Data centers will consume 8% of U.S. power by 2030. Nuclear plants are being restarted to feed the beast. Tech giants are scrambling for energy deals worth billions. Yet 83 years ago today, beneath a Chicago football stadium, 49 scientists proved that splitting atoms could power civilization.
December 2, 1942: Enrico Fermi pulls a control rod from a pile of uranium and graphite blocks.
The Geiger counters click faster. Then faster still. For 28 minutes, humanity controls atomic fission. The Manhattan Project team celebrates with a bottle of Chianti in paper cups. They've just unlocked the universe's most concentrated energy source in a converted squash court.
The government invested $2 billion (1940s dollars). That's roughly $30 billion today. Seems massive until you realize OpenAI alone raised $6.6 billion this year. Microsoft committed $100 billion for AI infrastructure. The energy crisis they're creating dwarfs the one nuclear solved.
Here's what Fermi's team understood that today's AI builders miss: Energy density matters more than energy volume.
One uranium pellet (fingertip size) = 1 ton of coal One nuclear plant = 3 million solar panels One reactor = Powers 750,000 homes
While AI companies chase renewable deals for PR, the math is brutal. Some estimates suggest training GPT-4 consumed 50 GWh. That's enough to power 10,000 American homes for a year. For one model. We're building hundreds.
The smart money sees it. Constellation Energy stock surged 61% this year. They're restarting Three Mile Island for Microsoft. Amazon bought a nuclear-powered data center. Google signed the world's first corporate small modular reactor deal.
💼 What This Means for Your Wallet:
The next decade's biggest returns won't come from AI companies. They'll come from whoever solves AI's energy problem. Nuclear power, after 83 years, is finally seeing the widespread adoption and innovation akin to the iPhone's impact.
While everyone debates AI safety, the more immediate bottleneck is a fundamental one: physics. You can't run tomorrow's intelligence on yesterday's grid.
Fermi's pile initially generated just 0.5 watts. Today's reactors generate 1,000 megawatts. That 2-billion-fold improvement took 80 years. AI needs that leap in 8.
What happens when infinite intelligence meets finite energy?
Note: Featured image is an AI-generated historical illustration, not a period photograph.
Inspired Money
Data centers will consume 8% of U.S. power by 2030. Nuclear plants are being restarted to feed the beast. Tech giants are scrambling for energy deals worth billions. Yet 83 years ago today, beneath a Chicago football stadium, 49 scientists proved that splitting atoms could power civilization.
December 2, 1942: Enrico Fermi pulls a control rod from a pile of uranium and graphite blocks.
The Geiger counters click faster. Then faster still. For 28 minutes, humanity controls atomic fission. The Manhattan Project team celebrates with a bottle of Chianti in paper cups. They've just unlocked the universe's most concentrated energy source in a converted squash court.
The government invested $2 billion (1940s dollars). That's roughly $30 billion today. Seems massive until you realize OpenAI alone raised $6.6 billion this year. Microsoft committed $100 billion for AI infrastructure. The energy crisis they're creating dwarfs the one nuclear solved.
Here's what Fermi's team understood that today's AI builders miss: Energy density matters more than energy volume.
One uranium pellet (fingertip size) = 1 ton of coal
One nuclear plant = 3 million solar panels
One reactor = Powers 750,000 homes
While AI companies chase renewable deals for PR, the math is brutal. Some estimates suggest training GPT-4 consumed 50 GWh. That's enough to power 10,000 American homes for a year. For one model. We're building hundreds.
The smart money sees it. Constellation Energy stock surged 61% this year. They're restarting Three Mile Island for Microsoft. Amazon bought a nuclear-powered data center. Google signed the world's first corporate small modular reactor deal.
💼 What This Means for Your Wallet:
The next decade's biggest returns won't come from AI companies. They'll come from whoever solves AI's energy problem. Nuclear power, after 83 years, is finally seeing the widespread adoption and innovation akin to the iPhone's impact.
Investment radar:
• Small modular reactor companies (finally getting regulatory approval)
• Uranium miners (supply can't meet coming demand)
• Nuclear plant operators (pricing power increasing daily)
• Grid infrastructure (everything needs upgrading)
While everyone debates AI safety, the more immediate bottleneck is a fundamental one: physics. You can't run tomorrow's intelligence on yesterday's grid.
Fermi's pile initially generated just 0.5 watts. Today's reactors generate 1,000 megawatts. That 2-billion-fold improvement took 80 years. AI needs that leap in 8.
What happens when infinite intelligence meets finite energy?
Note: Featured image is an AI-generated historical illustration, not a period photograph.
3 days ago | [YT] | 3