Energy supply is the key to the AI race with China

Energy abundance could be the deciding factor in whether America or China wins the race for global artificial intelligence leadership. Right now, China is surging in that race — and they’re winning it on the power grid, not just in the lab. 

Beijing has treated AI as a national mission since 2017, mobilizing its entire economy around one goal: building the energy to power the future. America’s crumbling grid is handing them that advantage. As a result, China generates twice as much electricity as America does and will add 60 percent additional capacity by 2040. 

Unfortunately, our grid cannot say the same. After decades of underinvestment in our physical backbone, our energy system is fragmented, outdated, and unreliable. Seventy percent of our transmission lines are at least 25 years old. Our grid also faces structural red tape, which limits its efficiency. Some 12,000 energy projects are awaiting regulatory approval for connection. 

Increased energy demand is a signal of growth and innovation. The issue is when the supply can’t keep up with demand. 

Our lack of supply has enabled China to build an early lead in the AI adoption race. Eighty-three percent of Chinese companies use generative AI, while just 65 percent of American companies do the same. Innovation requires a stable, high-capacity environment that the U.S. currently struggles to provide.  

Recognizing this challenge, the administration recently announced a “Ratepayer Protection Pledge,” creating a framework that allows technology companies to finance new power generation while protecting consumers from rising energy costs. The initiative reflects a growing recognition in Washington that AI leadership depends on abundant, reliable electricity.

But federal action alone won’t be enough. America’s national strength ultimately depends on the decisions made across the country. The choices made locally shape both America’s strength and community prosperity. 

Our home states offer proof that America can meet this moment. North Dakota produces more energy than it uses and is the third-largest producer of crude oil in the nation, making it capable of meeting AI’s power demands.

Georgia is also already showing the way on energy, adopting a strategy that incorporates all types of energy to meet the load growth of the AI era. In December, the Peach State approved 10,000 megawatts of new energy capacity from natural gas, while the Plant Vogtle nuclear plant is now the largest clean energy generator in the nation. 

Congress should follow the lead of our two states. To reclaim the lead in this race, ensuring that we have an uninterrupted power supply to operate our AI is key. This will take moving with urgency across three fronts:

• That starts with codifying the administration’s AI Action Plan: This framework creates a durable federal AI strategy, prevents future policy reversals, and stops state-level overreach, which slows American innovation. The plan also takes important steps to prioritize building and maintaining a vast AI infrastructure and the energy to power it. 

• Then, building more power generation and transmission infrastructure: This means accelerating “all-of-the-above” energy generation; onshoring turbine and transformer production; modernizing and securing the grid; streamlining permitting and expanding transmission. 

• And, upgrading America’s AI base: Maintain U.S. chip and compute advantage; securing critical mineral supply chains; enacting a multi-year freeze on restrictive state AI laws; rejecting regulations that weaken innovation.

America can still win this tech race. We have the talent and entrepreneurial spirit to out-innovate the Chinese. But even the most brilliant algorithms cannot run in the dark. Sustaining our competitive edge will require a grid built on the foundation of local ambition. 

Kent Conrad represented North Dakota in the Senate from 1986 to 2013 as a Democrat. Saxby Chambliss represented Georgia in the Senate from 2003 to 2015 as a Republican. Both serve as advisers to the American Edge Project.