Jensen Huang

Jensen Huang has run Nvidia since the day it was founded, in 1993, at a roadside restaurant. He spent thirty years betting that the chips designed to draw video-game graphics could do far more. When artificial intelligence arrived, they turned out to be exactly what it needed, and Nvidia became the most valuable company in the world.

Taiwan, Thailand, Kentucky

He was born Jen-Hsun Huang on 17 February 1963 in Tainan, Taiwan. His father was a chemical engineer and his mother a teacher. When he was a child, the family moved to Thailand for his father’s work.

Amid political unrest in Thailand, his parents sent him and his older brother to the United States, to stay with an uncle in Tacoma, Washington. He was nine and spoke little English. The uncle enrolled them at the Oneida Baptist Institute in rural Kentucky, believing it to be a good boarding school. It was in fact a strict school for troubled boys. Huang, the youngest pupil, was given the job of cleaning the toilets and shared a room with an older student who had a knife wound. He has said that he learned to be tough there, and that the boy taught him to read and he taught the boy to do push-ups.

His parents later moved to Oregon, and the family was reunited. He went to Aloha High School near Portland, was a nationally ranked junior table tennis player, and graduated at sixteen. As a teenager he worked at Denny’s, the restaurant chain, washing dishes and waiting tables.

Oregon State and Stanford

He studied electrical engineering at Oregon State University, graduating in 1984. His lab partner was Lori Mills. They married in 1985 and have two children, Spencer and Madison, who both later worked at Nvidia.

He designed microprocessors at AMD and then worked at LSI Logic, while studying part time for a master’s degree in electrical engineering at Stanford, which he completed in 1992.

The Denny’s

In 1993 he met two engineers from Sun Microsystems, Chris Malachowsky and Curtis Priem, at a Denny’s in east San Jose. They believed that personal computers would one day need special chips to draw realistic 3D graphics, especially for games. On his thirtieth birthday Huang agreed to be chief executive. Nvidia was founded in April 1993.

The first product, the NV1, released in 1995, was built on a way of drawing 3D shapes that the rest of the industry did not adopt. It failed. A contract with the game-console maker Sega went wrong. Nvidia laid off much of its staff and was, Huang has said, about thirty days from going out of business.

The company bet what was left on a new chip, the RIVA 128, released in 1997. It worked, and it sold. Nvidia went public in January 1999, and later that year released the GeForce 256, which it marketed as the world’s first “GPU”, or graphics processing unit. The whole story, from the company’s side, is told in the site’s Nvidia business profile.

The CUDA Bet

A graphics chip does something unusual. Instead of performing one calculation at a time very quickly, it performs thousands of simple calculations at once, one for each pixel. Researchers realised that this made GPUs very good at other kinds of maths too.

In 2006 Nvidia launched CUDA, software that let programmers use its graphics chips for general calculations. Huang put CUDA on every Nvidia chip, including cheap gaming cards, even though it added cost and most customers would never use it. For years Wall Street questioned the spending. Nvidia’s share price stayed flat for long stretches.

In 2012 a team at the University of Toronto used two Nvidia gaming cards to train a neural network, AlexNet, that won a major image-recognition contest by a wide margin. It showed that GPUs could make deep learning practical. Huang turned the company towards it. In 2016 he personally delivered Nvidia’s first DGX AI computer to a small research lab in San Francisco called OpenAI.

The Engine of AI

When OpenAI launched ChatGPT in November 2022, every large technology company raced to build AI systems, and nearly all of them needed Nvidia’s chips and the CUDA software that ran on them. Nvidia’s data-centre chips, the Hopper and then Blackwell generations, were in short supply for years.

Nvidia’s market value passed $1 trillion in 2023, $3 trillion in 2024, $4 trillion in July 2025 and $5 trillion in October 2025. In August 2026 it reported quarterly revenue of more than $96 billion, and its next generation of systems, Vera Rubin, began shipping. Huang, who still owns about 3 per cent of the company, became one of the richest people in the world.

He is instantly recognisable in his black leather jacket, and known for long, technical keynote speeches that he delivers without notes. He runs Nvidia with an unusually flat structure, with dozens of executives reporting directly to him, and says he prefers to share information widely rather than in small meetings.

Risks

Nvidia’s success has brought new pressures. The US government has restricted the sale of its most advanced chips to China, once a large market. Its biggest customers — Google, Amazon, Microsoft, Meta — are designing their own chips to reduce their dependence on it. And investors regularly debate whether the spending on AI data centres can continue at its current pace.

Networks and a Deal That Failed

Huang understood early that AI would need not just chips but whole systems: thousands of chips linked together and working as one. In 2020 Nvidia completed the purchase of Mellanox, an Israeli maker of high-speed networking equipment, for about $7 billion. It became central to Nvidia’s data-centre business.

Not every deal worked. In 2020 Nvidia agreed to buy the British chip designer Arm from SoftBank for about $40 billion. Regulators in the United States, Britain and Europe objected, and the deal was abandoned in 2022.

He has given generously to the universities that trained him. Stanford’s engineering centre bears his name, and in 2022 he and his wife gave $50 million to Oregon State University for a new research and supercomputing complex. He has said that when Nvidia’s share price first reached $100, he had the company’s logo tattooed on his arm.

How He Manages

Huang has described a management style that deliberately avoids some standard practices. He says he rarely holds one-to-one meetings, preferring to give feedback in front of groups so that everyone can learn from it, and he asks staff across the company to email him short lists of the most important things they are working on, which he reads in large numbers.

Nvidia also never abandoned gaming. It introduced real-time ray tracing for games with its GeForce RTX chips in 2018, and its chips power Nintendo’s Switch consoles. The gamers who bought its cards for decades paid for the research that later powered AI.

Huang has also argued that countries will want their own AI computing capacity, rather than relying entirely on American cloud providers, and has travelled widely to sell that idea. Governments and national telecoms companies in Europe, the Middle East and Asia have since placed large orders.

The measure of him is patience with a long bet. For more than a decade he invested in making Nvidia’s chips programmable for work that barely existed. When that work arrived, no one else was ready.

Common Questions

How This Was Written

Corrections are welcome and are made without argument: commissions@99founder.com.