How Nvidia CEO Jensen Huang Transformed a Graphics Card Company Into an AI Giant: 'One of the Most Remarkable Business Pivots in History'

A decade ago, Nvidia was a awesome graphics paper maker, vying pinch competitors for illustration AMD and Intel for dominance. Now it's an AI elephantine pinch 70% to 95% of nan marketplace stock for AI chips, and nan brains of OpenAI's ChatGPT. It's besides nan best-performing stock pinch nan highest return successful nan past 25 years.

Why did Nvidia put successful AI chips complete 10 years ago, up of nan competition? CEO Jensen Huang and committee personnel Mark Stevens, Nvidia's 2 largest individual shareholders, talked to Sequoia Capital partner Roelof Botha to explicate what Botha called "one of nan astir singular business pivots successful history."

Nvidia's original merchandise was 3D graphics cards for PC games, but institution leaders noticed by nan mid-2000s that nan PC marketplace was hitting a maturation limit.

Related: Nvidia CEO Jensen Huang Turned Down a Merger Offer successful nan Company's Early Days, According to Insiders. Here's Why.

"We felt we were ever gonna beryllium boxed into nan PC gaming marketplace and ever knocking heads pinch Intel if we didn't create a marque caller marketplace that cipher other was in," Stevens explained.

Jensen Huang, co-founder and main executive serviceman of Nvidia. Photographer: Lionel Ng/Bloomberg via Getty Images

That request for a caller marketplace intersected pinch a merchandise Nvidia already had connected hand: its graphics processor unit, aliases GPU, which could beryllium utilized to powerfulness tasks extracurricular of gaming. Researchers astatine universities crossed nan world began exploring nan graphics cards, yet building precocious computers pinch them.

Related: Is It Too Late to Buy Nvidia? Former Morgan Stanley Strategist Says 'Buy High, Sell Higher.'

Huang recalled gathering a quantum chemist successful Taiwan who showed him a closet pinch a "giant array" of Nvidia's GPUs connected its shelves; location fans were rotating to support nan strategy cool.

"He said, 'I built my ain individual supercomputer.' And he said to maine that because of our work… he's capable to do his activity successful his lifetime," Huang said.

Other researchers, for illustration Meta AI chief Yann LeCun successful New York, began reaching retired to Nvidia astir nan computing powerfulness of its chips. Nvidia began considering nan AI marketplace erstwhile AI had yet to participate nan mainstream and was a "zero cardinal dollar market" aliases a marketplace that had yet to materialize.

"There was nary guarantee that AI would ever really look because, support successful mind, AI had had galore stops and starts complete nan past 40 years," Stevens said. "I mean, AI has been astir arsenic a machine subject conception for decades. But it had ne'er really taken disconnected arsenic a immense marketplace opportunity."

Related: Nvidia Is 'Slowly Becoming nan IBM of nan AI Era,' According to nan Leader of a $2 Billion AI Startup

Huang and different institution leaders still believed successful AI and decided to put billions successful nan tech successful nan 2010s.

"This was a elephantine pivot for our company," Huang said. "The company's attraction was steered distant from its halfway business."

Huang highlighted nan other cost, talent, and skills Nvidia had to relationship for pinch nan pivot, arsenic it affected nan full company. It took 10 to 15 years of effort, but that business determination led to Nvidia powering nan AI gyration pinch an early ChatGPT partnership.

"Every CEO's occupation is expected to look astir corners," Huang said. "You want to beryllium nan personification who believes nan institution tin execute much than nan institution believes it can."

Related: How to Be a Billionaire By 25, According to a College Dropout Turned CEO Worth $1.6 Billion