The Architects of AI: Jensen Huang

Jensen Huang has been waiting for 30 years. And now, quietly and without fanfare, the wait is over.

Every AI system that has generated fear, excitement, controversy or capital in recent years, OpenAI’s GPT models, Anthropic’s Claude, Meta’s Llama, Musk’s Grok, runs on Nvidia chips. Every data centre that Altman is building under the Stargate initiative, every server farm that Zuckerberg is filling with $115 billion of capital expenditure, every model that Amodei is training while simultaneously warning about its consequences, all of it depends on the hardware that Huang’s company designs and manufactures.

Nvidia commands between 80-90% of the global AI chip market. There is no meaningful AI without Nvidia. And yet Jensen Huang, whose personal net worth as of early 2026 stands at approximately $180 billion dollars, making him one of the ten wealthiest people in recorded history, is the least discussed figure in any serious analysis of how AI is reshaping the world.

That absence is not accidental. Understanding it is the key to understanding Huang.

The Beginning: Taiwan, Thailand, and a Reform School in Kentucky

Jensen Huang, born Huang Jen-Hsun, came into the world on 17 February 1963 in Tainan, Taiwan’s fourth largest city and its former capital, a place of temples, street food and deep historical consciousness. His father, Huang Hsing-tai, was a chemical engineer. His mother, Lo Tsai-hsiu, was a grade school teacher who taught her sons English by selecting ten new words from the dictionary and drilling them every day, an act of practical foresight whose discipline would shape everything that followed.

When Jensen was five, the family relocated to Thailand, following his father’s engineering work. When he was nine, the situation changed dramatically. Regional instability connected to the Vietnam War, combined with limited educational options for their children in Thailand, prompted his parents to make a decision that Huang has described as agonising. In 1972, they sent their two sons, Jensen, nine, and his older brother, eleven, to live with an uncle in Tacoma, Washington, as unaccompanied minors while they remained behind to resolve the family’s affairs.

The uncle, newly arrived in Washington and unfamiliar with the American educational landscape, was searching for an affordable boarding school for his nephews. He found the Oneida Baptist Institute, located in Clay County, rural Kentucky, one of the poorest counties in the United States. He enrolled the boys believing it to be a prestigious boarding academy. It was, in fact, a reform school for troubled boys.

Jensen Huang was nine years old, spoke limited English, and had left his parents on the other side of the Pacific. His first night at Oneida, his roommate was a 17 year old whose body was wrapped in tape from a recent fight. The dormitories had no closet doors, no locks and a population of boys who smoked constantly and settled disputes with knives. To get to school each morning, students crossed a swinging rope bridge over a river, the wooden planks rotted and missing. Local boys would grab the ropes when Huang was crossing and shake them, trying to throw him into the water below.

Every student had a job. His brother was sent to the tobacco fields the school ran to fund itself. Jensen was assigned to clean the bathrooms for a hundred teenage boys every single day. He was called racial slurs. He was bullied relentlessly. He had no counsellor, no support structure and no immediate prospect of his parents arriving.

He was there for approximately two years before his parents, having finally immigrated to the United States, discovered the situation and removed their sons. The family reunited and settled in Oregon.

Huang has spoken about this period in interviews with a quality that observers consistently describe as striking: he does not narrate it as trauma. He narrates it as formation. “Back then, there wasn’t a counsellor to talk to,” he said. “Back then, you just had to toughen up and move on.” When he describes crossing the rope bridge and seeing the boys try to shake him off, he says: “It never seemed to affect me. I just shook it off.” In 2019 he donated $2 million to Oneida Baptist Institute and had a new building named after him. He spoke fondly of the footbridge at the ceremony, declining to mention the boys who had tried to throw him from it.

The self-editing is as revealing as the story itself. Huang’s relationship with his own suffering is one of transformation, a refusal to allow an adverse experience to become a self-limiting identity. “Greatness is not intelligence,” he told students at Stanford. “Greatness comes from character. And character isn’t formed out of smart people. It comes from people who have suffered.”

Once settled in Oregon, Huang attended Aloha High School, where he graduated two years early, and worked from the age of 15 at a local Denny’s restaurant as a busboy and waiter, scraping plates and serving coffee to the diners of suburban Portland. He placed third in junior doubles at the US Open Table Tennis Championship at fifteen, demonstrating the combination of practice, competitive intensity and pattern recognition that would later define his professional career.

He enrolled at Oregon State University, graduating in 1984 with a bachelor’s degree in electrical engineering. He then completed a master’s degree in electrical engineering at Stanford University while working full-time in Silicon Valley at Advanced Micro Devices and later at LSI Logic. At Stanford he met Lori Mills. They married in 1985 and have two children. By all available accounts it is a stable, private and genuinely contented domestic life.

The Money: From a Denny’s Booth to Three Trillion Dollars

In 1993, at the age of thirty, Jensen Huang met two colleagues, Chris Malachowsky, who had worked at Sun Microsystems, and Curtis Priem, who had worked at IBM, at a Denny’s restaurant in East San Jose to found a company. He had worked at Denny’s as a teenager. He would later say that the choice of venue was unconscious rather than symbolic. Whether or not that is true, it is the kind of origin story that acquires meaning in retrospect.

They founded the company with $40,000 in initial capital raised between them, working initially out of Priem’s townhouse with around a dozen colleagues who agreed to work without salaries while funding was secured. They initially called it NVision before discovering the name belonged to a toilet paper manufacturer. They settled on Nvidia, from the Latin invidia, meaning envy.

Their original vision was specific and modest: to build dedicated graphics processing chips that could handle the complex geometry calculations required for three-dimensional computer games, which were then becoming commercially significant. The CPU, central processing unit, that powered most personal computers was a general-purpose processor, efficient at sequential calculations but not at the massively parallel mathematical operations that three-dimensional graphics required. A specialised chip, Huang and his co-founders believed, could transform gaming.

They were right, though not in the way they anticipated. The early years were genuinely precarious. The company’s first major product was designed for Sega, which subsequently changed its hardware specifications, nearly destroying Nvidia before it had properly started. The company survived on a combination of Huang’s relentless customer development and a fortunate contract with Sony that provided enough revenue to reach its next milestone.

In 1999, Nvidia coined the term graphics processing unit, GPU, with the launch of the GeForce 256, the first chip to integrate transform and lighting calculations on a single processor. The chip transformed gaming. By the mid-2000s Nvidia had established itself as the dominant force in gaming graphics alongside ATI, which was subsequently acquired by AMD.

But Huang had seen something larger. In 2007 Nvidia launched CUDA, Compute Unified Device Architecture, a software platform that allowed GPUs to be used for general-purpose computing beyond graphics. The GPU’s strength was parallel processing: the ability to perform thousands of calculations simultaneously, which was useless for the sequential operations of most business software but extraordinarily powerful for scientific simulation, machine learning and, it would eventually prove, the training of large neural networks.

For years, CUDA was used primarily by academic researchers in physics, biology and climatology, for whom its parallel processing capabilities offered dramatic speedups in simulation. Nvidia was patient. Huang invested in the research community, offering academic pricing, sponsoring conferences and building software tools that made CUDA progressively more accessible. The company was, in effect, seeding a market that did not yet exist in commercial form.

When OpenAI released ChatGPT in November 2022 and the world suddenly understood what large language models could do, the market that Nvidia had been seeding for 15 years exploded. The H100 chip, Nvidia’s flagship AI accelerator, became the most sought after piece of hardware in the world. Larry Ellison of Oracle publicly admitted that he and Elon Musk had been “begging” Huang for H100s. The waiting list for Nvidia’s chips extended to years. The company’s revenue grew from $27 billion in fiscal 2023 to $60 billion in fiscal 2024. Its market capitalisation crossed $1 trillion in June 2023, $2 trillion in February 2024, and $3 trillion later that year, briefly making Nvidia the most valuable company in the world.

In the first quarter of 2026, Nvidia reported revenue of $81.6 billion, 85% from the same period the previous year. Data centre revenue alone reached $75.2 billion, up 92% year on year. These figures make Nvidia's earlier financial milestones look modest by comparison and confirm what Huang has been saying for several years: that AI infrastructure spending is not a bubble but a structural shift in how the world's computing resources are organised.

Huang’s personal wealth, almost entirely derived from his 3.5% percent stake in Nvidia, grew from approximately $3 billion in 2019 to over $180 billion in 2026, an increase of roughly 60 times in 7 years. It is, in proportional terms, one of the most extraordinary personal wealth accumulations in modern economic history.

He continues to sign women’s clothing at fan events. He wears a signature black leather jacket to every public appearance, which he has explained by saying it is one of the few materials that does not make his skin itch. The combination of extraordinary wealth and unpretentious personal style is genuine rather than performed. He is, by the accounts of people who have worked closely with him for decades, the same person he was when Nvidia was struggling to survive.

The Ideas: What Jensen Huang Actually Believes

Huang is measured in his public statements which makes the task of characterising his beliefs somewhat more demanding than for his peers. He does not write long essays or make alarming pronouncements from Senate hearing rooms. He speaks at conferences, gives commencement addresses and conducts interviews in which he is consistently thoughtful, occasionally profound and notably careful about the limits of what he is willing to claim.

On AI and what it will do

Huang’s understanding of AI is technical and first-principles rather than philosophical. He has described three scaling laws driving AI progress; pretraining, where more data, more parameters and more compute produce predictably better models; post-training, which refines models through compute-intensive techniques; and test-time compute, which provides dynamic compute for real-time reasoning. He has predicted, at the 2024 SIEPR Economic Summit at Stanford, that in approximately five years AI will be able to pass every professional exam a human can take, and that in ten years the computational capabilities of AI systems will be a million times larger than they are today.

He has described this not with alarm but with excitement. “I cannot imagine a more exciting time to begin your life’s work,” he told Carnegie Mellon graduates in May 2026. “When society engages technology openly, responsibly and optimistically, we expand human potential far more than we diminish it.”

His view on AI and jobs, repeated in various forms across multiple public appearances, is that: “You’re not going to lose your job to AI. You’re going to lose your job to someone who uses AI.” The responsibility, in Huang’s framing, lies with the individual to adapt, not with the technology to be constrained or with the institutions to manage the transition.

At the NTU commencement in 2023 he was more expansive: “AI will create new jobs that didn’t exist before. Automated tasks will obsolete some jobs, and AI will change every job. While some worry that AI may take their jobs, someone who is an expert with AI will.” The statement is empirically defensible in the long run and functionally silent on what happens in the transition, to the steelworker who cannot retrain, to the entry-level analyst whose position was eliminated before they developed the judgment to operate at a higher level, to the communities whose defining economic activity disappears before replacement activity arrives.

At GTC Taipei in June 2026, Huang went further. "AI is no longer a single breakthrough or application," he said. "It is essential infrastructure. Every company will use it. Every nation will build it." The statement is the clearest articulation of his positioning strategy in the public record of AI as infrastructure, as fundamental and as invisible as electricity or water. Infrastructure that, by definition, nobody questions and nobody turns off. And infrastructure that, by definition, belongs to whoever built it first and built it best.

📍 VIDEO: NTU Commencement Speech 2023

[Covers AI, entrepreneurship, near-death Nvidia stories and his philosophy of suffering and resilience.]

On responsibility and what it means

Huang’s most substantive public statement on the responsibilities that come with building foundational AI infrastructure came at the CMU commencement of 2026: “The responsibility of our generation is not only to advance AI but to advance it wisely.” He called on engineers and policymakers to advance capability and safety in step, pressing the same arguments used by Anthropic, OpenAI and Microsoft about the necessity of American AI leadership relative to authoritarian competitors.

What is notable about this statement, and about Huang’s public positioning on AI responsibility more generally, is what it does not say. It does not name the specific consequences of AI-driven job displacement. It does not engage with the question of how productivity gains will be distributed. It does not address the Frances Haugen problem, the structural tendency of powerful technology companies to know about harm and continue regardless. It speaks in terms of capability and safety as technical challenges rather than as political and distributional ones.

This is not evasion in any simple sense. It reflects a genuine intellectual position: that the primary responsibility of the people building transformative technology is to build it well and to ensure that democratic societies rather than authoritarian ones lead its development. The distributional questions, who benefits, who suffers, who decides, are, in this framing, questions for policymakers and societies rather than for technologists.

Whether that division of responsibility is adequate to the scale of the transformation Huang is enabling is the central question his profile raises.

📍 VIDEO: CMU Commencement 2026

[His most recent and most substantive public statement on AI responsibility, reindustrialisation and what the current generation of engineers owes to the societies they are transforming.]

On leadership and suffering

Huang’s philosophy of leadership is unusually honest about difficulty. At Stanford’s SIEPR summit he told students: “For all of you Stanford students, I wish upon you ample doses of pain and suffering.” He has said in multiple forums that “greatness comes from character, and character comes from people who have suffered.” He talks openly about Nvidia’s near-death experiences, the wrong architecture choice for Sega, the bet-the-company decision on CUDA, the retreat from mobile, as formative rather than as embarrassments to be minimised.

This philosophy, shaped unmistakably by the Oneida experience that a nine year old processed not as trauma but as formation, produces a leadership style that several observers have described as unusually resilient and unusually demanding. He runs Nvidia with no direct reports, all senior leaders report to him simultaneously, and is known for the directness and rigour of his feedback. He has said that “the most efficient way to live is to have transparency” and that he tries to tell people exactly what he thinks rather than managing their feelings.

📍 VIDEO: Stanford SIEPR Economic Summit 2024

[Covers the ten year AI roadmap, predictions about professional exams and his signature “I wish you pain and suffering” message to students.]

There is a particular kind of power that does not announce itself. It does not seek the stage or the controversy. It does not found political movements or fund the dismantling of democratic norms. It does not ask whether your life has meaning without work or warn of white-collar bloodbaths from a CNN studio. It simply builds the infrastructure upon which everything else depends, and then waits.

The Character Question: The Infrastructure of Everything

Jensen Huang’s story is of a nine year old sent alone to a reform school in rural Kentucky who cleaned bathrooms, was threatened with knives, was called racial slurs daily, crossed a sabotaged rope bridge every morning and processed all of it not as injustice but as material for character formation and whose professional story is, in its quiet way, the most consequential of all. His power is one that build the infrastructure upon which everything else depends, and the waits.

Without Nvidia’s chips, none of the other members of the Silicon Seven could do what they do. Altman cannot train GPT models. Amodei cannot develop Claude. Zuckerberg cannot build his Superintelligence Lab. Musk cannot run xAI. Every dollar of the $500 billion in Stargate funding requires Nvidia hardware to be meaningful. Every warning about AI displacement, every essay about machines of loving grace, every prediction of white-collar bloodbaths, every question about whether your life has meaning without work, all of it is downstream of the chips that Huang’s company designs.

This is what makes his relative silence about the consequences of AI for ordinary people a significant form of silence. When Huang says “you won’t lose your job to AI, you’ll lose it to someone who uses AI,” he is making a claim that is simultaneously accurate, reassuring and incomplete. It is accurate in the long run. It is incomplete about the transition, about the person who loses their job before they have developed the skills to use AI, about the community whose defining industry disappears before a replacement emerges, about the distributional question of who captures the productivity gains that Nvidia’s chips are enabling.

He is aware of these questions but his choice not to engage with them is positionin. It is a decision about what role the maker of the hardware should play in the debate about its consequences. But is this positioning adequate?

Nvidia commands 80-90 % of the AI chip market. Every significant AI system in the world depends on its hardware. The concentration of that much consequential infrastructure in the hands of one company, led by one person who believes that the responsibility for AI’s consequences lies with policymakers and societies rather than with technologists, is either the appropriate division of labour in a democratic society or the most significant accountability gap in the entire AI ecosystem.

Huang is the quiet one. The one without the provocations, the political projects or the apocalypse preparations.

It turns out he was the one who built the ground everything else stands on.