Remember 1965 when a single relay in Niagara Falls blacked out 30 million people? Turns out, we’re about to replay that movie—except this time, the culprit isn’t air conditioners. It’s AI.
Here’s the thing nobody wants to admit: the electricity grid wasn’t built for what’s coming. Data centers are already consuming insane amounts of power, and we’re just getting started. The hyperscalers—Amazon, Microsoft, Google—are spending tens of billions on compute capacity, but there’s a problem. The bottleneck isn’t chips anymore. It’s juice.
This is where things get weird. Tech companies are seriously debating whether the answer is nuclear reactors or data centers floating in space. I’m not joking. This is an actual engineering conversation happening right now, and it’s going to determine which companies mint money over the next decade.
The Geopolitical Plot Twist
The Iran blockade threw a wrench into global energy markets. Oil spiked. Inflation expectations shifted. But here’s what most investors missed: defense stocks, energy infrastructure, and AI compute companies were the real winners during the chaos. Now that tensions are cooling (ceasefire talks are actually progressing), the sector rotation is shifting again. The playbook is changing, and most people are still reading the old one.
Orbital Compute: The Sleeper Play
Amazon’s acquisition of Globalstar wasn’t just about satellite internet. It was a chess move. Amazon just locked down the infrastructure to build data centers in space—before SpaceX goes public and Elon Musk gets tens of billions to do the same thing. Think about that for a second.
The full stack of winners spans from compute owners (Amazon, Microsoft, Alphabet, Meta) to rocket companies (Rocket Lab), orbital network operators (Globalstar, AST SpaceMobile), space compute suppliers (Nvidia, GlobalFoundries), and optical interconnect players (Coherent, Lumentum). Whether the future runs on nuclear or orbital compute—or both—the suppliers selling the picks and shovels are positioned to win either way.
The Prediction Market Revolution
Robinhood just became the front door to a new Vegas. Prediction markets are exploding as a real-time probability engine for geopolitical events and macro outcomes. They’re faster than polls, more honest than punditry, and increasingly liquid. This could become the new forward-looking signal that financial markets price around. Robinhood could hit $200 by year-end if prediction markets keep scaling.
The Real Risk Nobody’s Pricing In
Here’s the contrarian take: the biggest threat to AI stocks isn’t valuations or competition. It’s populist backlash. Negative headlines are piling up. Bubble fears. ROI skepticism. The social sentiment is awful and worsening. Once the general public realizes how much capital is flowing into AI, legislation could follow. That’s the real risk.
But here’s why it probably doesn’t matter: the fundamentals are too strong, the capital commitments too deep, and the competitive dynamics too urgent. TSMC just raised guidance to 30%-plus growth and maxed out capex at $56 billion. Companies don’t spend that kind of money on demand they expect to slow.
The Bottom Line
The AI infrastructure buildout is still in early innings. The energy question will reshape economics. Orbital compute is moving from theory to execution. And the companies supplying the infrastructure are positioned to benefit regardless of which path wins.
Stay long, stay selective, and recognize that the next 12 months belong to patience. The thesis has never been stronger.
