The Man Behind the AI Gold Rush Is Selling Shovels

Jensen Huang has become the most consequential CEO in America, and he didn't do it by inventing a chatbot. He did it by selling the shovels. Nvidia, the chip company Huang co-founded in a Denny's booth in 1993, is now worth more than $3 trillion. Its graphics processing units power nearly every major AI model on Earth. When OpenAI trained GPT, it used Nvidia chips. When Google built Gemini, Nvidia chips. Meta, Microsoft, Amazon, xAI — same story. Huang sits at the center of the biggest technology boom since the internet, and he's the one collecting rent. But here's the uncomfortable part. The entire AI economy is basically a loop where the same handful of companies buy from each other. Nvidia sells chips to cloud providers. Cloud providers sell compute to AI startups. AI startups raise money from venture firms and, increasingly, from the very cloud providers that need to buy more Nvidia chips. Microsoft backs OpenAI. Amazon backs Anthropic. Google backs both sides. It's less a market than a circulatory system, and Huang's company is the heart. That doesn't make it a fraud. It makes it fragile. Huang's genius isn't just engineering. It's positioning. He saw a decade ago that GPUs weren't just for video games. He built CUDA, a software layer that trapped developers in Nvidia's ecosystem, and he spent years giving away hardware to researchers who would later demand it at scale. When the AI moment arrived, Nvidia was the only company ready. That's not luck. That's a two-decade bet finally paying off. But every empire invites challengers. AMD is closing the gap. Google and Amazon are designing their own chips. China's Huawei is building alternatives for a market Nvidia can't legally serve anyway thanks to export controls. And Nvidia's biggest customers — the hyperscalers — have every incentive to stop paying Huang's margins, which run north of 70 percent. That's not a sustainable tax forever. There's also the demand question nobody wants to ask out loud. Companies are spending tens of billions on AI infrastructure, and the returns are still mostly theoretical. If the AI bubble deflates — not pops, just deflates — Nvidia is the first domino. Huang knows this. That's why he's been on a global tour, courting sovereign wealth funds, governments, and anyone with a budget and an AI ambition. And let's be honest about who benefits from the hype. Huang does. Every earnings call that beats expectations sends the stock higher. Every breathless headline about AI's transformation is, functionally, a free ad for Nvidia's product line. The man wears a leather jacket to keynote events like a rock star because he understands that in a gold rush, the person selling shovels gets to be the celebrity. None of this means Nvidia is a bad company. It's one of the most impressive businesses of the last thirty years. Huang is a genuine visionary who built something real. But the story being sold to retail investors — that AI is an infinite runway and Nvidia is the only toll booth — deserves more skepticism than it usually gets. The real test isn't whether Huang can keep selling chips. It's whether the customers buying them can ever make the math work without each other. **The bottom line:** Nvidia built the picks and axes for the AI gold rush, and Jensen Huang is laughing all the way to the bank. But when everyone's fortune depends on the same supplier, the supplier's fortune depends on everyone else's delusions holding up. Watch the customers, not just the chips.