With AdQuick, you can now easily plan, deploy and measure campaigns just as easily as digital ads, making them a no-brainer to add to your team’s toolbox.

You can learn more at www.AdQuick.com

Listen, I need to tell you about something wild that just happened on the Joe Rogan podcast. NVIDIA CEO Jensen Huang—the guy whose chips power basically every AI system you've heard of—just casually dropped that tech companies are going to be building their own nuclear reactors within six or seven years. Not looking into it. Not exploring options. Building them.

Here's the thing: this isn't some futurist fantasy. This is happening because the alternative is that AI development just... stops.

The numbers here are absurd. A single NVIDIA Blackwell GPU pulls 1,200 watts. Modern AI racks are hitting 240 kilowatts—enough to power 200 American homes from one rack. A serious AI training cluster needs 500 megawatts running 24/7. That's a mid-sized city's worth of continuous power.

By 2030, US data centers will jump from 4% to potentially 12% of national electricity consumption. The deployed H100 chips alone will consume as much electricity as the entire country of Georgia. The entire country.

And here's the kicker: the grid can't deliver it. Not won't—can't. There's plenty of power generation out there, but the distribution infrastructure is maxed out. Upgrading transmission lines takes five to ten years. AI companies need power now.

Yeah. And they're already signing the checks.

Microsoft is reopening Three Mile Island. Google signed a deal with Kairos Power for 500 megawatts of molten salt reactors by 2035, with the first one online in 2030. Amazon dropped $650 million to buy a data center campus right next to a nuclear plant, securing access to 960 megawatts. Oracle's Larry Ellison is planning a one-gigawatt data center powered by three small modular reactors.

This is real money moving right now. The technology exists—NuScale just got regulatory approval for their 77-megawatt design in May 2025. They're the only SMR vendor with full NRC approval, and they're on track for deployment by 2030.

Here's what Huang understood that makes this whole thing click: companies will "build as much power as you need, and you can contribute back to the grid." Tech giants are becoming energy companies. They're not just consumers anymore—they're producers who can sell excess capacity back to utilities.

This privatizes what used to be public infrastructure. When Microsoft owns a nuclear plant and Google operates a fleet of reactors, that's a different world than everyone sharing the same electrical grid. It's about energy independence, sure, but it's also about competitive advantage. If your AI training runs on reliable nuclear baseload and your competitor is fighting for grid capacity, you win.

The geopolitical angle matters too. Huang explicitly tied this to US competitiveness: "If the United States doesn't grow, we will have no prosperity." China is building 150 new reactors by 2035. This is an arms race where the ammunition is electricity.

There's one problem: uranium. Global production only meets 80-90% of current reactor demand. The US needs 50 million pounds annually but produces maybe 5 million domestically. Industry projections show a potential one billion pound deficit over the next 15 years.

So while tech companies are racing to build reactors, they might be racing into a supply bottleneck that nobody's priced in yet.

Six to seven years. That's the timeline. By 2032, your cloud services might be powered by a reactor that Google built. Wild times.