Jensen Huang’s chips power both sides of the AI race — that’s becoming a problem

Jensen Huang

Photo: Photographer: Peter Dasilva / CC BY 4.0, via Wikimedia Commons

Born	February 17, 1963 Tainan
Occupation	Entrepreneur
Awards	IEEE Founders Medal u00b7 Queen Elizabeth Prize for Engineering u00b7 Time 100

In June 2026, Jensen Huang turned down Senator Elizabeth Warren’s invitation to explain his company’s China business before a Senate committee. The polite refusal made visible something that had been accumulating since the AI boom started: NVIDIA‘s chips power the American AI buildup and, until recent export controls changed the picture, powered much of China’s as well. Huang argues this is not a problem his company created. Washington disagrees on at least one point — who should be deciding.

Huang was the second son of a chemical engineer and a schoolteacher in Taipei, Taiwan. His parents sent nine-year-old Jensen and his brother to the United States under a misapprehension: they believed Oneida Baptist Institute in rural Kentucky was an elite boarding school. It was not. Huang arrived knowing little English, an undersized immigrant in a school with no one who looked like him, and distinguished himself through sheer competitive focus — in table tennis specifically, well enough to appear in Sports Illustrated at fourteen.

He finished high school in Oregon at sixteen, then spent years working Denny’s graveyard shifts as a dishwasher, busboy, and waiter while completing an electrical engineering degree at Oregon State. He would later earn a master’s at Stanford during evenings while working at LSI Logic, where he designed graphics hardware alongside Chris Malachowsky and Curtis Priem. In 1993, the three of them met at another Denny’s — the choice of venue was not random — and founded NVIDIA, pooling six hundred dollars between them.

The company nearly died before it found its footing. NVIDIA’s initial chips were built around a graphics geometry different from the industry standard. Sega’s five-million-dollar investment bought enough runway to correct the mistake. By August 1997, when the RIVA 128 launched and became the fastest-selling graphics card in history, the company had roughly one month of operating capital left. That proximity to the edge never fully left Huang’s management philosophy. He has described the company’s unofficial motto in those years as “our company is thirty days from going out of business” — and treated urgency as a permanent operating state rather than a response to crisis.

The GPU — NVIDIA coined the term — arrived in 1999, the same year as the company’s IPO, and turned NVIDIA into the dominant force in PC gaming graphics. The GeForce series defined what consumer graphics cards meant for two decades. The next pivotal decision came in 2006 with the launch of CUDA, a platform that made NVIDIA’s GPUs available for general-purpose parallel computing. Within the company, the investment was contested. Outside it, it was largely ignored — until a deep learning model called AlexNet ran on NVIDIA hardware in 2012 and won the ImageNet competition by a margin that forced the AI research community to reconsider what gaming chips could do. The data center business that followed made the gaming business look modest.

What happened after ChatGPT launched in 2022 requires a different vocabulary. NVIDIA went from supplying AI infrastructure to becoming the effective bottleneck that determined how fast the AI industry could move. Demand for H100 chips exceeded supply by ratios that made compute allocation a geopolitical instrument. In October 2025, NVIDIA became the first publicly traded company to exceed five trillion dollars in market capitalization.

The export-controls question is where Huang’s position becomes genuinely difficult to inhabit cleanly. He has called U.S. chip restrictions on China a “loser’s mentality,” arguing that China already controls sixty percent of global chip manufacturing capacity and that cutting it off from American hardware accelerates the development of a parallel Chinese technology ecosystem rather than preventing one. In May 2026 he acknowledged that NVIDIA had “largely conceded” the China AI chip market to Huawei as a result of those restrictions. When Senator Warren invited him to testify on the matter in June 2026, he declined. The same year, he accepted appointment to President Trump’s Council of Advisors on Science and Technology — a position that places him inside the administration whose chip export policy he publicly calls a failure, institutionalizing a tension rather than resolving it.

His GTC 2026 keynote in March outlined five hundred billion dollars in projected GPU demand for the Blackwell architecture and identified a potential two-hundred-billion-dollar market in CPUs for AI agents. IEEE awarded him its Medal of Honor in January 2026; Time named him among the Persons of the Year for 2025, as one of the “Architects of AI.” He delivered Carnegie Mellon’s commencement address in May, telling graduates they were entering the workforce “at the beginning of the AI revolution.” His philanthropic foundation, seeded with NVIDIA shares in 2007, now holds over twelve billion dollars in assets and distributed more than a hundred million dollars in grants in 2024.

Huang has been married to Lori Mills since 1985; they met as lab partners in an Oregon State engineering class. Their two children, Spencer and Madison, both work at NVIDIA — a detail that reads either as dynastic or as the most literal possible expression of the company as life project. A late-2024 investigation revealed he had structured estate planning instruments estimated to reduce his family’s estate tax liability by eight billion dollars, a disclosure that generated headlines without substantially altering his public standing.

His PCAST role means Huang now advises on U.S. AI policy from inside the same administration whose chip export rules he calls a mistake. The Blackwell and Rubin chip architectures extend NVIDIA’s computational lead several years ahead. Whether the man who spent thirty years building the world’s most indispensable hardware company can navigate its geopolitical position with equal skill is the problem his next chapter will answer.

Tags: artificial intelligence, business, nvidia, ceos, business leaders, Jensen Huang

Related People

Related People

Discussion