Clean energy is crucial to winning AI competitions with China

Free updates are notified at any time

The author is former U.S. Treasury Secretary and President of Paulson College

The US-China AI competition is one of the most urgent and unwelcome energy challenges facing the US. The United States leads, but Beijing is investing heavily to close the gap. The question for Washington is: Can the United States stay ahead without the national energy strategy powering it?

In recent years, the energy landscape has undergone tremendous changes. Russia's invasion of Ukraine reshaped it overnight. Prices soared, and the government rushed to reduce its dependence on Russian gas. Energy security becomes crucial. As Europe and other regions of non-independent energy sources seek to address these vulnerabilities, they are increasingly looking to reduce dependence on fossil fuels.

China is working to combine its long-term industrial strategy with substantial investment in AI infrastructure and the energy that supports it. Its data center market is expected to grow by nearly $2.75 billion between 2025 and 2029. It invests more renewable energy in 2024 than the U.S., the EU and the UK combined. Beijing’s clear ambitions are technology that will rule the future, and understanding energy policy will be key.

Meanwhile, in the United States, as AI models become more complex and deployed more, and cloud power grows, power demand rises faster than the capacity utilities can build. Some data centers now consume as much power as mid-sized cities. In Virginia, they consumed a quarter of the state's electricity load in 2023. This has increased focus on the pressure on the system and higher residential bills, resulting in new regulations and effectively suspending the state’s building data centers.

The United States must develop and deploy a clear national energy strategy to prioritize speed, flexibility, and cost-effectiveness. Since the demand for electricity greatly outweighs the supply, it makes sense to view clean electricity as an “increase” of existing energy supply. In fact, nearly all other electricity in the United States came from solar and wind last year – traditional energy sources couldn’t meet demand.

Due to turbine shortages, gas cannot fill voids in the short or medium term. Coal cannot solve the problem because decommissioned utilities cannot increase quickly enough to meet urgent needs (and air pollution can harm health). The expansion of nuclear energy is an important part of the solution, but the United States lags a few years behind, while China has made progress in developing state-of-the-art nuclear power.

Given the rich gas in the United States, many data centers will still rely on it, but smarter, faster and faster models are emerging. For example, due to the delivery time of equipment orders and components, air allows and interconnect, hybrid models using solar and battery storage or peak gas backups expand faster during high demand. These also require less upfront investment, and the “full cost” is equivalent to the base carrier gas.

The model is not without risks. The near-term shock of solar or storage ecosystems will leave a big hole, just like the demand for AI power soars. Without existing solar incentives, these incentives drive a large upside to level the playing field with China, and the full cost of the hybrid model will rise as the U.S. government’s tariff regime. But sticking to gas will be a sacrifice of development speed and betting on a rapid decline in solar and battery storage costs, which fell rapidly by about 20% last year and fell more than 80% in the past decade.

The correct answer is not to choose one energy source over another, but to quickly create conditions that extend valid conditions. This means low taxes that make U.S. renewable energy boom during the first Trump administration, light-regulating the environment. We should also exempt technologies such as long-term batteries, while double the support for practical solutions based in the United States. As for how utilities recover costs, we must encourage investment in technology that reduces energy prices for everyone.

AI is the main driving force for power demand. But this also has the potential to unlock substantial supply gains. We can't miss this moment. More capacity, lower costs and higher energy security are within our control.

The United States still has established advantages in energy independence. Testing is whether we can meet the energy needs of the next era of innovation. If so, we will be stronger than other countries’ AI premises in the coming decades.