The AI Energy Paradox: Can the Grid Keep Up?

Alex Rivera 4

Welcome to the show, everybody! I'm Alex, here with Emily. And Emily, I want to start today with a number that frankly makes my head spin: 30.8 million megawatt-hours.

Emily Nguyen 4

Wait, 30.8 million megawatt-hours? For context, that's more electricity than entire small countries use in a year. What is consuming that much power?

Alex Rivera 4

That is exactly what Google's data centers consumed in twenty twenty four. Just one company. And that number is roughly double what they were using just four years earlier. We are looking at a fundamentally new era of energy demand, and it's all being driven by artificial intelligence.

Emily Nguyen 4

Doubling in four years. That completely upends the traditional grid-plus-renewables model they've championed for the last decade. Because you can't just slap a few solar panels on a server farm that size and call it a day. AI models require continuous, uninterrupted training. They need twenty four seven, three hundred sixty five-day power.

Alex Rivera 4

Exactly. And because of that need for constant power, we are seeing some wild shifts in strategy. Just this past January, a company called Crusoe Energy filed a permit to build a nine hundred thirty three-megawatt natural gas power plant right on site at a new Google data center campus in Texas.

Emily Nguyen 4

A nine hundred thirty three-megawatt gas plant just for one campus? That's the size of a traditional municipal power station. But here's the part that really concerns me from an environmental science perspective: according to the permit documents, this specific Texas plant has no carbon capture technology attached to it. None. The turbines could emit up to four point five million tons of CO2 a year.

Alex Rivera 4

Four and a half million tons of CO2. And this is Google—the company that essentially pioneered the concept of twenty four seven carbon-free energy. They are spending up to one hundred eighty five billion dollars in capital expenditures in 2026, mostly on AI infrastructure. They are so desperate for electricity that they are pivoting to what analysts are calling an "Everything Everywhere All at Once" power strategy. Wind, solar, nuclear, geothermal, and yes... straight, unfiltered natural gas.

Emily Nguyen 4

Which brings us to the central paradox of our time. We have this cutting-edge technology that promises to optimize our world, maybe even help us solve climate change, but its immediate power hunger is dragging us right back to traditional fossil fuels.

Alex Rivera 4

And this dynamic isn't just playing out in Texas. It's literally redrawing the global energy map. If you look at Africa right now, they have about 1.2 gigawatts of total installed data center capacity. But they account for nearly 20 percent of the world's population.

Emily Nguyen 4

So 20 percent of the population, but only 0.6 percent of global data center capacity. That is a massive infrastructure gap.

Alex Rivera 4

Massive. But the demand for data centers there is projected to increase up to 5.5 times by 2030. And how are they planning to power it? Natural gas. Nigeria, for example, is sitting on one of Africa's largest reserves—over 200 trillion cubic feet of gas. For decades, the strategy was to export that gas. Now, the African Chamber of Energy is pushing to use it domestically to power AI and digital infrastructure.

Emily Nguyen 4

Using that 200 trillion cubic feet locally to break external digital dependence. It makes complete economic sense for them. Intermittent sources like wind and solar struggle with the unstable grids in many emerging markets. They need what we call "baseload" power—the reliable, steady foundation of electricity that doesn't drop off when the sun sets.

Alex Rivera 4

"Baseload." That really is the word of the year in the energy sector right now. And it's creating massive friction here in the U.S. too. Get this: earlier this year, between April 29th and May 6th, the price of natural gas at the Waha Hub in West Texas actually dropped below negative two dollars per million British thermal units. Negative two dollars!

Emily Nguyen 4

They were literally paying people two dollars to take the gas away? How does that even happen?

Alex Rivera 4

Because they're pumping so much oil in the Permian Basin, and the gas comes up with it as a byproduct. But they don't have the pipeline infrastructure to move the gas to where the data centers and the demand actually are. So you have a glut of gas in Texas, while places like New England are starved for pipeline capacity and paying premium prices. It's a complete logistical bottleneck.

Emily Nguyen 4

And that bottleneck is exactly where the conversation needs to pivot. Because we're stuck in this binary mindset: either we stall the AI revolution because we can't build renewable grids fast enough, or we accept millions of tons of new carbon emissions from traditional fossil gas to power these 900-megawatt campuses. But there is a third path.

Alex Rivera 4

The third path. This is what we spend our days working on at 3 Rivers Energy Partners. Emily, tell them about Renewable Natural Gas, or RNG.

Emily Nguyen 4

RNG is the bridge. Biologically and chemically, it is exactly the same as traditional natural gas—methane. But the difference is where it comes from. Instead of drilling miles into the earth to extract new fossil carbon that hasn't seen the atmosphere in millions of years, RNG captures methane that is already being emitted right now by organic waste.

Alex Rivera 4

Like agricultural waste, dairy farms, distilleries—stuff that's just rotting and releasing methane into the sky anyway.

Emily Nguyen 4

Exactly. Methane is roughly 25 times more potent as a greenhouse gas than carbon dioxide. By capturing it from a dairy farm or a landfill, refining it, and putting it into a pipeline, you are taking a massive climate liability and turning it into that prized 24/7 baseload power. You get the operational stability of gas, but with a net-neutral, or sometimes even net-negative, carbon footprint.

Alex Rivera 4

And from a business perspective, the beauty of RNG is that it is a pure plug-and-play solution. Remember those pipeline constraints we just talked about? Building a new interstate natural gas pipeline takes years of fighting through the Federal Energy Regulatory Commission—FERC. But RNG facilities can often be decentralized. You process the gas right where the agricultural waste is, and you inject it into existing local utility networks.

Emily Nguyen 4

Bypassing the FERC interstate bottleneck entirely, while utilizing the three million miles of pipeline we already have in the ground.

Alex Rivera 4

Exactly. At 3 Rivers Energy Partners, we are working with major agricultural and industrial players to scale this technology up. Because when you look at that 7.3 percent projected growth in natural gas electricity generation by twenty twenty seven—driven almost entirely by data centers—we have to ensure as much of that as possible is coming from renewable sources.

Emily Nguyen 4

Exactly. We've been looking at the AI energy boom as a climate disaster waiting to happen. But if tech giants like Google, channel even a fraction of those billions in capital expenditures toward renewables like RNG, AI could actually become the financial catalyst that scales the circular energy economy.

Alex Rivera 4

That's the frontier we're building toward. It's not about turning off the servers; it's about fundamentally changing what keeps them humming. Thanks for joining us for this deep dive, everybody. We'll catch you on the next one.