Building the Energy Capacity We Need for an AI World

Innovations in the supply and distribution of energy has always been one of the key limits, and opportunities, in allowing society to evolve. Steam power, for example, provided significant energy increases which allowed textile factories in the19th century to operate more efficiently through increased use of machinery. Digging back into history, there are examples as far back as Roman aqueducts and beyond as to how innovations in energy transmission and distribution allowed society to evolve.

Since the launch of Chat-GPT3 back in November of 2022, the artificial intelligence wars continue all over the world among tech giants including Google, Microsoft, Amazon and NVIDIA. In turn, as consumer interest increased, the opportunity to capitalize on AI-demand has led to a significant, abrupt spike in the need for data centers. While data center growth was marginal over the previous decade, the sudden boom may shock the existing power grid not presently equipped to handle the increased volume. All the while, Cisco, Oracle, and others fight for cloud compute and telecommunications market share.

What does it all mean? Although the power required to operate data centers has not significantly changed in the last decade, reports—including one from Goldman Sachs—now predict that power output will need to double by 2030, the next five years, to keep up with the growth driven by these companies.

Power Infrastructure Comes Center Stage

Power infrastructure over the next five years in particular comes to the center stage to play a pivotal role in providing the additional energy required, significantly driven by data centers,. In April, the federal government launched what is known as the Coordinated Interagency Transmission Authorizations and Permitting Program which is intended to expedite permitting and approval processes for power transmission projects. The ability to efficiently plan, permit, and execute critical infrastructure projects will be critical to keep up with rising demands.

At the end of each of those transmission lines, power fulfillment will likely come from an increasing mixture of natural gas and renewable energy alternatives. While LNG (liquid natural gas) is predicted to grow significantly to match demand, investments in solar and wind power as leading renewable energy alternatives will also be required to supplement growth.

As long as demand for AI continues on its current trajectory, the need for additional energy infrastructure through both upgrades and new distribution channels will be considered critical. It is a complex challenge, but with expedited permitting and planning, stakeholder collaboration, and public and private investment, can be achieved.

Can We Get There?

Public infrastructure projects like this are notoriously complex, ranging from massive transportation projects through rural America to water treatment plants to underground energy and utilities work. Historically speaking, these projects are slow to plan and develop, difficult to finance and, after all of those hurdles, difficult to execute properly due to unforeseen conditions or complex designs.

However, there are microcosms of project successes all over the place which indicate the capacity and ability to quickly and effectively address infrastructure projects through stakeholder collaboration and focus. Initial reports of the cleanup of the Francis Scott Key Bridge indicated it would take a year to re-open the water way— - last month, officials announced that the work was completed in 11 weeks. Not only that, but the planning process for the new bridge is already in progress, with a project team to be selected this summer—a case study in what may be considered a “streamlined” approach and example for other projects around the country.

If these expedited delivery efforts can be successfully applied to the growing energy needs, the difficulties AI has posed may make way for innovations in power assets to succeed. These initiatives are critical to society’s ability to evolve to an AI-powered future, whatever that may look like.