On AI, Energy, and America's Determination to Lose
A critique of our approach to winning the first technological race of the 21st century.
When we talk about the new technologies entering our world and what their ramifications might be, I wonder how our forefathers handled such seismic shifts in their own lifetimes. What were the consequences of their actions?
For most of human history, progress was slow, with technological advances often restrained by dogmatic systems of thought that feared upsetting society’s delicate balance. Refusing to embrace progress was tolerable, even necessary at times to keep the peace… But the twentieth century changed everything. The pace of innovation exploded, and a handful of breakthroughs reshaped the entire world.
Consider electrification: the Tennessee Valley Authority, launched in the 1930s, brought power, dams, and industrial capacity to impoverished rural regions- transforming lives overnight. The internal combustion engine and mass automobile production propelled mobility and suburban growth. Aviation turned the planet into a smaller, more interconnected space, altering commerce and warfare. And the Manhattan Project, with its spectacular fusion of science, engineering, and national purpose, delivered nuclear power, both destructive and constructive, in record time. These technologies were never optional; they formed the backbone of national strength. Countries that mastered them surged ahead, while those that lagged were often left behind.
America emerged as the leader of the world because we chose to embrace these technologies and build them at scale. If we hadn’t, we’d be living in a world where Soviet Russia or perhaps Germany had the privilege of being a unipolar superpower.
Our century’s equivalent is Artificial Intelligence. It is the defining technology of our era. Just as no twentieth-century superpower could fall behind in steel, oil, or nuclear physics, no twenty-first-century power can afford to fall behind in AI. This consensus holds across party lines, and I think relatively is uncontroversial. No matter how much you hate generative art and em-dashes, the powers of AI cannot be denied- while the specific effects remain to be seen, it is unquestionably a technology that will transform modern economies and militaries.
What is not so straightforward, however, is how we stay competitive in this high-stakes race.
I am not an artificial intelligence scientist, but I do know a lot about power- and AI training demands an enormous amount of it. Training a large language model like Chat GPT‑4 is estimated to consume around 25 megawatt-hours in 3 months- roughly what 20,000 U.S. homes would use in a comparable timeframe. Deploying this software for consumer use consumes similar scales of power, and every high-cap tech company from San Francisco to Shanghai is trying to perfect their own model. Suffice to say, we need power- and a lot of it. Without abundant, affordable energy, AI progress collapses under its own weight.
That’s America’s key vulnerability right now. Our electrical grid is already strained; seasonal blackouts and brownouts in places like California and Texas are a reality, and a cyberattack on the grid is high on the nation’s security threat list. Our ability to build energy at scale has eroded. We spent the twentieth century generating power from oil, gas, and coal- but those sources are no longer viable for rapid expansion. They are finite, volatile, and geopolitically fraught. OPEC’s oil embargo of the 1970s, which crippled industries and drove energy prices sky-high, stands as a stark warning. Depending on such resources leaves us exposed to both climate disruption and the leverage of foreign suppliers.
The Trump administration recognized this urgency and took action. On May 23, 2025, the President signed executive orders aimed at significantly boosting nuclear power production, targeting a fourfold increase within 25 years. The orders streamline licensing processes (now aiming for under 18 months for new reactors and 12 months for renewals) and initiate pilot programs for small modular reactors to support national security and AI infrastructure. These directives build on existing legislation like the ADVANCE Act of 2024, designed to accelerate the development, licensing, and export of advanced nuclear reactor technologies.
Yet industrial capacity challenges remain. America’s heavy industry- once capable of building turbines, ships, locomotives- is severely diminished. Rebuilding it will take years, and we do not have years to spare. No matter what President Trump or his cabinet may wish, there is no large-scale American manufacturing base to call upon to meet this new demand for energy right now. Just signing a piece of paper and throwing government money will not stand up a new natural gas or nuclear plant tomorrow- and we need them today.
In stark contrast to fossil fuels and nuclear, solar power stands out as the only reliably scalable source we currently have. It’s relatively cheap and easy to build, and the energy it produces is cheaper than anything else on the market. It is the natural solution to our problem while we wait for re-industrialization efforts to materialize. However, a burgeoning solar industry threatens fossil fuel interests, which thrive on energy scarcity. These companies poured hundreds of millions into the 2024 election- some estimates suggest between $219 million and as much as $450 million in direct and indirect influence, mostly into President Trump’s campaign. They push to stymie solar through tariffs, permit delays, and more... and the President seems amicable to their desires. Recently, the Administration slammed heavy tariffs- more than 3,000 percent- on imported solar panels and posted a tirade on Truth Social, claiming the U.S. will not approve new projects.
That should raise alarm for anyone serious about America’s future.
Delaying solar and renewables is not just bad policy; it is strategically perilous. It creates an AI bottleneck, ceding advantage to nations that choose abundance. While America doubles down on legacy energy, our competitors are embracing scalable, future-facing technology.
We need energy everywhere we can get it. Refusing solar risks repeating the fate of our abandoned capacity in steel, shipbuilding, and automotive manufacturing- industries once world-dominant, now shadows of their former selves and struggling to even exist in our market economy.
Artificial Intelligence may be the frontier, but energy is the foundation. If we want to lead this century as we led the last, we must reject reactionary policies that constrain our solutions. Our future- and that of our children- depends on the choices we make now. We cannot wait years for nuclear power; we must scale what’s available today, or we will lose before the race even begins. We have the solution in front of us, but we must overcome our dogma to realize it.