Artificial intelligence, in case you hadn’t heard, is about to take your job, upend society and, after that, perhaps destroy all humanity. And the man who has thrust us into this head-spinning new age? A Taiwanese immigrant who used to clean lavatories.

Until last month, most people outside techland would not have heard of Jensen Huang. Nor would most people have heard of Nvidia, the company he founded. Yet Nvidia is now one of the most valuable companies in the world. And Huang, if the hype is to be believed, is in the process of joining the software billionaires Bill Gates and Steve Jobs as a titan of the modern age.

Who is this overnight sensation? Huang is the 60-year-old Taiwanese-American whose Californian company is the maker of the ultra-powerful computer chips that have emerged as the engine of the AI revolution. In a single day last month, Nvidia’s shares shot up as much as 29 percent, gaining $260 billion in value. It was a staggering rise, one of the largest yet seen on Wall Street.

Days later the company’s value passed $1 trillion, joining the handful of companies such as Amazon and Apple to have reached that threshold before its stock fell slightly.

In a single day last month, Nvidia’s shares shot up as much as 29 percent, gaining $260 billion in value.

The surge came after Huang said that he expected Nvidia sales to outstrip market expectations by a whopping 50 percent. For companies as big as Nvidia, increases of that magnitude just don’t happen. Unless, of course, you are the sole provider of the specialized — and very expensive — hardware required to build a potentially world-altering technology.

Everyone is in a ferment about how ChatGPT and similar bots will change the world. But those bots need powerful, specialized chips to function. And Huang has the chips.

The surge capped quite the fortnight for Huang, known for his characteristic black leather jacket and a humility that has endured despite his success. His net worth has ballooned to $35 billion, wedging him in the global rich list between Phil Knight, who founded Nike, and MacKenzie Scott, the ex-wife of the Amazon founder Jeff Bezos.

Nvidia’s Grace Hopper Superchip will help meet the surging demand for generative A.I.

The story of how Huang ended up here, selling the picks and shovels in the AI gold rush, stretches back decades to civil conflict in Thailand. Huang was born in Taiwan but his parents relocated to work in Thailand, where a bloody uprising had broken out. In an interview with The Sunday Times in 2021, he recalled “tanks were rolling down the streets … grenades are going off. It’s a full-on battle.”

So in 1973, when Huang was nine and his brother was ten, his parents made the difficult decision to send their children halfway across the world and follow them later. The Huang boys landed at a small boarding school called Oneida Baptist Institute, in the mountains of eastern Kentucky. The children also had to take jobs; Huang’s was cleaning the bathrooms.

Yet Huang has credited it as the place that taught him grit, to get back up when he was knocked down. “The kids were really tough. They all had pocketknives — and when they get in fights, it’s not pretty,” he recalled in an interview. “The ending of the story [though] is I loved the time I was there.”

Huang, a gifted student, was lured to Silicon Valley. In 1993, aged 30, and with the engineers Curtis Priem and Chris Malachowsky, he started Nvidia over a meal at Denny’s, a cheap and cheerful diner chain on the wrong side of San Jose. The start-up invented the graphics processor unit, niche chips that allowed PC gaming to take off by enabling richer graphics.

Nvidia nearly went bankrupt early on when its first chip flopped, but found its footing and began to thrive in its one, specialized lane.

Huang, though, had grander plans. He saw potential in AI and went all in. It turned out that the same chips that were needed to render a 3D video game were well suited to training AI’s so-called “deep-learning” systems. “We risked everything to pursue deep learning,” Huang said.

The risk has paid off. Everyone from Google to Meta and Amazon began snapping up Nvidia’s chips as they began to dive deeper into AI. But it wasn’t until the November release of ChatGPT that the world truly realized that the age of AI had suddenly arrived.

Nvidia nearly went bankrupt early on when its first chip flopped.

Nvidia was perfectly poised. AI models are “trained” by being fed vast amounts of data, and then drawing on the lessons they have learned to spit out answers to queries. These systems require a symphony of chips processing information in parallel to pull this off. It is as if an entire data center is working together as one computer. Nvidia has spent years developing the hardware and software to power these systems, which has given it a big lead over rivals now scrambling to catch up.

In 1973, when Huang was nine, his parents sent him halfway across the world, from Thailand to boarding school in eastern Kentucky.

Huang’s role in the AI revolution does not come without caveats, however. In September, the US government banned Nvidia from selling its chips to China because of concerns about its geopolitical rival winning the AI arms race.

Huang has complained about the company having its “hands tied behind our back” by America’s restrictions. “If [China] can’t buy from the United States, they’ll just build it themselves,” he said in an interview.

There’s also the problem that no one, including the companies building these systems, truly understands how Nvidia’s AI models work. They have begun to exhibit “emergent properties” — industry parlance for unexpected or inexplicable actions.

Recently, the Center for AI Safety, a campaign group, published a letter signed by Gates and Geoff Hinton, the psychologist and computer scientist known as the “godfather of AI”, warning that the technology should be viewed as a “societal-scale” risk alongside pandemics and nuclear war.

This was not the first such letter from experts warning of our imminent demise at the hands of robot overlords. But Huang’s name has not appeared on any of them.

In a recent speech, Huang reflected on Nvidia’s long and remarkable journey. “At Nvidia, I experienced failures,” he said. “Great big ones. All humiliating and embarrassing. Many nearly doomed us.”

If the AI pessimists are to be believed, it is Huang’s success that could doom us all.

Danny Fortson is the West Coast correspondent at The Sunday Times of London and the host of the weekly podcast Danny in the Valley