CFJ Op-Ed: Powering AI: Why Markets Matter More than Mandates
- Jeffrey Depp
- May 19
- 2 min read
Artificial intelligence is rapidly transforming the American economy. But behind every breakthrough model, advanced semiconductor, and data center lies a less glamorous reality: electricity.
As AI adoption accelerates, the United States faces a growing challenge. The computational infrastructure required to support advanced AI systems demands enormous quantities of reliable power. Increasingly, America’s ability to lead in AI will depend not only on software innovation, but also on whether it can generate enough electricity to support it.
In a recent article published in the National Law Review, CFJ Senior Fellow Jeffrey Depp argues that policymakers should approach this challenge with caution. While there is growing pressure to address rising electricity demand through government planning and industrial policy, history suggests that markets are often better equipped than regulators to determine how best to expand energy supply.
The article notes that many technology firms are already investing heavily in infrastructure, including transmission upgrades, grid improvements, and new generation capacity. Companies such as Meta have also expressed support for permitting reform and increased investment in nuclear energy. These developments suggest that private actors recognize the scale of the challenge and are responding accordingly.
The real question is not whether additional electricity generation will be needed—it unquestionably will be. Rather, the question is whether policymakers will allow markets and engineering realities to guide investment decisions, or whether political preferences will dictate which technologies are permitted to compete.
This distinction is particularly important because AI data centers require reliable, continuous-duty power. Unlike many other industrial users, advanced computing infrastructure cannot simply scale back operations when weather conditions change. Reliability matters. That reality has renewed interest in energy sources capable of providing consistent baseload generation, including nuclear power, natural gas, and other dispatchable resources.
The article also cautions against viewing energy policy solely through the lens of climate politics. While renewable technologies may play an important role in the nation’s future energy mix, policymakers should avoid imposing rigid mandates that ignore concerns about intermittency, storage limitations, transmission requirements, lifecycle costs, and supply-chain vulnerabilities. The concentration of solar-panel, battery, and critical-mineral production in China raises additional strategic considerations that deserve attention.
Drawing on insights from Austrian economics and Public Choice theory, the article emphasizes the importance of decentralized decision-making and incentive structures. Markets aggregate information through prices, investment decisions, and entrepreneurial experimentation in ways that centralized planners cannot easily replicate. Likewise, Public Choice economics reminds us that regulators and political actors respond to incentives just as market participants do, creating opportunities for rent-seeking and regulatory distortion.
The article concludes that the most effective AI energy policy may be one that focuses less on directing investment and more on removing barriers to it. Streamlining permitting, facilitating transmission development, and allowing diverse energy technologies to compete can help ensure that supply expands to meet growing demand.
America’s AI future will depend on abundant, reliable, and affordable energy. Achieving that goal will require innovation not only in computing, but also in the institutions that govern energy production and infrastructure development.
Read the full article, “Powering AI: Why Markets Matter More Than Mandates,” in the National Law Review.
