If you're worried about an AI bubble, Kevin O'Leary has news for you: you're looking at the wrong problem entirely.

The "Shark Tank" star and prominent investor addressed bubble concerns in a video posted on X this Monday, and his take was pretty straightforward. AI isn't some speculative frenzy anymore — it's already woven into the fabric of the American economy, from real estate to financial services and everything in between.

Why O'Leary Isn't Worried About a Bubble

"No, I don't think we're in a bubble," O'Leary said when asked whether AI's rapid ascent resembles past market manias. His reasoning? The technology has already demonstrated real value across all 11 economic sectors.

The past two years have transformed AI from a buzzword into a legitimate productivity tool that actually improves margins, he explained. Companies aren't just experimenting anymore — they're deploying AI at scale because it works.

But here's where things get interesting. That success has created an entirely different crisis.

The Power Grid Can't Keep Up

"Now, we got a really big problem and it's called the grid. There's no power left," O'Leary warned, pointing to the explosive demand coming from data centers and AI computing infrastructure.

Think about it: AI models require massive computational power, which means massive electricity consumption. As adoption accelerates, the bottleneck isn't chips or software or talent — it's finding enough electricity to keep the lights on.

This isn't theoretical hand-wringing. It's already reshaping where and how companies can build AI infrastructure.

O'Leary's Energy Play: Betting on Bitzero

O'Leary is putting his money where his mouth is. He's invested in Bitzero, a company he describes as part real estate operator, part energy supplier, and part permitting specialist.

The company's CEO, Mohammed Bakhashwain, has been locking in long-term electricity contracts at under six cents per kilowatt-hour in Nordic countries like Finland and Norway. With those agreements secured, Bitzero can supply cheap, reliable power for operations ranging from Bitcoin (BTC) mining — which generates cash flow — to AI data centers.

It's a classic infrastructure bet. While O'Leary still holds positions in chip makers like Nvidia Corp (NVDA), he sees energy as the more fundamental asset in the AI revolution.

Energy as the New Infrastructure Play

"I also want to own the infrastructure of AI and that is power… if you don't have power, you got nothing. Bitzero has Power," O'Leary stated.

He's framing electricity as the modern equivalent of picks and shovels during a gold rush. Everyone needs it, supply is constrained, and whoever controls access to cheap, reliable power has a structural advantage.

It's a compelling thesis, especially as traditional power grids strain under unprecedented demand.

China's Power Subsidy Strategy

Meanwhile, Beijing is taking a different approach to the same problem. Earlier this month, reports emerged that China has ramped up subsidies for large data centers, slashing power bills by as much as 50%.

Local governments in data-center-heavy regions like Gansu, Guizhou, and Inner Mongolia are rolling out these incentives, primarily benefiting tech giants like ByteDance, Alibaba Group Holding Ltd. (BABA), and Tencent Holdings Ltd. (TCEHY).

Why the sudden generosity? These companies have been dealing with surging electricity costs after Beijing banned them from purchasing Nvidia chips. The ban forced them to use domestically produced processors from companies like Huawei Technologies and Cambricon, which are significantly less power-efficient.

Tech firms complained loudly about the higher operating expenses tied to these homegrown chips, and the government responded with aggressive energy subsidies to keep them competitive.

China's Grid Advantage

Here's the kicker: even with the added energy demands of less efficient domestic chips, China's centralized power grid still delivers electricity that's cheaper, cleaner, and more reliable than what's available in the United States.

That advantage has transformed power-rich remote regions into major hubs for large-scale data center development, giving Chinese companies a geographic and economic edge in the AI infrastructure race.

It's a reminder that in the AI era, national competitiveness might come down to something as unsexy as electrical capacity and grid stability.

The infrastructure underneath the AI boom matters just as much as the algorithms on top of it. And if O'Leary is right, the real money might be in the wires and power plants nobody's thinking about.