Earlier this year, when President Trump abruptly declared a national energy emergency, my reaction was probably the same as yours. "He is just trying to lower energy prices to fight inflation." But now that we are in June, a chilling reality has set in. Taming inflation was never the main goal. It was a convenient cover story for a much larger, stealthier war.
When the news first broke, the people around me brushed it off. Friends would ask, "Is tweaking the price of crude oil and natural gas really that big of a deal?" Now I am ready to tell them the hidden truth behind that headline. The moment I publish this piece, I plan to share this link directly with them. Because what I discovered is far more severe than I ever anticipated.
This was no routine economic policy. It was a desperate, all-in gamble to secure the single most critical weapon of the AI era: raw electricity. Why else would President Trump sign executive orders demanding the construction of data centers and power grids at a literal wartime pace? The mainstream media remains fixated on the AI rivalry between the US and China. Yet they are paying dangerously little attention to the real underlying threat: skyrocketing electricity bills and the looming specter of massive power outages.
The Invisible War for Electricity
The race to develop AI is essentially an electricity war. A single conversation with ChatGPT consumes over ten times more power than a traditional Google search. I use AI constantly to organize data and ask questions, not just once, but hundreds of times a day. Many of you probably do the same.
The power consumed by a single massive data center is equivalent to the electricity needed to run 1,000 Walmart stores.
The International Energy Agency issued a stark warning. By 2026, the electricity consumption of global data centers will rival the total annual power usage of the entire country of Japan. We are entering an era where individual corporate facilities are devouring as much energy as major cities or entire nations.
The winner of the AI race will not be decided by who launches the smartest model. It will be determined by who can reliably feed colossal amounts of electricity into these data centers. Without enough power, data centers grind to a halt, rendering even the most advanced AI completely useless. The shocking reality is that our infrastructure is vastly unprepared for this surge in demand. And the situation is most dire in the United States, the very heartland of artificial intelligence.
If you want to fully grasp the madness of this tech infrastructure war, I highly recommend reading Supremacy. While I was researching this column, this exceptional book provided the exact details on why these corporate giants are acting so recklessly. Bloomberg tech journalist Parmy Olson exposes the ruthless, behind-the-scenes battle between Google and OpenAI in a way that reads almost like a thriller. If my post made you wonder what is truly driving them to burn through the planet's resources at such a terrifying speed, this book will give you the complete picture.
Why Big Tech Broke Its Climate Promise
Back in 2023, Google revealed that their greenhouse gas emissions had spiked by 48% compared to 2019. Microsoft admitted their carbon emissions had climbed by over 30% since 2020.
"Big Tech's 2030 carbon reduction targets are incredibly ambitious, but we are already witnessing various phenomena driven by the climate crisis." — Amanda Smith, Climate NGO Scientist
 |
| Big Tech's green promises versus the stark reality of fossil fuel dependency |
What happened to these tech giants who solemnly pledged carbon neutrality by 2030? The answer is brutally simple. When the existing power grid could not feed their starving AI data centers, they resorted to generating their own electricity, which sent their carbon footprint soaring. Amazon signed a deal to build a massive 650-megawatt natural gas power plant in Oregon. Meta is pushing a similar initiative in Louisiana.
Their pivot back to natural gas is purely logical. Solar and wind fluctuate with the weather, making them poorly suited for data centers that must run 24 hours a day, 7 days a week. Battery storage technology is not yet economically viable. So the fastest and most reliable fix was fossil fuels, specifically natural gas. The ultimate irony is that policy is actively encouraging this detour. States like Texas, Oklahoma, and Georgia are rolling back environmental regulations just to win the fierce bidding wars for these data centers.
The Bridge Fuel Dilemma and Global Inequality
This AI-triggered explosion in natural gas demand has completely rewritten the rules of the energy market. According to the US Energy Information Administration, natural gas prices surged 40% year-over-year in 2024, and data center-driven demand is projected to triple by 2030.
Since 2023, shale gas companies have seen their stock prices soar by an average of 80%, with over $200 billion pouring into new LNG terminal construction. But the myth of "clean fossil fuel" has a fatal blind spot. Methane, the primary component of natural gas, is a greenhouse gas 80 times more potent than carbon dioxide. A Cornell University study revealed that methane leaks during US natural gas production are actually three times higher than official government statistics. Worse, the gas infrastructure being built today is designed to operate for 30 to 40 years, putting it on a direct collision course with 2050 net-zero targets. The IEA has bluntly warned that all investments in new fossil fuel infrastructure must stop immediately.
The geopolitical fallout is equally severe. Europe, which ramped up US natural gas imports fivefold to escape its reliance on Russia, is now forced into a bidding war with AI data centers.
Europe used to buy American gas at standard rates. Now they are paying a massive premium.
By winter 2024, LNG prices in Germany and Japan shattered all-time records, while developing nations plunged deeper into energy poverty. The AI revolution is rapidly accelerating global energy inequality.
The Desperate Pivot to Nuclear Power
Backed into a corner, Big Tech is turning to a technology it once avoided: nuclear energy.
In September 2024, Microsoft signed a 20-year power purchase agreement to revive the Three Mile Island nuclear plant, shuttered since 1979. Amazon is pouring $960 million into building a data center next to the Susquehanna nuclear plant in Pennsylvania. The real buzz, however, surrounds Small Modular Reactors, or SMRs.
In October 2024, Google struck a deal with Kairos Power to secure 500 megawatts of SMR energy by 2030. More than 120 startups have entered this space, including Oklo, backed by OpenAI's Sam Altman, and TerraPower, founded by Bill Gates.
The appeal is clear. SMRs are roughly one-tenth the size of traditional reactors, can be mass-produced in factories, and have a construction timeline of just 3 to 5 years. They generate zero carbon emissions while providing round-the-clock power.
But this is not a silver bullet. The economic viability of SMRs remains unproven, radioactive waste is still an unresolved issue, and the first commercial SMR will not be operational until at least 2030. Until then, our dependence on natural gas appears unavoidable.
The Hidden Cost Nobody Talks About
I was planning to wrap up here. But I recently came across another reality just as critical as the electricity crisis, and far less discussed: the staggering amount of water required to cool these data centers.
 |
| Massive water intake for data center cooling systems |
Exchanging just a few simple prompts with ChatGPT evaporates a 500ml bottle of drinking water.
In 2022, Microsoft's data center in Iowa consumed 1.7 billion liters of water in a single year, enough to fill 2,300 Olympic-sized swimming pools. Tragically, many of these facilities are being built in regions already suffering from severe water shortages, like Phoenix, Arizona, and Salt Lake City, Utah. In 2023, Google had to scrap a data center project in Santiago, Chile, due to fierce protests from local residents already living through a brutal drought. Recent studies project that by 2027, the global water consumption of AI data centers will surpass the total annual water usage of the entire population of Denmark.
We are watching a scenario where the resources needed to run AI are swelling to the scale of sovereign nations.
I have rarely harbored pessimistic thoughts about this turbulent AI era. But seeing how our finite resources, first natural gas and now water, are being aggressively consumed and creating severe friction between nations and communities, deeply worries me.
Intangible AI is devouring the tangible Earth.
AI has undoubtedly expanded human intellect. But in return, it is fiercely testing the physical limits of our planet. The natural gas boom might act as a band-aid for the short term, but in the long run, it will only pour fuel on the climate crisis.
Experts argue we must tackle this on three fronts. First, drastically improve the energy efficiency of AI models themselves. Second, accelerate investments in safe, zero-carbon energy sources, particularly SMRs and nuclear fusion. Third, if we must rely on natural gas, implementing Carbon Capture and Storage technologies and strictly regulating methane leaks should be legally mandated.
The AI revolution has already begun. It is an unstoppable tide. But true innovation is not just about creating something new. It is about ensuring it can be sustained. As AI continues to evolve, these environmental dilemmas will persist alongside it. I hope we all keep a close eye on this unfolding reality together. Let me know your thoughts in the comments below.
Curious about the real stories behind big tech, crypto, and everyday economics? 👉👉👉Subscribe to The Techtonic for your weekly dose of easy-to-read business trends.
Latest Insight: [How Uber is Navigating the AI & Autonomous Era]
Disclaimer: As an Amazon Associate, I earn from qualifying purchases.
댓글 쓰기