America’s ‘Golden Age’ Requires Energy Dominance

Encouraging AI proposals will require vast amounts of energy we’re currently leaving in the ground

President Trump announced the arrival of a “Golden Age of America” in his Second Inaugural Address, and he might just be right. Blending his signature hyperbole and pro-American rhetoric with some techno-optimism, Trump promised to make the United States energy dominant, and authorized a swath of executive actions to put that pledge into effect. Much debate on U.S. energy policy has been spent on the metrics of electric vehicle mandates and offshore drilling, but the most pressing issue is how to meet energy demand from the artificial intelligence boom. Trump made his first big infrastructure event all about AI, announcing a $500 billion investment over the next four years for the new Stargate initiative run by OpenAI, SoftBank, and Oracle, starting with a new data center in Texas. But this simply won’t be possible without huge amounts of additional power.

Goldman Sachs estimates that the global energy demand for data centers will increase by 165% by 2030 because of the accelerating growth of AI, which would result in data centers using 8% of all U.S. power. AI currently uses around 30 times more computing power than standard search engines like Google, and that means using more electricity. The capabilities and performance of AI have scaled up dramatically in recent years (just look at how much the AI videos of Will Smith eating spaghetti have gone from pure nightmare fuel to looking almost real). While we do not know exactly how quickly AI will develop, leading voices such as Anthropic CEO Dario Amodei believe AI could actually surpass human intelligence in the coming years. What we do know is that the U.S. is home to some of the best talent and most innovative companies in the world with vast sources of private capital to back them up, making AI an immensely valuable asset for American prosperity and security.

The huge leaps forward in AI that likely await us over the next few years will demand a massive expansion in these data centers that are incredibly energy intensive. DeepSeek’s dramatic entry has already sent US tech stocks tumbling, but Jevons Paradox would suggest that energy demand will still dramatically increase. While the energy efficiency of AI will improve, big increases in computing power from increased consumption will inevitably require corresponding increases in conventional power. And recent upticks in electric vehicles, cloud storage, and crypto mining have already put stress on an outdated domestic power grid with lines that are sometimes 50 years old, burdened by restrictive permitting rules and clean-energy targets…

Please continue reading at Commonplace where I have joined as a contributor