Commercialization Still Far Off... Nvidia Invests in Fusion Power

Nvidia Leads AI Innovation in Fusion Energy
Big Tech Races to Solve Future Power Shortages

With NVentures, the venture investment arm of Nvidia, investing in Commonwealth Fusion Systems (CFS), there is growing interest in fusion power-related stocks in the Korean stock market. Although there are still many technical challenges to overcome and commercialization is expected to take considerable time, there is increasing anticipation that the use of artificial intelligence (AI) could help reduce trial and error.

According to the investment banking industry on September 3, CFS, headquartered in Massachusetts, recently raised a total of $863 million in funding. The funds will be used to complete the SPARC fusion demonstration device project and to develop ARC, the first commercial power plant being built in Virginia, United States.

CFS aims to develop ARC, the world's first commercial fusion power plant, by 2030. While it will take at least five years to determine whether the project will succeed, Nvidia's investment is fueling hopes for early commercialization. On August 22, Jensen Huang, CEO of Nvidia, who visited Taiwan, stated, "I sincerely hope that various forms of energy will be developed for the advancement of the AI industry."

In addition to Nvidia, Counterpoint Global under Morgan Stanley, former Duquesne Capital Chairman Stanley Druckenmiller, and others have also invested in CFS. A consortium of 12 Japanese companies led by Mitsui & Co. and Mitsubishi Corporation also participated. Among existing investors, Google and former Google CEO Eric Schmidt made additional investments.

Former CEO Schmidt pointed out, "The natural limit of AI is not the chip, but electricity," adding, "The United States will need an additional 92 gigawatts (GW) of power to support AI innovation." In June this year, Google signed a purchase agreement with CFS to secure 200 megawatts (MW) of power supply.

The fact that global big tech companies such as Nvidia and Google are rushing to invest in fusion power indicates that the looming power shortage is a serious issue. While the U.S. government is expediting nuclear power development, nuclear power alone will not be sufficient. In addition to the power shortages at data centers resulting from advances in AI technology, the increase in power consumption due to the commercialization of humanoids must also be considered.

The importance of securing a lead in fusion power technology is growing. John Smith, Senior Director at General Atomics, which is developing fusion technology, said, "The country that achieves fusion first will have the power to shape the future." General Atomics recently developed the Central Solenoid module, the largest and most powerful pulsed superconducting magnet in history, which will be supplied as a core component for driving fusion reactions at the International Thermonuclear Experimental Reactor (ITER).

As AI technology advances, it is expected that technical challenges to commercializing fusion power will be resolved more rapidly. Simulation-based automated design optimization can dramatically reduce the time and cost required for commercialization. If Nvidia invests in fusion startups and provides AI technology, it can maintain strong influence even in the era of fusion energy.

As the commercialization of fusion power approaches, there is growing interest in domestic companies such as Iljin Power, Vitzro Tech, Mobis, and KNR Systems. Korea's fusion power technology is globally advanced. The achievement of the Korea Superconducting Tokamak Advanced Research (KSTAR) device in maintaining high-temperature plasma above 100 million degrees Celsius for over 30 seconds has attracted international attention.

Im Sangkuk, a researcher at KB Securities, stated, "Given that stable power supply is essential for AI growth, investments in fusion power by global big tech companies such as Nvidia and related firms are expected to drive the development of a new axis of power energy." He added, "A re-evaluation across the entire related value chain is expected. However, given the uncertainties surrounding development time, costs, and the likelihood of success in fusion power commercialization, continuous monitoring is necessary."