The Man Selling AI's Shovels Just Became a Trillion-Dollar…

Jensen Huang walked onto the Computex stage in Taipei last week wearing his signature black leather jacket, the same one he's worn for years like a rock star who never got the memo that he's now running the most important company on Earth. Nvidia's market cap had just crossed $3 trillion. The crowd cheered. The stock chart looked like a hockey stick glued to a rocket. Here's the uncomfortable part nobody on that stage wanted to say out loud: nobody actually knows if any of this is real yet. Let me be clear about what Huang has built. Nvidia makes the chips that train basically every serious AI model on the planet. OpenAI, Google, Meta, Microsoft, Anthropic—they're all lined up at the same trough. When the AI gold rush started, Huang was the guy selling shovels, and he sold them at a markup that would make a pharmaceutical executive blush. Gross margins north of 70%. A near-monopoly on the hardware that matters. It's one of the great business positions in modern history, and Huang earned it by betting on AI years before it was fashionable. But shovels only sell if someone's still digging. The math getting uncomfortable is this: Nvidia's biggest customers are also its biggest question marks. Microsoft, Meta, Google, and Amazon are collectively spending well over $100 billion a year on AI infrastructure, and the revenue to justify that spending is, to put it gently, still arriving. Meta's AI investments are propping up a stock price more than a profit line. OpenAI reportedly loses money on many of its own products. Somebody eventually has to show a return, or the shovel orders slow down. And when they slow down, they don't slow down gently. Chip demand is notoriously cyclical. Ask anyone who held Cisco stock in 2001. Cisco sold the routers that built the internet, and it was absolutely right about the future. It just got priced for a future that arrived slower than the market wanted. Nvidia is not Cisco—the margins and the moat are genuinely different—but the pattern of "correct about the technology, wrong about the timeline" has buried a lot of smart money. There's also the competition question Huang keeps getting asked and keeps deflecting. AMD is shipping credible AI chips. Google has its own silicon. Amazon, Microsoft, and Meta are all designing custom chips specifically to stop paying Nvidia's prices. That's not a threat tomorrow. It's a threat in the three-to-five-year window that the stock's valuation is already borrowing against. So who benefits from the current narrative? Wall Street analysts who need the momentum to continue. Fund managers who can't afford to be underweight the hottest name. And Huang himself, who has turned a chip company into a cultural phenomenon—keynotes that feel like Apple events, a leather jacket that's become a brand, a personal fortune that's ballooned past $100 billion. None of that makes him a fraud. It makes him a brilliant operator riding a wave that may or may not break exactly where everyone's standing. The honest answer to "is Nvidia a bubble" is that nobody knows, and anyone who tells you they do is selling something. What we do know is that Huang built a real business with real technology and real customers. What we don't know is whether those customers can keep spending like drunken sailors on shore leave forever. My take: Huang deserves every dollar he's made, and the technology is genuinely transformative. But the stock is priced for a world where AI revenue materializes on schedule, competition never catches up, and nothing goes wrong. History suggests at least one of those assumptions will crack. When it does, the man in the leather jacket will still be fine. The people who bought at the top might not be.