Here's the thing about the AI race that nobody wants to talk about: it's not really about who has the most money or the smartest engineers. According to "Shark Tank" star Kevin O'Leary, it's about who can keep the lights on.

Energy Is the Real Bottleneck

O'Leary took to X over the weekend with a straightforward message: building next-generation AI data centers requires massive amounts of low-cost, sustainable power, plus the land and skilled workforce to make it happen. And those conditions? They're extremely rare and difficult to secure.

The nations that secure dependable and affordable energy supplies will ultimately win the AI race, he argues. This reframes the entire competition as something far bigger than just buying hardware or training models. You can have all the capital in the world, but if you can't power the infrastructure, you're stuck.

Canada's $70 Billion Megaproject Leads the Pack

O'Leary backed up his argument by sharing a ranking of the world's 20 largest upcoming data center projects. Sitting at number one? His own Wonder Valley AI Data Center in Greenview City, Canada, with a planned value of $70 billion. That makes it the largest AI data center project currently in the planning stage globally.

To put that scale in perspective, it dwarfs other major international efforts. Group 42's Project Stargate in the UAE comes in at $40 billion. South Korea's Jeonnam AI Data Center is targeting $35 billion. A joint project involving Aker Solutions, Nscale, and OpenAI in Norway is planned at $20 billion. Most of these massive developments remain in planning or pre-execution phases, with only a handful actively under construction.

China's Power Advantage Keeps O'Leary Up at Night

O'Leary has been sounding this alarm consistently. He recently warned that China's enormous power infrastructure—not its semiconductor production capabilities—represents the greatest challenge to America's position in the global AI race. The U.S. could fall behind China in AI development due to one crucial factor: electricity availability.

He's not alone in this concern. Nvidia Corp (NVDA) CEO Jensen Huang has repeatedly cautioned that China's AI development could accelerate thanks to lower energy costs and looser regulatory restrictions. When the chip maker most responsible for powering AI infrastructure shares that worry, it's worth paying attention.

The implication is stark. You can restrict chip exports, limit technology transfers, and impose sanctions, but if your competitor can simply plug in more power at lower cost, they've got a structural advantage that's hard to overcome. Energy infrastructure isn't something you can build overnight or import on short notice. It requires decades of planning, massive capital investment, and political will to execute. That's what makes this race particularly challenging for policymakers trying to maintain technological leadership.