A week and a half ago, an AI earthquake hit.

DeepSeek, an unknown Chinese firm, built AI software that rivals the best US models—ChatGPT, Claude, Gemini. But here’s the kicker: DeepSeek’s latest model, R1, was reportedly developed for just $5.6 million—a rounding error compared to the $600 million OpenAI allegedly spent on ChatGPT.

It’s like a souped-up Camry outrunning a NASCAR at the Daytona 500.

No one saw this coming, and now the industry is scrambling to understand how the Chinese did it. But there’s no debate—DeepSeek is a breakthrough, a shot across the bow in the AI race.

The market responded violently. Nvidia cratered 17% in a day, wiping out $600 billion in market cap—nearly the size of Brazil’s entire stock market. Meanwhile, DeepSeek’s AI Assistant—completely unheard of three weeks ago—is now the #1 most downloaded app on the Apple Store. And unlike ChatGPT’s $20/month subscription, it’s free.

A tech investor friend told me he would have “backed the truck up” to invest in OpenAI at a $150B valuation six months ago. Now? He’d short it if he could. His thesis? AI models like ChatGPT and Claude will soon be ubiquitous and free—just like Netscape and Firefox are today. AI will change the world, but the companies that spent hundreds of billions building it may not be the ones who profit.

This entire episode is another reminder of the breakneck speed of technological change.

I recently saw a chart tracking the Top 20 global stocks from 1989 to today (recreated below). The turnover is astounding. Just one company—Exxon—has managed to stay on the list. In 1989, Japanese banks and energy giants reigned supreme. IBM was the only tech company in the mix. By 1995, Japan had largely vanished—only Toyota remained, hanging on at #15.

Technology breeds creative destruction, and AI is accelerating that cycle. AI was born in the U.S., but it will be copied and commoditized worldwide. My bet? At least 2-3 of the Top 20 global companies a decade from now haven’t even been created yet—and at least a quarter will be based outside the U.S. Energy, medicine, agriculture, and transportation are all ripe for AI-driven transformation, and the companies leading those revolutions will be the next trillion-dollar giants.

So how does an investor stay ahead of the shifting sands?

If black swans like DeepSeek are emerging with increasing frequency, trying to pick the winners—even just 3-5 years from now—is a fool’s errand. In 1989, you might have bet on IBM mainframes, GE appliances, and Japanese banking dominance. You would’ve been wrong.

The solution? John Bogle, the founder of Vanguard, said it best: “Don’t look for the needle in the haystack. Just buy the haystack!”

Because here’s the truth: No one knows which companies will dominate AI, just like no one saw Apple, Amazon, or Nvidia coming decades ago. The best defense against disruption isn’t trying to outguess the future—it’s broad, global diversification.

History proves that innovation destroys incumbents, but markets keep growing. AI will create unimaginable wealth, just not always where we expect. Stay invested, spread your bets, and let time do the heavy lifting.