The US and China have starkly divergent visions of the future of energy, and they’re bringing them to play in the race to power artificial intelligence.
Running the computers needed to train and run the ChatGPTs and DeepSeeks of the world requires a lot of electricity. Total consumption from data centers is set to quadruple during the next decade to 1,600 terawatt-hours a year, according to BloombergNEF.
That’s more than Japan’s annual usage.
President Donald Trump has stressed the importance of using coal, natural gas and nuclear to ensure the US can satisfy that rising demand, most recently at an event in Pennsylvania, where he heralded more than $92 billion of investments in AI and energy infrastructure.
By contrast, Chinese President Xi Jinping wants to propel his country’s AI boom with renewables. Earlier this month, Beijing issued rules requiring new data centers to rely on clean energy for at least 80% of their electricity consumption.
The differing agendas play to each country’s strengths. The US has vast shale resources that help make it the world’s biggest oil and gas producer, while China has assumed the dominant position in global supply chains for manufacturing solar panels and batteries.
Both approaches have advantages and disadvantages.
In the US, cheap gas is plentiful right now, but the expanding power sector and a coming wave of export plants promise to create a bidding war for the molecules.
Chinese solar and wind energy is cheaper than gas-fired electricity in the US. However, the intermittent nature of clean generation means Beijing is investing more in battery storage and grids to ensure it can keep adding panels and turbines.
Neither country’s journey will be black and white: Many American companies still want to buy renewable power to run their data centers, and China will still rely on coal to some extent.
But both see AI as a competition between them, and their energy policies will go a long way toward deciding the winner.
--Dan Murtaugh, Bloomberg News
