Powering the AI Revolution in the United States: A Discussion on Resources and Strategies

The Federal Energy Regulatory Commission (FERC) has more than tripled its forecast for growth in U.S. power demand by 2030, suggesting the need for between 50 GW and 130 GW in extra generating capacity. This surge is largely driven by the advance of artificial intelligence (AI) and the growth of data centers that support it.

The market is overweighting the pursuit of natural gas to meet the energy demands of these data centers. However, the widespread recognition of the "bottomless well" of underlying energy resources means that other options, such as geothermal, natural gas, and nuclear, are also being considered.

The issue is no longer about the exhaustibility of resources but about the need to weigh choices and tradeoffs, while minimizing government friction. This is particularly important as companies such as Anthropic, OpenAI, Google, and other tech giants have national energy plans to ensure power for AI development.

Anthropic's July 2025 roadmap notes that AI will require a broader-based effort to unlock energy and data center buildouts around the country. One key aspect of this effort is the acceleration of permitting for geothermal, natural gas, and nuclear projects.

Each AI-focused data center requires a significant amount of power. A single data center hosts hundreds or thousands of racks, with a single AI server rack, refrigerator-sized, weighing as much as a car and using as much energy annually as 100 cars. The telecommunication network supporting these data centers, made up of hundreds of thousands of miles of physical cables and wireless "roads", is power hungry.

Caterpillar announced a partnership to supply some 4 GW of diesel engine generation for data centers, while the industry that supplies diesel truck engines manufactures 150 GW of power capacity every year. Construction is underway on dozens, and soon likely hundreds, of AI-centric data centers in the U.S., each requiring 200 MW to 1,000 MW of power.

A detailed national survey found a total of some 55 GW of new data-center projects already committed or under construction. The three major vendors for utility-scale gas-fired turbines are sold out through 2030, indicating the scale of the challenge.

The concern about ensuring that America can power the AI and data-center revolution is not a new one. As far back as 1865, economist William Stanley Jevons's paper, "The Coal Question," was animated by worries that the "exhaustion of coal" would end the prosperity wave driven by the invention of steam engines. Today, the focus is on ensuring a sustainable and reliable energy supply to drive the next wave of technological innovation.

Some have framed the challenge of delivering power as a partisan issue, but the "digital sobriety" crowd is eager to emphasize that it is a shared responsibility. The key will be to balance the need for growth with the need for sustainability, and to ensure that America remains at the forefront of the AI revolution.

Powering the AI Revolution in the United States: A Discussion on Resources and Strategies