How Nvidia CEO Jensen Huang Turned His Video Game Graphics Company Into an AI Titan

Nvidia’s revolutionary innovation – the Graphics Processing Unit – powers humanoid robots, the design of virtual movie sets and the creation of protein-based drugs to fight disease.

Insatiable demand for Nvidia’s GPU technology, used for artificial intelligence, has pushed the company into the stock market stratosphere, where it joined Microsoft, Apple and Alphabet – Google’s parent company – to become one of four American companies to have ever reached the top. $2 trillion in market valuation.

Nvidia followed this up in March, when CEO Jensen Huang unveiled Blackwell, the company’s latest GPU. Designed in America, but made in Taiwan, Huang said Blackwell was the fastest chip ever designed and tailored for AI.

“We hope it will do things that surprise us,” Huang said. “Exactly.”

Nvidia’s surprising debut at a Denny’s

Nvidia’s futuristic campus sits right next to the company’s modest birthplace: a Denny’s in San Jose, California. As a teenager, Huang, now 61, worked as a dishwasher at a Denny’s. Thirty years ago, it was at a Denny’s in San Jose, where he and two friends, co-founders Chris Malachowsky and Curtis Priem, imagined a whole new way of dealing with video game graphics and decided to create a company.

At the time, Huang was a 30-year-old electrical engineer, married with two young children. He and his two co-founders at Nvidia didn’t know how to start a company, but they went for it anyway.

Their big idea was to speed up the processing power of computers with a new graphics chip. Their initial attempt failed and nearly bankrupted the company in 1996, but they were able to change course and ultimately create their revolutionary GPU.

Just eight years after Nvidia’s emergence at Denny’s, the company earned a spot in the S&P 500. Huang then set his sights on developing the software and hardware for a revolutionary GPU-driven supercomputer that would bring business well beyond video games. It seemed like a risky bet for Wall Street, but for early AI developers it was a revelation.

“It was a chance based on a vision,” Huang said.

The technological capability they invented was perfect for AI researchers, he said. In 2016, Huang delivered Nvidia’s AI supercomputer, the first of its kind, to then-OpenAI board member Elon Musk, who used it to create the building blocks of ChatGPT. When AI took off, so did Huang’s reputation.

He is now a Silicon Valley celebrity. As someone who immigrated to the United States from Taiwan at age 9, Huang said he could never have imagined his success.

“It’s the most extraordinary thing, Bill, that a normal dishwasher boy could become this,” Huang told 60 Minutes correspondent Bill Whitaker. “There’s no magic, it’s just 61 years of hard work every day.”

People who work with Huang describe him as demanding, a perfectionist and not easy to work for. Huang said all of these descriptions ring true to him.

“It should be like this,” he said. “If we want to do extraordinary things, it shouldn’t be easy.”

AI advances made with Nvidia technology

At Nvidia’s annual developer conference in March, the mood wasn’t just optimistic, it was downright giddy. More than 11,000 enthusiasts – software developers, tech moguls and lucky shareholders – descended on San Jose’s professional hockey arena to kick off a four-day AI extravaganza.

Huang showed some of what AI has made possible in just the last few years, such as an AI-powered simulation of Earth’s weather that will ultimately be able to calculate and predict the weather 3,000 times more faster than a supercomputer while using 1,000 times less energy. » said Huang.

Pinar Seyhan Demirdag, co-founder of Cuebric, leverages Nvidia GPUs to instantly transform simple text prompts into virtual movie sets for a fraction of the cost of some current backgrounds. She said her business gets a lot of love from Hollywood.

At Generate:Biomedicines, Dr. Alex Snyder – head of research and development – ​​uses Nvidia’s technology to create protein-based medicines. She was initially skeptical about using AI for drug development. Then she looked at the lab data and changed her mind.

His team now asks its AI models to create proteins to fight diseases like cancer and asthma. AI generates proteins that do not exist in nature, which are then rigorously tested in the laboratory. A drug to beat the coronavirus is in clinical trials.

“We don’t put Frankensteins into people,” Snyder said. “We take what is known and we really push the field, we push the biology.”

Figure, a Silicon Valley startup funded by Nvidia, has developed a humanoid robot driven by Nvidia GPUs. CEO Brett Adcock says the robot is designed to address labor shortages and wouldn’t be possible without Nvidia’s technology.

“We think they are arguably the best in the world at this,” Adcock said.

Its prototype isn’t ready yet, but early results are so promising that German automaker BMW plans to begin testing the robot at its South Carolina factory this year. Adcock envisions a future with billions of robots working alongside humans.

Addressing concerns about AI

Investors are bullish on Nvidia, but others worry that AI technology is flying too high. Last year, more than 600 AI scientists, ethicists, and other leading AI scientists signed a statement urging caution, warning of the risk of AI to humanity. Some fear that AI will take jobs as technology advances.

“I think over time AI and robotics will start to do more and more of what humans can do and better,” Adcock said.

According to Huang, when companies become more productive, profits increase, which means they can hire more workers.

“You always want humans to know, because we have good judgment,” Huang said. “Because there are circumstances that machines just don’t understand, they just don’t.”

Huang sees a future of progress and prosperity for AI, not one where machines are our masters. For him, AI is a revolutionary technology.

“We need artificial intelligence to help us explore the universe in places we could never do on our own,” Huang said.