The Nvidia Doctrine

Welcome to The Closer, where we reveal how power really works—the hidden dynamics, the unwritten rules, the patterns that repeat across industries and eras.

This week at CES, Jensen Huang took the stage to declare that "the ChatGPT moment for physical AI is here." But to understand why that matters—and why Nvidia is now the most valuable company on Earth—you have to understand the man behind it. This is the story of how Jensen Huang built his empire not by chasing hype, but by having the conviction to endure silence.

Every Friday, we go deep on one story that reveals how power really works. Subscribe to never miss a Deep Dive. Subscribe with your e-mail for free.

When Jensen Huang launched CUDA in 2006, the audience was complete silence. Nobody wanted it. Nobody asked for it. Nobody understood it. Nvidia's valuation fell from $12 billion to somewhere between $2 billion and $3 billion. The company had just doubled the cost of its chips to include this new software layer, and the market's verdict was swift and brutal: silence.

Today, nearly every major AI model on the planet runs on hardware that depends on CUDA (Compute Unified Device Architecture). OpenAI's GPT models, Google's Gemini, Anthropic's Claude, Meta's Llama—all of them train on Nvidia GPUs using CUDA. The silence turned into a $5 trillion company, the most valuable on Earth.

That silence is the key to understanding Jensen Huang, and the source of his immense power. In a world that chases hype, Huang has built the world's most valuable company by making bets that draw no applause, orders, or even immediate understanding. He invests in what he calls "zero-billion dollar markets."

This is the story of the Nvidia Doctrine: a playbook for building power by having the conviction to withstand the quiet. It's a story that begins not in Silicon Valley, but in a rural Kentucky boarding school, and it offers a profound lesson for anyone trying to build something that lasts.

The Janitor and the Knife Fight

The CEO of the world's most valuable company didn't learn about America through elite universities or tech incubators. His education started at the Oneida Baptist Institute in Clay County, Kentucky—one of the poorest counties in the country, then and now.

Jensen Huang arrived there in the mid-1970s, sent by his parents from Thailand to escape political instability during one of the country's periodic coups. He was nine years old. He and his brother were alone without their parents. The boys had been sent to live with an uncle in Tacoma, Washington, who was tasked with finding them a school. He found Oneida Baptist Institute, believing it to be a prestigious boarding school. It was not.

At nine years old, Jensen was the youngest student on campus. His assigned job was to clean the toilets for a hundred teenage boys. His roommate was a 17-year-old wrapped in tape from a recent knife fight—"the toughest kid in school," Huang later recalled. His brother was sent to work the tobacco fields the school ran to fund itself. "Kind of like a penitentiary," Huang told Joe Rogan in a recent podcast.

The students smoked constantly. Huang tried it too, at age nine, for about a week. The dorms had no closet doors or locks. Disputes were settled with knives. And through all of this, Jensen and his brother were navigating alone. International phone calls were a luxury they couldn't afford. Communication with their parents in Bangkok was a monthly cassette tape, mailed back and forth across the Pacific. Their parents would record a message, mail the tape to Kentucky, and the boys would record over it and mail it back.

In one of those tapes, a young Jensen described his first visit to McDonald's to his parents. "Mom and Dad, we went to the most amazing restaurant today," he said. "This whole place is lit up. It's like the future. And the food comes in a box, and the food is incredible. The hamburger is incredible."

Two years later, his parents finally made it to America with suitcases and almost no money. His mother worked as a maid. His father, a trained engineer, found work by circling job openings in newspaper classifieds and calling whoever picked up. "They left everything behind," Huang said. "They started over in their late thirties."

He still carries one memory from those early years that, he says, "breaks my heart." Shortly after his parents arrived, the family was living in a rented, furnished apartment when Jensen and his brother accidentally broke a flimsy particleboard coffee table. "I just still remember the look on my mom's face," he said. "They didn't have any money, and she didn't know how she was going to pay it back."

This early experience—of being an outsider, of navigating a world he didn't understand, of having to make sense of things for himself—forged the core of what could be called the Nvidia Doctrine.

The Denny's Booth and the Sega Bet

Nvidia was born in a Denny's booth in San Jose in 1993. Huang, who had worked as a dishwasher and busboy at Denny's while in school, met with two friends—Chris Malachowsky and Curtis Priem—to sketch out an idea for a new kind of computer chip. They wanted to build a chip that could render 3D graphics on a personal computer.

It was a bet on a zero-billion-dollar market. At the time, the market for 3D graphics on PCs was essentially nonexistent. Video games were played on consoles. Professional graphics work was done on expensive workstations. The idea that ordinary people would want—or need—3D graphics on their home computers seemed far-fetched.

The company's early years were a struggle. Nvidia burned through its initial funding building chips that the market didn't want. By the late 1990s, the company was on the brink of bankruptcy. They had a contract to build chips for Sega's Dreamcast console, but Huang realized that the architecture they were developing was a dead end. The technology wouldn't scale. It was the wrong bet.

What happened next is a masterclass in negotiation from a position of apparent weakness.

Huang went to the CEO of Sega and told him they were going to stop working on the chip. Then he asked to be paid the full amount of the contract anyway—for a product Sega would never receive. The CEO agreed. That money saved Nvidia and allowed them to build the GeForce 256, the chip that put them on the map. When you have no leverage, sometimes the only move is to ask for the impossible. Huang wasn't negotiating from a position of strength. He was negotiating from a position of conviction.

"If he said no, we would have gone out of business," Huang later explained. The Sega CEO got nothing tangible in return. But Huang's conviction—his absolute certainty that he was right about where the technology needed to go—was persuasive enough to win the deal.

The CUDA Silence

The Sega bet was just a warm-up. The defining moment for the Nvidia Doctrine came in 2006 with the launch of CUDA. It was a software platform that would allow developers to program Nvidia's GPUs for general-purpose computing, not just graphics.

The idea came from an unexpected place. Researchers at universities had started using Nvidia's graphics cards for scientific computing. They had figured out that the parallel processing architecture of GPUs—designed to render millions of pixels simultaneously—was also excellent for certain kinds of mathematical calculations. They were hacking the graphics cards to do physics simulations, molecular modeling, and other compute-intensive work.

Most companies would have ignored this. It wasn't the intended use case. The market for scientific computing was tiny compared to gaming. But Huang saw a pattern. If researchers were going to this much trouble to use his chips for computing, maybe the future of computing itself was parallel.

So Nvidia built CUDA. It roughly doubled the cost of their chips. It required a massive investment in software development. And when they launched it, the market's reaction was brutal.

"When I launched CUDA, the audience was complete silence," Huang recalled. "Nobody wanted it. Nobody asked for it. Nobody understood it."

The company's stock cratered. Nvidia's valuation fell from about $12 billion to between $2 billion and $3 billion. Wall Street analysts questioned whether Huang had lost his mind. Why was a graphics card company investing in scientific computing software?

Huang held his nerve. He knew that the future of computing wasn't just about graphics; it was about parallel processing. He was betting that the world would eventually need to process so much data, so quickly, that traditional CPUs wouldn't be able to keep up. He was right—but it would take nearly two decades for the world to catch up to his vision.

The DGX-1 Delivery

Ten years after CUDA, in 2016, Huang was still making bets that nobody understood. He launched the DGX-1, billed as the world's first AI supercomputer. It was a $129,000 machine designed specifically for training deep learning models.

Again, silence. No purchase orders. The AI winter had only recently thawed. Deep learning was still an academic curiosity. The idea that companies would pay six figures for a specialized AI training machine seemed absurd.

Then, a single phone call. It was from Elon Musk, who said he needed a machine for a small nonprofit AI lab he was involved with.

"All the blood drained out of my face," Huang told Rogan. "A nonprofit is not buying a $300,000 computer."

But Musk insisted. So Huang did something remarkable: he boxed up one of the first DGX-1 units, loaded it into his car, and drove it to San Francisco himself. He walked into a cramped upstairs office filled with a handful of researchers—including Berkeley robotics pioneer Pieter Abbeel and a young researcher named Ilya Sutskever. That cramped office was OpenAI, long before it became the most discussed AI organization in the world.

Huang left the DGX-1 with them and drove home. He had just hand-delivered the infrastructure that would help launch the AI revolution.

The Physical AI Pivot

At CES this week in Las Vegas, Huang took the stage flanked by two BD-1 droids from the Star Wars universe. It was theatrical, but the message was serious.

"The ChatGPT moment for physical AI is here," he declared, "when machines begin to understand, reason and act in the real world."

Huang unveiled a new generation of chips and software designed to power autonomous vehicles, robots, and factories. He announced Alpamayo, which he called "the world's first thinking, reasoning autonomous vehicle AI"—trained end-to-end, "literally from camera-in to actuation-out." He showed a demo of a Mercedes-Benz CLA driving itself through San Francisco, avoiding pedestrians and taking turns, running entirely on Nvidia's software stack.

He announced the Rubin platform—six new chips that will be available in the second half of 2026. He talked about Caterpillar construction equipment and Agibot humanoid robots, all running on Nvidia technology.

"There's no question in my mind now that this is going to be one of the largest robotics industries," Huang said. "Our vision is that someday every single car, every single truck will be autonomous."

He's no longer just building the infrastructure for AI in the cloud. He's building the infrastructure for AI in the physical world. He's betting on another zero-billion-dollar market. And this time, the world is listening.

The Doctrine

What can we learn from Jensen Huang's approach to building power?

The Nvidia Doctrine isn't about being smarter than everyone else. It's about having a different relationship with time and uncertainty. It's about being willing to make bets that won't pay off for years—or decades—and having the conviction to hold them through the silence.

Mark Stevens, a Sequoia Capital partner who has known Huang for decades, summarized it this way: "Jensen always likes to say, 'We're investing in zero-billion-dollar markets.'" The logic is counterintuitive but powerful. If you invest in a market that already exists, you're competing with everyone else who can see the same opportunity. If you invest in a market that doesn't exist yet, you have time to build the infrastructure, the ecosystem, the moat—before anyone else realizes what's happening.

But this only works if you're right about where the world is going. And being right requires a kind of first-principles thinking that most people find uncomfortable. It means ignoring what the market is telling you today and focusing on what physics, or mathematics, or human nature tells you must eventually be true.

Huang learned this in a Kentucky boarding school, cleaning toilets while his roommate recovered from a knife fight. He learned it mailing cassette tapes to parents he couldn't afford to call. He learned it watching his mother's face when they broke a coffee table they couldn't pay for.

He learned that the world doesn't owe you anything. That silence isn't rejection—it's just the sound of being early. That conviction, held long enough, becomes its own kind of leverage.

If you’ve read this far, we want to hear from you! We hope The Closer is becoming an essential part of your week – an insider guide to the deals that are changing our world. Ideas, feedback, or just want to chat? Write us.

Please get in touch. Until next time, keep making moves and thinking outside the box. Deals change the world – and so can you.