Nuclear Power: Fission Or Fusion?

Introduction

Nuclear energy is regaining investor and media attention, as more than 20 countries have pledged to triple their nuclear capacity by 2050.¹ The boom in AI data centers is incentivizing tech giants like Google,² Amazon,³ Meta,⁴ OpenAI,⁵ and Microsoft to explore the possibility of nuclear power as a clean source of energy. On the heels of a Power Purchase Agreement (PPA) with Helion, a private fusion company,⁶ Microsoft recently announced another PPA, this one for 20 years, to restore the Three Mile Island Unit 1 fission nuclear facility.⁷

The two basic forms of nuclear energy are fission and fusion, as illustrated below. Based on ARK’s research, fusion still is facing significant challenges, making fission—which has been powering the US grid since the 1950s—the more viable alternative.⁸

Source: ARK Investment Management, LLC, 2024. For illustrative purposes only.

Relative to fission, fusion energy has been heralded for its cleanliness, safety, and higher energy density, but it also has been “30 years away” for the past ~40 years, as shown below.⁹

Source: ARK Investment Management, LLC, 2024 based on data from Takeda et al 2023.¹⁰ Forecasts are inherently limited and cannot be relied upon. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

According to ARK’s research, fusion will have to achieve three milestones before hitting prime time: scientific breakeven, engineering breakeven, and commercial breakeven, as illustrated below.

Source: ARK Investment Management, LLC, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Scientific And Engineering Breakeven

Scientific breakeven—also known as "Q>1”—is the ratio of power output to the power injected into the fuel to generate that output. In December 2022, the National Ignition Facility (NIF) achieved the first scientific breakeven in nuclear fusion. Deploying laser-based inertial confinement,¹¹ it achieved a Q of ~1.54, later increasing to ~2.36.¹² Although they did hit a major scientific milestone, the fusion reactions lasted only a few billionths of a second and lost ~99% of the energy required to cause the reaction,¹³ as shown below.¹⁴ Experts agree that this fusion method is unlikely to achieve the efficiency necessary for commercial viability.¹⁵

Source: ARK Investment Management, LLC, 2024, based on data from Osolin 2024.¹⁶ For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 

An approach that could become commercially viable, at least theoretically, is magnetic confinement fusion (MCF).¹⁷ Despite ~75 years of research, however, it has not achieved scientific breakeven (Q>1), as shown below.¹⁸ The current record is a Q of ~0.67,¹⁹ set by the Joint European Torus (JET) in 1997. Prospectively, the large scale International Thermonuclear Experimental Reactor (ITER) project is aiming for a Q of 10 by 2039.²⁰ That said, while the NIF has achieved scientific breakeven, no project has reached engineering breakeven.

*Excluding Laser-Based Inertial Confinement. Note: The "Triple Product" displayed on the Y axis above is a single metric that measures nuclear fusion progress by considering density, confinement time, and temperature.²¹ Source: ARK Investment Management, LLC, 2024, based on data from Wurzel and Hsu 2022.²² Forecasts are inherently limited and cannot be relied upon. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Commercial Breakeven

Even if both the scientific and engineering challenges were to be resolved, commercial breakeven would prove dauntingly difficult for several reasons.

First, a key fusion fuel—tritium—is almost nonexistent in nature and is primarily the by-product of nuclear fission in heavy water reactors.²³ In fact, based on regulatory and technology forces, the heavy water reactors and associated tritium are likely to phase out of existence by 2060.²⁴ While fusion reactors also can generate tritium during operations, they require some supply to start up and are unlikely to produce enough to be self-sustaining.²⁵ Another regulatory consideration is that the Nuclear Regulatory Commission (NRC) has yet to finalize guidance for nuclear fusion reactors,²⁶ and current research suggests that fusion is likely to be uncompetitive with existing energy sources based on costs.²⁷

Despite those hurdles, ~45 private fusion companies are pursuing the opportunity.²⁸ Among them, Commonwealth Fusion Systems (CFS) is the most promising. With its SPARC reactor, CFS aims to achieve Q>1 in ~2027,²⁹ more than a decade ahead of ITER. High-temperature superconducting magnets could enable SPARC to match ITER’s performance at one 40th of the volume.³⁰

In our view, engineering breakeven is the crucial variable to monitor at this time. While private company pursuit of scientific breakeven is exciting, the history of fission suggests that commercializing a fusion plant could take another ~15 years, as shown below. 

Source: ARK Investment Management LLC, 2024. This ARK analysis draws on a range of external data sources, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

That said, because of lower proliferation risks, fusion could commercialize faster than fission, especially given more rapid research breakthroughs and more streamlined regulations.³¹

Conclusion

Given the considerable scientific, engineering, and commercial hurdles facing nuclear fusion, fission seems to offer a more promising nuclear solution to the world’s burgeoning power needs. Having powered the US grid reliably since the 1950s, nuclear fission is likely to be the nuclear technology that could help propel the global economy toward net-zero emissions by 2050. Of course, we will continue to monitor fusion's progress toward the same goal in the meantime.