Nvidia Just Became the World's Most Valuable Company. Here's…

Nvidia hit a milestone this week that would have sounded absurd five years ago: it briefly became the most valuable company on Earth, leapfrogging Microsoft and Apple with a market cap north of $3 trillion. The stock is up roughly 170% this year alone. If you bought $10,000 of NVDA in early 2019, you'd be sitting on something close to $300,000 today. So the hype is real. The question is whether the business is. Let's start with what Nvidia actually sells. Its H100 and newer Blackwell chips are the shovels in the AI gold rush — the hardware that trains and runs models like ChatGPT, Gemini, and Claude. Demand has been so ferocious that Nvidia can't make them fast enough. Data center revenue exploded 427% year-over-year last quarter. That's not a typo. But here's where a skeptical observer should start squinting. First, the customer concentration problem. A huge chunk of Nvidia's revenue comes from a handful of buyers: Microsoft, Meta, Google, Amazon, and a few others. These are also Nvidia's competitors. Microsoft and Google are designing their own AI chips. Amazon has Trainium. Meta has MTIA. Right now, Nvidia's software ecosystem, CUDA, makes switching painful. That moat is real — but moats erode, and these are the richest companies in history with every incentive to dig. Second, the circular-money smell. Nvidia invests in AI startups. Those startups buy Nvidia chips. Nvidia reports the revenue. Some of this is normal ecosystem-building. Some of it starts to look like a company buying its own demand. When you see Nvidia's stock rise on news that it's funding more AI companies, ask yourself who's really paying whom. Third, the expectations baked into the price. At recent levels, Nvidia trades at a forward price-to-earnings ratio in the mid-40s. That's not dot-com crazy, but it assumes enormous growth continues for years. If AI spending cools — if companies decide the returns aren't there and pull back on capex — Nvidia gets hit first and hardest. Remember, Cisco was the Nvidia of 2000. It made the routers that built the internet. The internet was real. Cisco's stock still took two decades to recover its peak. Fourth, the geopolitics. Roughly 20% of Nvidia's revenue has come from China. U.S. export controls have already cut into that. If tensions escalate, that's a chunk of the business that simply disappears, and no amount of domestic demand fully replaces it overnight. None of this means Nvidia is a bubble or a bad company. It's an extraordinary business with real products, real profits, and a genuine lead. Jensen Huang saw the AI wave years before most of Silicon Valley and bet the company on it. That deserves respect. But the stock price is a different thing from the company. The company can be great and the stock can still be overpriced. Those two facts coexist all the time, and the crowd cheering the loudest usually isn't the one holding the bag when the music stops. **The bottom line:** Nvidia is winning the AI hardware race, but the people winning the most from this frenzy are early investors, insiders selling into strength, and the analysts collecting fees on the way up. If you're buying today, you're not early. You're buying the story everyone already knows, at a price that assumes nothing goes wrong. Do with that what you will — but at least know which seat you're in.