Three years ago, ChatGPT didn’t exist.

It now has 800 million weekly users, making it the fastest-growing product in history.

But few people realize just how much electricity ChatGPT—and other artificial intelligence (AI) tools like it—consume.

And even fewer are prepared for the second-order effect AI is going to have on the energy market.

See, when you use ChatGPT, you’re tapping into vast data centers filled with specialized chips that run so hot, they need constant cooling just to keep from melting down.

One of Nvidia’s (NVDA) new AI chips draws as much power as two households.

And a modern AI data center needs 5X more power than a traditional one. Even a simple ChatGPT query uses at least 10X more electricity than a Google search.

Now, everyone is asking, “How are we going to power all these AI factories?”

• The answer is obvious: nuclear energy.

Ditching nuclear was a luxury.

For the past two decades, America’s energy consumption has been in a bear market. “Why do we even need nuclear?” was a legitimate question.

Source: Our World in Data

But we’re entering a new era of demand—one where AI, robotics, reindustrialization, and electrification all need more power.

And not just clean power, but “always on” power. Energy you can bet a $10 billion data center on.

And that is forcing everyone to do a U-turn on atomic energy.

• We’re already seeing signs of this major sentiment shift...

Recently, the World Bank lifted its ban on funding nuclear power projects—the first such move since 1959.

New York, once a poster child for bad energy policy, is also reverting course. After shutting down its largest nuclear plant in 2021, the state now plans to build a new nuclear power plant capable of producing enough electricity for 1 million homes.

Meanwhile, states like Illinois, Montana, and Wisconsin have lifted longstanding bans on the construction of new nuclear plants. Texas just approved a $350 million fund to build new reactors.

And in the biggest shift yet, President Trump signed four executive orders aimed at breaking nuclear energy from its regulatory shackles.

Regulation has gotten so ridiculous, it’s forced nuclear plants to limit radiation exposure to 0.0001 millisieverts per year—less than the amount you ingest when you eat a banana. No wonder it’s been nearly impossible to build atomic energy.

After decades of red tape and bureaucratic nonsense, the dam is finally breaking.

AI’s insatiable appetite for energy threw a bucket of cold water over regulators and forced them to wake up.

• AI is nuclear’s best friend.

AI data centers have an urgent need for massive amounts of clean, reliable energy.

Nuclear provides that.

It’s to AI what oil was to the industrial age: the fuel behind a new era of exponential progress.

Why else do you think big tech companies are sprinting around, checkbooks in hand, hoovering up all the atomic energy they can get their hands on?

For example, in the past year:

• Microsoft (MSFT) inked a deal to help restart a reactor at the iconic Three Mile Island plant in Pennsylvania.
• Google (GOOGL) signed an agreement with startup Kairos Power to build several SMRs to power its data centers.
• Amazon (AMZN) entered a 17-year contract with Talen Energy Corp. (TLN) to power its data centers with nuclear energy. It’s also among a consortium investing $500 million into SMR startup X-energy.
• Meta Platforms (META) inked a 20-year deal to buy power from Constellation Energy Corp.’s (CEG) Clinton nuclear plant in Illinois. The deal ensures the Illinois plant will stay operational past its previously scheduled 2027 closing.

A few years ago, I wrote we’d soon see nuclear-powered AI factories. It sounded a little kooky at the time.

Now, it’s happening!

• We’re entering the second atomic age. So, how do we profit from it?

You could buy a uranium producer like Cameco (CCJ). I first wrote about Cameco in 2018.

But Cameco isn’t a great business—and it never will be. It sells uranium, a commodity. That means its fate is tied to the price of uranium, which it doesn’t control.

In the past decade, uranium prices have traded as low as $18 and as high as $100. That’s like running a lemonade stand where the price of lemons changes by 80% every few years.

No matter how well you run the business, your margins are at the mercy of the market.

Instead, I recommend owning picks-and-shovels companies that will benefit from long-term, structural tailwinds. Businesses with pricing power and the ability to grow earnings regardless of where the commodity winds blow.

If you’d like a front-row seat to the nuclear renaissance as it unfolds—as well as other disruptive megatrends reshaping entire industries—The Jolt is where I cover these seismic shifts and the opportunities they create.

Each week, I unpack where the world is headed… and how to invest accordingly.

Go here to join us.

______________

Stephen McBride is Chief Analyst, RiskHedge. To get more ideas like this sent straight to your inbox every Monday and Friday, make sure to sign up for The Jolt, a free investment letter focused on profiting from disruption.