Can America keep up with AI’s energy appetite in the race with China? | Utility Dive

Even with companies like Amazon, Google and Meta making big moves on clean energy, without an improved permitting process, their data centers might struggle to get the power they need to go online.

Divyansh Kaushik is a non-resident senior fellow at American Policy Ventures.

Artificial intelligence may be defining our era, but its insatiable appetite for energy is competing to share the spotlight. AI systems, which are transforming industries and sparking new innovations, require staggering amounts of power to function. Data centers, which serve as the backbone of AI, are projected to consume nearly 8% of global electricity by 2030.

As AI’s energy needs soar, a key question is materializing: Can the United States keep up with The People’s Republic of China, its near-peer competitor in AI research and development, or will we be left behind?

It’s a question the White House is hoping to tackle through its new task force focused on strengthening AI infrastructure and ensuring we have the energy capacity to power American data centers. The Department of Energy has also taken several measures and shared resources that industry can benefit from to meet this rising electricity demand from AI.

With both nations chasing the lead in AI, meeting the massive energy demands of this technology will be critical to maintaining leadership. If the U.S. doesn’t act quickly, we risk ceding ground to China in a cornerstone element of emerging technology. This was the subject of a recent Senate Energy and Natural Resources Committee hearing, where I highlighted the need to modernize energy infrastructure to keep the U.S. ahead of China in the AI race.

Since 2010, the amount of computing power needed to run AI models has been doubling every six months, and there’s no sign of this slowing down. In just six years, AI’s computational demands have exploded — from GPT-1 to GPT-4 — putting immense pressure on our energy infrastructure. Without a plan to modernize our energy systems and permitting processes, we’ll struggle to keep pace with this exponential growth.

To remain competitive, the U.S. must ensure its energy infrastructure can handle this rapid expansion, especially as global rivals like China are ramping up their efforts. China, for example, has increased its research and development spending by 10% every year for the past several years, clearly aiming to dominate the AI space by 2030. If the U.S. doesn’t take action to meet AI’s growing energy needs, we risk losing our leadership in global innovation.

One way to meet this challenge is by reforming how we approve energy projects. Right now, long delays in getting energy infrastructure built are making it hard to respond quickly enough to AI’s exploding demands. Legislative efforts like the Energy Permitting Reform Act are highlighting potential approaches to speeding up the approval process while maintaining environmental safeguards.

Today, it can take years to get a new energy project connected to the grid. In Virginia, a major hub for data centers, it can take years for a new facility to secure the 100 MW or more needed to power up. These kinds of delays slow down business and put us at risk of falling behind in the AI race.

Streamlining the permitting process isn’t about cutting corners or ignoring environmental concerns — it’s about finding a way to meet our energy needs responsibly and efficiently. The ongoing policy discussions show that we don’t have to choose between environmental protection and energy progress. By removing unnecessary delays, we can accelerate the development of new energy projects, including cleaner energy sources, while maintaining strong environmental standards. It’s about finding a balance that ensures we can meet the urgent demands of AI and stay true to our long-term environmental goals.

The private sector is already signaling the urgency of this challenge. Oracle, for example, recently announced plans to build a 1 GW data center to meet the rising demand for its AI and cloud services. Hyperscalers like Microsoft, Google, Amazon and Meta are making big bets on nuclear and advanced geothermal energy projects to meet the moment. But even with companies making big moves, without an improved permitting process, these data centers might struggle to get the power they need to go online.

It’s also important to recognize that AI could be part of the solution. AI can improve the efficiency of our electric grid, predicting and managing energy supply and demand in ways we couldn’t before. While AI is driving energy demand, it also has the potential to make our energy systems smarter and more efficient.

Investing in energy-efficient AI technologies is going to be key. Data centers are already major consumers of electricity, and by investing in more energy-saving technologies for AI systems, we can reduce AI’s overall energy footprint. The U.S. Department of Energy has already begun to explore how AI can make the grid more resilient and improve energy forecasting, helping us stay ahead of the curve.

According to the International Energy Agency, demand for electricity from data centers could double by 2026, driven largely by AI workloads. That presents both a challenge and an opportunity. If we can modernize our permitting systems and continue to invest in energy-efficient AI technologies, we can meet the growing demand and turn it into a competitive advantage for the U.S.

The conversation around permitting reform shows just how critical it is to address our energy infrastructure as we move deeper into an AI-driven future. By ensuring we have the resources and infrastructure to meet these new demands, we can secure both economic growth and national security. Striking the right balance between regulatory streamlining and environmental protection will allow the U.S. to stay at the forefront of AI innovation and remain a global leader in the energy landscape of tomorrow.