence," relying on human understanding of the underlying physical processes, which is precisely where AI technology can play a role.

In this regard, Sun Xuan proposed a groundbreaking vision: "Can we generate or create an AI that can learn on its own, without needing our existing experimental data, but instead exploring based on physical rules, just like the previous Go software AlphaZero, learning by itself and designing an entirely new fusion reactor?" With the assistance of AI technology, the field of nuclear fusion is seen as being on the brink of commercialization. According to a report released by the American Fusion Industry Association in July last year, global nuclear fusion companies have cumulatively received a total investment of $7.1 billion, an increase of $900 million year-on-year, with capital market financing reaching new highs; among the 35 companies surveyed, 89% are optimistic about achieving grid-connected power generation before the end of the 2030s.

China's nuclear fusion sector is also developing rapidly. Sun Xuan pointed out that China has evolved from the early single Tokamak fusion to various forms of fusion today, with many research institutes, universities, and enterprises making numerous breakthroughs in the field of nuclear fusion.

"The advancement of AI represents the evolution of human intelligence from carbon-based to silicon-based, while fusion signifies a transition from utilizing existing energy forms on Earth to cosmic energy forms. Achieving both can be said to herald the arrival of a new era, and if the two can work together, it may accelerate the arrival of this great era."

Sun Xuan candidly stated, "When the future arrives, we may not even realize it."