Can The US Triple Its Nuclear Capacity By 2050?

The nuclear industry is transforming rapidly as the US and 30 other nations commit to tripling nuclear capacity by 2050.¹ Hyperscalers like Microsoft, Google, and Amazon are catching the nuclear energy wave in an attempt to meet the rising power demands associated with Artificial Intelligence (AI).² Commitments to net-zero carbon emissions have spurred new and renewed interest in nuclear energy, as shown below. ARK’s research suggests that the US must deploy microreactors, Small Modular Reactors (SMRs), and large-scale reactors to triple nuclear capacity by 2050. 

Source: ARK Investment Management LLC, 2025, based on data from Capital IQ Pro as of December 18, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 

For fifty years, excessive regulation has hampered the development of nuclear energy, leading to cost overruns and delays. Our research shows that regulatory hurdles instituted during the 1970s³ derailed the decline in US nuclear construction costs projected by Wright’s Law.⁴ That regulatory derailment has deprived the US of its share of global nuclear construction projects—with none currently underway—while countries like China are expanding their nuclear capacity aggressively, as shown below. 

Source: ARK Investment Management LLC, 2025, based on data from PRIS 2024.⁵ For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 

Today, the US nuclear industry is beginning to revive and could be on the threshold of a renaissance. In July 2024, the ADVANCE Act directed the Nuclear Regulatory Commission (NRC) to streamline licensing,⁶ and now the Department of Energy (DoE) is investing billions in nuclear energy. Meanwhile, as AI boosts demand for power, hyperscalers like Meta Platforms, Alphabet, Amazon, and Microsoft seem willing to pay higher rates for power purchase agreements (PPAs)⁷ to fund first-of-a-kind (FOAK) designs. As risk capital encourages new designs and regulators lower hurdles to innovation, Wright’s Law suggests that the cost competitiveness of nuclear power will be restored, ultimately lowering electricity prices and reigniting the powerful trends aborted by regulation in 1974.⁸

To triple capacity by 2050, the US must add ~200 gigawatts (GW) of nuclear power. By 2035, the US plans to deploy 35GW and then add 15GW per year by 2040.⁹ Some older reactors may restart,¹⁰ but advanced reactors are likely to be the primary path to expansion, as shown below. Moreover, the US must invest heavily in the nuclear supply chain to reduce reliance on Russian imports.¹¹

Note: 2024 capacity assumes that Diablo Canyon remains operational, and all existing fleets renew licenses where applicable. Source: ARK Investment Management LLC, 2025, based on data from IAEA PRIS Power Reactor Information System 2025 and World Nuclear News 2024, 2025.¹² For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 

Advanced reactors are designs—ranging in size from microreactors to SMRs to large-scale reactors—that offer better efficiency, safety, and versatility than conventional reactors.¹³ Currently, the US is developing more than 30 advanced reactor designs, as shown below.  

Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, 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.

While SMRs seem to be attracting the most attention, our research suggests that a mix of microreactors, SMRs, and large-scale reactors will be essential to tripling capacity. Smaller reactors with shorter construction times and modular designs have the potential to help meet near-term demand, particularly defense and behind-the-meter applications, as shown below.  

*Microreactor and Small Modular Reactor timelines are estimates. Large-scale reactors have historically taken ~6-8 years on average to construct. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of February 25, 2025, which may be provided upon request. 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. 

Longer term, as they have since the 1950s, large-scale reactors are likely to provide most of the reliable baseload power to the US grid. Selecting and iterating on a few key designs from more than 30 options in the US will help reignite the cost declines associated with Wright’s Law, ultimately lowering electricity prices. 

In our view, the geopolitical landscape today is much like that of the 1950s, including concerns about excessive regulation and losing a global nuclear race. Industry and government collaborated to create a surge in nuclear plant construction in the 1960s.¹⁴ Now, with bold targets, rising demand, and advanced reactors nearing deployment, the US has another opportunity to act decisively and reclaim its leadership in the nuclear energy race.