Going all-in with Nvidia: How Jensen Huang’s high-stakes bets paid off
Nvidia cheated death several times. Now it's powering next generation technology from Google, Meta, Amazon and more.
Jensen Huang plays to win or die — and always has.
"When somebody goes all-in, they can do things that you can't," the Nvidia (NVDA) CEO and co-founder told Oregon State University students in 2013. "Go all-in. Don't play to come back. Don't play as if there's another hand. If there's another hand, let that be a surprise, and we'll figure out what to do next."
Huang, speaking from the steps of his alma mater, had already made his next great gamble — this time on AI. A decade later, to say Nvidia has had a good 2023 would be an understatement.
The company’s name — drawn from the Latin word for envy, “invidia” — has never been more apt, as Nvidia became a key player enabling and benefiting from the AI boom. The chipmaker’s graphics processing units, or GPUs, are in dizzying demand from giants like Microsoft (MSFT), Meta (META), Amazon (AMZN), and Alphabet (GOOG, GOOGL).
The problem is simple: If you want to train a large language model, like ChatGPT or Meta’s Llama 2, you need the processing power to do it, and that processing power comes from their graphics chips. These AI models would take decades to train without advanced GPUs, which Nvidia spent years perfecting. While competitors like Intel (INTC) and AMD (AMD) have been caught flat-footed, Nvidia is the only real player in town.
(Right now, there’s such a high demand for their GPUs, that even they are running out of chips.)
Over the last 12 months, Nvidia stock is up more than 300%. And over the past five years, the company's shares are up over 650%.
But Nvidia could have died at the table years ago. Huang, famed for his love of leather jackets, has an almost swashbuckling instinct that's led the company to staggering highs — and dangerous lows.
Amidst the AI gold rush, Nvidia’s past brushes with mortality seem distant. However, according to Acquired podcast co-host Ben Gilbert, they anchor Nvidia’s idiosyncratic identity, which is only now being recognized. (Nvidia was so criminally undercovered that no one has yet to publish a book about the company.)
“Nvidia has gotten to where they are by at least two, maybe three, times completely going all in and betting the company on a new idea that could have killed the company, but they were right every single time,” said Gilbert.
Huang wasn't available to be interviewed for this article.
An existential crisis: 1997
In 1997, James Cameron's "Titanic" was a megahit and the PC market was almost as hot, dominated by the likes of Compaq, Dell, and Hewlett-Packard.
Nvidia, founded four years before, was looking for a win. Started with just $40,000 in the bank and conceived of in meetings at Denny's, Nvidia had yet to release a product that would secure its future.
“They got off on the wrong foot,” said Gilbert, about Nvidia's earliest chips. The NV1, released in 1995, tried to do too much too fast. It was an off-kilter shape, and though the 3D graphics weren't bad, the chip was expensive. Its other functions, like audio, were merely acceptable.
By 1997, Huang and his co-founders, Chris Malachowsky and Curtis Priem, were at an impasse. With just nine months of runway left, Nvidia had to pull one out of the hat — or fold. Huang decided to shell out a third of their remaining cash up front to speed up the development of the Riva 128 chip, the first consumer GPU for 3D graphics.
Soon, Nvidia was out of money and out of time. The Riva 128 went straight into production, with the company's fate in the balance.
It worked.
The chip, which combined 2D and 3D graphics for revolutionary picture quality, changed the standard for gaming graphics. The Riva 128 became Nvidia's first mainstream success, and the company went from near-disaster to listing on the Nasdaq two years later.
Huang and his Nvidia co-founders had pushed in all their chips — and won another hand to play.
Betting the farm: 2006-2017
According to investors, for the better part of a decade, Nvidia wasn’t growing. Between 2009 and 2015, Nvidia’s market cap held relatively flat, lingering between $8.5 billion and $10 billion.
But quietly, Nvidia was stretching its muscles. During this period, Huang made his most controversial gamble: CUDA, or “Compute Unified Device Architecture.” The platform enables GPUs to be used beyond video games for general processing in a range of fields, from physics to auto manufacturing to AI.
Reading this in 2023, that might seem obvious. But turn back the clock by a decade, and central processing units, or CPUs, reigned in computers and phones. Graphics cards like Nvidia’s were siloed off and dismissed as gamer tech. The 2012 ImageNet competition, a vaunted software contest, showed that GPUs could ultimately do more than CPUs, with the right infrastructure. But no one cared… except for computer scientists and Huang.
When Nvidia first released the CUDA in beta in 2006, it made waves in computer science circles but was hardly a conversation starter at cocktail parties, said Arthur Carvalho, a Miami University professor.
"It was revolutionary. It was this breakthrough moment — but on the technical side,” he said via phone.
Wall Street didn’t get it, but Huang bet the farm on CUDA anyway. SEC filings show that Nvidia invested nearly $12 billion in R&D from 2006 to 2017 while reeling in revenues in the single billions. In 2014, for example, Nvidia spent over 30% of its $4.13 billion revenue on research and development. Much of the investments during that period went towards CUDA.
Shareholders and analysts complained and scoffed, and many backed away from the company entirely.
“Some investors were big Nvidia fans in the late 2000s and gave them the benefit of the doubt for the first five years of the CUDA investment," said Gilbert. "But in the mid-2010s, market demand still wasn't showing up in a big way, and it was becoming a bigger and bigger investment."
Bryan Catanzaro, Nvidia's vice president of applied deep learning research, has been with the company since 2008. His sense of that time was that Nvidia and Huang were the underdogs, that Wall Street and company watchers fully expected CUDA to be a company-changing — translation: ending — failure.
"It made all of our products more expensive since we were selling these gamer cards while putting computing acceleration into them," he told Yahoo Finance on a video call. "It took a lot of commitment to follow through. … I would say it was about 10 years before Wall Street really started to believe this investment was worth anything."
Today, that company-changing "failure," CUDA, is what drove Nvidia to a trillion-dollar valuation — and made them the go-to for AI companies.
An investor’s eye view
Like Taylor Swift, Nvidia's had distinct eras — but a number of long-term investors have stuck it out, waiting for a moment like this.
Greg Halter, Carnegie Investment Counsel's research director, first bought Nvidia stock on the firm's behalf in 2011. The firm increased its position substantially in 2018 and, today, Nvidia is among Carnegie's 10 largest holdings.
But Halter remembers tougher times. For example, in 2018 and 2022, Nvidia found itself with a "crypto hangover," as sliding cryptocurrency prices led to stark declines in demand for Nvidia's chips, commonly used for crypto mining.
"There are times where a stock price will move dramatically, as opposed to the actual business, actual revenues, and actual cash flow," said Halter. "That’s something you have to tune out if you have a good company.”
Nvidia, throughout its 30-year history, has also weathered life-threatening lawsuits from competitors, questionable acquisitions, and a sputtering push into mobile. A chief criticism of Nvidia is that the company trend-chases, an idea that Halter and others push back on. They produce processing power, and processing power takes the shape of the flavor of the month.
But AI was their end game: The company delivered the very first GPU to a fledgling OpenAI in 2016, through a long-standing relationship with its co-founder Ilya Sutskever — one of the winners of that 2012 software competition.
And now, their believers are rewarded for their patience. Nvidia’s tech and competitive moat in AI are real and the byproduct of risky, long-term investment, said Wealth Consulting Group CIO Jim Worden, who has been investing in Nvidia for a decade.
"Nvidia has got the lead time in terms of competition, and I don't see AMD or Intel coming close to where they're at," said Worden over the phone. "It's a race, it's a sprint, and if someone's a few laps ahead of you in that sprint, it's going to be hard to catch up."
'When an airplane is falling out of the sky, every foot matters'
Despite its appetite for dangling from the precipice, Nvidia has maintained workplace stability for its crew of 26,000. There were some layoffs throughout its history — for example, in 2008 — but mass firings are few and far between.
Huang has retained much of his staff over the years. Nvidia confirmed to Yahoo Finance that, of staff directly reporting to Huang today, seven have been at Nvidia for more than 20 years, four have been with Huang for more than 15 years, and three for more than 10 years. Nvidia’s institutional memory might just be its greatest asset.
"Jensen often says that when an airplane is falling out of the sky, every foot matters," said Catanzaro.
The lack of turnover at Nvidia is unusual and unexpected — both because it’s a tech company and because of Huang’s willingness to go to the edge. Or perhaps, it's that kind of in-the-weeds conviction among Huang and his team that has gotten Nvidia this far.
“I think he has conviction to bet so boldly because he does understand the technologies, the markets, and the use cases soup to nuts,” said Gilbert.
And Huang doesn’t hedge. He doesn’t make multiple bets. The employees at Nvidia are so used to living on the brink of disaster that in the face of overwhelming success, their behaviors and philosophies aren’t changing.
“Working at Nvidia feels existential every day,” said Catanzaro. “It doesn’t matter what the market cap is. We feel like the company is going to go bankrupt tomorrow. We all feel that way.”
The existential dread isn't entirely misplaced in a fast-evolving sector like tech. Nvidia, for all its preparations, isn't yet meeting the generational demand shock for its chips, leaving customers scrambling for any and all GPUs they can find.
That gap — between the need for chips now and Nvidia's ability to provide them en masse — could make way for a future competitor to grab a piece of its pie. Analysts have previously warned Yahoo Finance that a surge in competition for Nvidia is inevitable. There are also concerns about an AI bubble from banks like Morgan Stanley and investors like Alan Patricof. Nvidia may have gained the upper hand, but the game is just beginning.
Then, there's the final risk. Anyone willing to bet big over and over will someday lose — it's not a question of if, but of "by how much?" and when.
Today, with a skyrocketing share price, Huang could cash in, walk away from the table, and make a killing. But he’s won another hand, and he’s going to play it.
“What is the definition of success? A good share price is important, of course,” Huang told those college students 10 years ago. “However, are you doing work that is relevant? Are you doing work that is exquisite? Or are you doing work that results in a great stock price? I have not figured out how to do all three at one time. However, in the long term, it all seems to work out.”
Allie Garfinkle is a Senior Tech Reporter at Yahoo Finance. Follow her on Twitter at @agarfinks and on LinkedIn.
Click here for the latest technology news that will impact the stock market.
Read the latest financial and business news from Yahoo Finance
