It's All About Energy

How Energy Became the Ultimate Commodity in the AI Arms Race

In case you missed it, here is the previous article:

Dear Merchants,

If you’re not positioned in energy infrastructure right now, you’re about to miss the biggest wealth transfer since the railroad boom

Here’s something that will make you rethink everything about the tech boom the AI revolution It’s about power. Raw, unglamorous, power.

While everyone’s been obsessing over which company has the smartest algorithms or the fastest chips, a seismic shift has occurred beneath our feet. The constraint on AI dominance is whether you can keep the lights on.

When NVIDIA’s Jensen Huang arguably the most powerful CEO in tech today publicly praises someone, who does he celebrate? Not a software genius or chip designer. He lavishes praise on Energy Secretary Chris Wright for bringing “a surge of energy, a surge of passion” to the Department of Energy. And Secretary Wright himself puts it even more bluntly: “the next big thing MUST be power” (video here)

That tells you everything you need to know about where we are in this story.

Let’s start with the present reality and build from there, because the evidence is overwhelming once you know where to look.

The United States operates 10 times more data centers than Germany, the 2nd largest market globally, with these facilities currently consuming 4% of total US electricity a share projected to exceed 10% by 2030. That’s equivalent to powering entire nations. We’re talking about restructuring the entire energy economy.

Now here’s where it gets really interesting, data center construction spending has tripled since 2022, hitting a record $14 billion in July 2025 and pushing spending to $26.9 billion, representing a 366% increase from 2021 levels. Average construction costs per facility have soared to $220 million.

But that’s just scratching the surface. Alphabet, Amazon, Meta, Microsoft, and OpenAI have announced $800 billion in commitments for new data centers in 2025 alone. Eight hundred billion dollars. In one year. For buildings that do nothing but consume electricity and process data.

That should make every investor sit up straight, for the first time in history, data center construction is on track to surpass general office construction spending, accounting for 32% of new office construction spending in the US compared to just 5% a decade ago. We’re witnessing a fundamental reordering of infrastructure priorities, where productivity is measured not by how many humans you can pack into glass towers, but by how much compute power you can deploy and how many electrons you can push through silicon.

The winners in this new world aren’t necessarily in Silicon Valley anymore. Louisiana leads all states with over $10 billion in data center investment over the past twelve months, driven primarily by Meta’s $27 billion Hyperion campus, followed by Texas with $7.1 billion, Virginia with $6.4 billion, Wisconsin with $3.7 billion, and Arizona and Georgia each exceeding $2 billion.

Notice the pattern? These aren’t tech hubs. These are energy hubs. The bottleneck has shifted from access to coding talent to access to electrical infrastructure!

Location is once again becoming destiny, but for entirely different reasons than in previous industrial revolutions.

Now, I know what you’re thinking. This sounds like every other tech bubble. How is this different from the dot-com crash or the crypto mania?

Here’s the fundamental difference, and it’s why I believe this boom has real staying power … the money is real

Today’s AI infrastructure is being built primarily with operating cash flow, not speculative capital or debt, with the Magnificent Seven spending over $400 billion annually on data center capacity while continuing to generate massive free cash flow. Meta alone generated approximately $50 billion in free cash flow over the last twelve months. These aren’t startups burning venture capital on the promise of future profitability. These are cash-generating behemoths building hard assets.

This is General Electric and US Steel in 1900. This is infrastructure that will define economic power for generations.

But here’s where the story takes an interesting turn that actually makes me more bullish, not less, debt is starting to enter the equation.

Before you panic, let me explain why this is good news. Meta structured a $27 billion joint venture with Blue Owl Capital for the Hyperion Data Center campus the largest private-credit deal in history!

BlackRock’s ETFs invested over $3 billion in the associated bonds, while PIMCO purchased approximately $18 billion.( Institutional money the smart money is pouring in)

Why does this matter? Because debt creates discipline. It forces companies to be strategic, to prioritize, to avoid the kind of overbuilding that characterized previous bubbles. Meta is sharing risk, preserving cash flexibility, and keeping massive infrastructure investments off its balance sheet. This is smart capital allocation for the long game.

Now let’s zoom out to the global picture, because this is where energy, AI, and geopolitics converge into something far bigger than a simple technology story.

AI has become the new global arms race, with capital expenditures eventually to be funded by governments, particularly in the US and China. If you’ve been wondering why gold, silver, and bitcoin have soared to record highs, this is your answer: markets are pricing in the monetary debasement required to fund this arms race.

The real constraint are the Physical infrastructure. The question is no longer about capital availability it’s about whether the US can build the hundreds of new nuclear power plants necessary to power the AI infrastructure demanded by these commitments by 2028-2030.

China added 429 gigawatts of new power capacity in 2024. China is adding power generation capacity at nearly four times the US rate. OpenAI warned the White House in October 2025 that China’s commitment to building new energy generation could give it an edge in the AI race.

China’s cumulative installed power generation capacity reached 3.72 billion kilowatts by the end of September 2025.

The country that controls the energy infrastructure will control the technology, and the country that controls the technology will write the rules for global commerce and security.

So how is America responding? With a return to the only baseload power source that can realistically meet the demand: nuclear.

The Trump administration announced an $80 billion partnership with Westinghouse, Cameco, and Brookfield Asset Management to construct at least 10 new large AP1000 reactors by 2030, representing the most ambitious nuclear build in decades and producing over 10 gigawatts of new nuclear capacity. The US government will secure financing and permits in exchange for a 20% share of future profits. This structure aligns incentives perfectly.

But we can’t wait a decade for new reactors. Google and Microsoft are entering direct agreements to restart dormant nuclear plants.

Yes, you read that right, tech companies are literally restarting nuclear power plants. That’s how serious the energy constraint has become.

Nuclear is the destination, but natural gas is the bridge we’re crossing right now.

Energy Secretary Chris Wright declared that “this is what unleashing American energy and American entrepreneurship looks like,” with the Trump administration on track to double LNG exports by 2030, and potentially double that again in the next 5-10 years if demand is there.

US LNG exports hit a record 9.33 million tons in August 2025, with Europe importing 6.16 million tons that month alone, while Energy Secretary Wright granted final approval to Venture Global’s CP2 plant in Louisiana to export 28 million metric tons per year.

There’s no free lunch, of course.Now electricity costs as much as 267% more than five years ago in areas near significant data center activity. Residential consumers are feeling the pain.

This creates political pressure, which creates opportunity. The companies that can deliver reliable, affordable electricity to both AI infrastructure and residential consumers will be the winners. Those that can’t will be regulated out of existence.

Even the oilfield services giants see which way the wind is blowing. As drilling activity slows with oil prices in the low $60s per barrel, SLB, Halliburton, and Baker Hughes are pivoting to data centers and AI infrastructure, now supplying distributed power generation equipment, turbines, grid systems, and digital platforms to hyperscale data centers.

Baker Hughes’ Industrial & Energy Technology division boasted $4 billion in new orders and a $32 billion backlog in Q3 2025. These companies are pivoting toward opportunity.

💰The Merchant’s Play:

So how do you actually play this?

Let me walk you through the most interesting stocks