A year ago, I wrote about the symbiotic growth of AI and energy—how AI’s rise is directly tied to the power industry’s ability to keep up. (Read the original here).
Some industries grow in parallel, but others are locked in a fate where they either rise together—or collapse together. IMO, AI and energy belong to the latter. AI demands more and more power; the energy sector scrambles to supply it. If energy lags, AI slows. If AI grows faster than energy infrastructure can handle, data centers become bottlenecks.
It’s easy to see these relationships in hindsight. Smartphones and semiconductors. Cloud computing and broadband expansion. Social media and digital advertising. But the real insight—the one that creates fortunes—is recognizing these dependencies before they become obvious.
Twelve months ago, it was clear AI would need more energy. What I didn’t imagine was the sheer scale of the demand surge (and the gap it’ll create).
AI’s Power Consumption: A Five-Year Surge in One Year
AI’s hunger for power has accelerated beyond every forecast. In 2024, data centers, AI workloads, and cryptocurrency mining consumed about 460 terawatt-hours (TWh) of electricity—nearly 2% of global electricity demand (International Energy Agency, 2024). By 2026, AI’s power needs alone could surge by 550%, reaching 286 TWh, and by 2030, the number could hit 652 TWh—more than the entire country of Japan consumes today (Forbes, 2024).
The chart below illustrates just how fast AI's electricity needs are expected to rise:
The Data Center Expansion Race is only accelerating: Tech giants are racing to build out infrastructure. Microsoft plans to spend $80 billion to scale its AI-ready data centers across North America. Google has placed orders for small modular nuclear reactors (SMRs) to power its AI-driven infrastructure (The Guardian, 2024). Amazon and Meta are investing heavily in natural gas plants to ensure power stability.
Why is this happening? AI models like GPT-4, Gemini, and Claude have made massive computational leaps. Training these models requires thousands of high-performance GPUs, each consuming 300-700 watts per chip under full load. A single AI model training cycle can use as much energy as 100,000 U.S. homes do in a month. The rise of real-time AI applications—like ChatGPT-powered search engines, AI video processing, and self-driving car simulations—has made energy costs a business-critical factor for tech companies.
The Energy Industry’s Response: Scrambling to Keep Up: This AI-driven demand has forced an unprecedented transformation in the energy sector.
Nuclear energy, once considered outdated, is making a return. Microsoft has signed deals to reopen Three Mile Island’s nuclear plant, tapping into stable, carbon-free energy. Google is betting on small nuclear reactors, securing contracts to build AI-dedicated power sources.
Natural gas, often seen as a transitional fuel, is seeing renewed investments. Microsoft has partnered with utility providers in Wisconsin to build gas-powered AI data centers. Meta and Amazon are backing large-scale gas plants in Louisiana and Mississippi (Business Insider, 2025).
Renewable energy remains in focus, but challenges persist. Shell is using AI to optimize power grids, reducing energy waste and improving efficiency in industrial processes. But solar and wind projects alone cannot yet meet AI’s demand spikes—especially as AI workloads require continuous, always-on power rather than intermittent energy sources.
The Financial Shift: AI is Making Utilities Hot Again: This is where the fun begins. For years, the stock market viewed utilities as boring, low-growth investments. AI has flipped that narrative. Hedge funds that once chased high-growth tech stocks are now moving into power generation companies. Vistra and Constellation Energy—historically slow-growth utility stocks—have become top performers in the S&P 500 .
Power companies are now AI enablers, making them some of the most valuable assets in the tech supply chain.
The Next Big AI-Driven Resource Battles
AI and energy are just one part of the equation. What’s the next big industrial connection that will shape the future?
AI and Water: Data centers need enormous amounts of water for cooling. With each AI data center consuming millions of gallons per day, the industry is already putting pressure on global water supplies. Google’s data centers consumed 4.3 billion gallons of water in 2023, a 20% increase over 2022. Microsoft’s AI training pushed its water consumption up by 34% in a single year (Nature, 2024). Major AI hubs like Phoenix, Dallas, and Las Vegas are in drought-prone areas, making water supply a long-term constraint. Prediction: The water-tech industry will boom as companies scramble for better cooling solutions, desalination, and closed-loop water recycling systems.
AI and Rare Earth Metals: The AI hardware boom is fueling a global rush for rare earth elements like lithium, cobalt, and neodymium—materials essential for GPUs, superconductors, and quantum computing. A single Nvidia H100 AI GPU requires 30+ rare materials, many of which come from China or politically unstable regions. The price of lithium has surged by 500% over the last five years, driven by demand from AI and EVs. The U.S. and EU are investing billions in rare earth mining to reduce dependence on China, setting up a geopolitical battle over mineral supply chains. Prediction: Expect massive investments in rare earth mining, alternative chip materials, and AI-designed resource extraction to counteract shortages.
Seeing the Future Before the Market Does: The AI-energy link was obvious in hindsight. The real challenge—the real goldmine—is predicting the next great industrial dependency before it plays out. AI and energy are now locked together. The next wave of AI-driven industries is forming. If you see the right pairings before the world does, you win big. And if you’re wrong? The market will remind you.