Old Man Coal Lost Its Merry Old Soul — But Big Tech Knows Better

We’ve got to keep every plant open.

That’s more than a mantra these days out of President Trump’s White house regarding our upcoming power crisis. 

And do you want to know just how desperate they are to meet tomorrow’s grid demand? 

Last week, the administration announced a $700 million initiative to support coal plants across America.

Naturally, the announcement came with all the fanfare, from the Defense Production Act authority to 13 existing plants getting the cash to modernize their plants, along with two brand-new coal plants in Alaska and West Virginia (mind you, they’d be the first ones built since 2013), and Maryland’s Warrior Run coal plant restarting after shutting down in June of 2024.

And that leads us back to the mantra, which was reiterated recently by Interior Secretary Doug Burgum: 

“We’ve got to keep every plant open. If there have been units at a coal plant that have been shut down, we need to bring those back on.”

You know just as well as I do that the White House is framing it as responding to AI data center power demand, with military power purchase agreements locking in coal generation and more job creation in coal country — the whole nine yards. 

But here’s the problem… 

Warrior Run shut down in June 2024 because it couldn’t compete economically. 

You see, it turned out that the independent market monitor for PJM — the grid operator serving 65 million people — said the plant couldn’t provide energy economically.

Folks, restarting the plant with up to 200 million in taxpayer funds doesn’t change that fact. 

Old Man Coal Lost Its Merry Old Soul

Today, the 102 coal units that have been retired in the last five years have a median age of 56 years.

Never forget this fact, no matter what bullish headline pops up in the coming weeks and months. 

Typically, coal plants retire around after 40 to 50 years of operation.

That makes sense, right? By design, five decades of churning out power leads to higher maintenance costs, less efficiency, and operators that start looking closer at balance sheets. 

To be blunt: Our ancient fleet of retired coal plants aren’t exactly perfect candidates for a restart. 

And if the average operating coal plant in America is 45 years old, you can see the clock ticking for the coal industry… can’t you?

Now throw in the fact that the recently retired units are more than a decade older than the fleet average. 

And you can be sure that these coal units weren’t shut down because of politics. 

Even more concerning is that as time goes on, our coal retirements will accelerate. Between now and 2030, another 173 coal units are scheduled for retirement. 

Then, we’ll see another 55 retire between 2031 and 2040. 

Yes, retirements slowed in 2025, with only 2.6 GW retired (the lowest since 2010). 

However, that’s only because the Trump administration dipped his hand into this arena by forcing delays. 

Look, restarting a 56-year-old coal plant extends its life maybe five to ten years  — before maintenance costs and regulatory pressure force closure again.

So, we’re essentially betting that every future administration will keep throwing subsidy money at uneconomical power plants. 

Do we really want to gamble on political hegemony for the next 15 years? 

I don’t think so. 

Unfortunately for President Trump, he can’t know who the next guy in the White House will be, and it’s a good bet a swing to the other political aisle means that federal support for those coal restarts will be in jeopardy. 

Yet, there’s something else happening that nobody’s paying attention to.

Trump’s Coal Bet Won’t Win the AI Race 

We both know that natural gas just got so abundant that it’s worth negative money.

That may sound incorrect, but it’s true. Last Friday, the daily spot pricing for natural gas at the West Texas Waha hub was trading at around negative $0.37 per MMBtu. 

That’s not to mention the fact that the Marcellus and Utica Shale formations in Appalachia are the largest gas-producing region in the country. 

Between the massive natural gas output in the northeast and the associated gas production that comes from our tight oil wells in the Permian Basin, the U.S. is absolutely drowning in natural gas.

And that’s a good thing!

It’s also the sole reason why coal has become so uneconomical. Remember, it wasn’t solar, wind, or any other renewable source that led to the death of the coal industry over the last 15 years — it was cheap natural gas. 

Truth is, it’s become the perfect bridge fuel for the next five to ten years.

Of course, natural gas plants can ramp up quickly, and are also cleaner and cheaper to build than coal. 

Natural gas also plays well when it comes to grid stability. 

But here’s the catch… 

Political pressure on fossil fuels will accelerate as time goes on. More importantly, coal will always play the villain in the headlines. 

And neither is the answer tech companies are actually building toward.

We’ve talked before about the flood of deals coming out recently. 

If you recall, Meta signed a power agreement in January 2026 to build a 1.2 gigawatt nuclear power campus in Ohio specifically to power data centers.

Early site work starts this year, and the first phase could produce power by 2030.

Oklo and other small modular reactor developers have commitments from Meta, Microsoft, Amazon, and Google totaling over $50 billion for nuclear projects.

This doesn’t sound like companies hedging their bets, does it? It sounds like Big Tech is pouring billions in capital to establish their power infrastructure for the next several decades. 

They’re not picking coal, nor are they betting the house on natural gas.

In fact, there are more than 127 global small modular reactor designs in various stages of development.

NuScale’s design got certified by the Nuclear Regulatory Commission; ENTRA1 and TVA are deploying up to 6 gigawatts of SMR capacity — that’s 72 individual modules.

And what about the 11 advanced reactor projects that got selected by the DOE? Well, the ten companies that won those awards are fast-tracking commercial deployment. 

SMRs are factory-built, modular, and provide between 5 and 300 megawatts each. They can also be co-located directly with data centers, which eliminates grid constraints entirely. 

Granted, that’s not to mention the fact that SMRs operate 24/7 without external power (passive safety systems mean no electricity required for cooling in an emergency), and could provide the cleanest baseload in the energy sector. 

If a medium-sized AI data center consumes electricity equivalent to powering 100,000 households, then an SMR could handle that natively.

Sure, coal and natural gas plants handle it too, but they come with political baggage that grows worse every year.

Let’s not mistake Trump’s $700 million coal subsidy as confidence — this is life support for a dying industry, plain and simple. 

Those plants will run a few years longer, then we’ll watch maintenance costs spike and closures resume.

That’s why we’re seeing real capital flowing into nuclear power, led by the biggest tech giants out there planning for the long-term. 

Believe me, they’re not waiting for government subsidies. 

They’re betting on these must-own energy stocks for tomorrow’s power.