“The application of AI for Science these years, along with the use of new materials, especially high-temperature superconductors in the field of fusion, has made fusion no longer seem distant,” said Chen Rui, CEO of Startorus Fusion, to the attendees.
On June 19, Chen Rui participated in Unity Ventures' 12th-anniversary annual meeting, where he introduced the commercialization of controlled nuclear fusion to investors, company founders, and industry experts present. He noted that for a long time, the R&D of fusion energy has been viewed by many as an exclusive task for the "national team", and the notion of “always 50 years away” has seemed like an unreachable scientific dream. However, an increasing number of commercial fusion companies worldwide are accelerating the R&D process of fusion energy.
Taking Startorus Fusion as an example, the company's SUNIST-2 device took only 279 days from design and construction to igniting the first plasma, making it the first commercial company in China to operate such a device. During this period, the company independently developed all systems, including the power, diagnostic, integrated control, and preprocessing systems. The performance parameters of the device are also continuously and rapidly improving; the plasma current has been raised to the design level, and stable operation of repetitive reconnection has been achieved.
The productivity transformation brought by AI technology has accelerated the pace of fusion energy R&D. Through the analysis and learning of massive data generated during the fusion reaction process, AI can reveal the intrinsic laws of plasma movement and predict the progress and outcomes of reactions. It can also enhance plasma control during fusion reactions, ensuring stability and safety. Furthermore, the continuous improvement of AI computing power requires a constant supply of fusion energy; computing power equates to electrical power. With the rising demand for computing power from large models and applications, addressing the growing energy needs of AI is an important pathway through nuclear fusion.
Chen Rui elaborated further on this point. He indicated that with the support of AI technology, the R&D progress of Startorus Fusion is advancing rapidly. The magnetic field strength and plasma parameter levels of the SUNIST-2 device currently rank first in the spherical tokamak field in China and third globally. Startorus Fusion has already begun constructing the next-generation device, which will heat plasma to 100 million degrees Celsius, achieving all parameters needed for fusion conditions and realizing an equivalent Q value greater than 10, thoroughly validating the engineering feasibility of fusion.
Finally, Chen Rui noted that fusion is a highly complex industrial chain encompassing traditional engineering fields such as materials, machinery, electricity, and electronics, as well as complex AI for Science. This year's Unity Ventures annual meeting, themed "Immortal Computing," explores the significance of AI computing for societal and industry development. He expressed hope for opportunities to collaborate with Unity Ventures' portfolio companies to jointly explore the prospects of deep integration between controlled nuclear fusion and AI technology.
