That student was Jensen Huang.

In a recent podcast appearance with Joe Rogan, the Nvidia CEO traced that improbable starting point back to his parents, who had sent him and his brother to the United States in the mid-1970s with almost nothing. The family had been living in Bangkok during one of Thailand’s periodic coups, and his parents decided it was no longer safe to keep the children there. They contacted an uncle they had never visited in Tacoma, Wash., and asked him to find a school in America that would accept two foreign boys with almost no savings.

He found one: Oneida Baptist Institute in Clay County, Ken., one of the poorest counties in the country then and now. The dorms had no closet doors, no locks, and a population of kids who smoked constantly–Huang said he also tried smoking for a week, at 9 —and settled disputes with knives. Huang’s roommate was a 17-year-old wrapped in tape from a recent fight; the “toughest kid in school,” he said. Every student had a job. His brother, was sent to the tobacco fields the school ran to fund the school—“kind of like a penitentiary”—while Huang became the janitor, cleaning the bathrooms for a hundred teenaged boys (“I just wished they would be a bit more careful” in the bathroom, he joked.)

That indefatigable cheerfulness, even when describing scenes that sound brutal to almost anyone else, ran through the entire interview. 

Huang said most of his memories from that period were good, and remembers the time he told his parents his amazement after eating at a restaurant: “Mom and dad, we went to the most amazing restaurant today. This whole place is lit up. It’s like the future. And the food comes in a box and the food is incredible. The hamburger is incredible.”

“It was McDonald’s,” Huang laughed. 

Indeed, these memories were relayed to his parents late; the boys were navigating all of this alone. International phone calls were too expensive, so his parents bought them a cheap tape deck. Once a month, they recorded an audio letter describing their lives in coal country and mailed it back to Bangkok. Their parents taped over the same cassette and mailed it back.

Two years later, Huang’s parents finally made it to America, with just suitcases and only a bit of money. His mother worked as a maid. His father, a trained engineer, looked for work by circling openings in the newspaper classifieds and calling whoever picked up. He eventually found a job at a consulting engineering firm designing factories and refineries.

“They left everything behind,” Huang said. “They started over in their late thirties.”

He still carries one memory from those early years that he said “breaks my heart.” Not long after his parents arrived in the U.S., the family was living in a rented, furnished apartment when he and his brother accidentally broke a flimsy particle-board coffee table. 

“I just still remember the look on my mom’s face,” he said. “They didn’t have any money, and she didn’t know how she was going to pay it back.”

For Huang, moments like that define the stakes his parents accepted when they came to the U.S. “with almost no money”.

“My parents are incredible,” he said. “It’s hard not to love this country. It’s hard not to be romantic about this country.”

Jensen Huang’s humble beginnings inspired Nvidia principles

That way of seeing America—as a place where people will give you a chance if you’re willing to take one—is how Huang explains Nvidia’s early, unlikely bets. 

Huang came up with the idea for Nvidia while sitting in a booth at a Denny’s, where he had worked first as a dishwasher and then a busboy. He wanted to build a chip that could power 3D graphics on a personal computer, and it was at that Denny’s booth that he met two friends to sketch out what would become the company.

Long before the company became synonymous with the AI boom, Huang kept steering it toward ideas that few people understood and even fewer believed in. CUDA was one of them. When Nvidia introduced it in 2006, the cost of the chip roughly doubled, revenue did not move, and the company’s valuation fell from about $12 billion to between $2 and $3 billion.

“When I launched CUDA, the audience was complete silence,” he said. “Nobody wanted it. Nobody asked for it. Nobody understood it.”

CUDA is the software layer that turns the graphics chips into general purpose compute engines, making them capable of large neural networks. Now, of course, nearly every major AI model today runs on hardware that depends on CUDA. 

The same thing happened when he introduced Nvidia’s first AI supercomputer, the DGX1. The launch drew “complete silence,” he said, and there were no purchase orders. The only person who reached out was none other than Tesla CEO Elon Musk, who told him he had “a nonprofit AI lab” that needed a system like this.

Huang assumed that meant the deal was impossible.

“All the blood drained out of my face,” he told Rogan. “A nonprofit is not buying a $300,000 computer.”

But Musk, the world’s richest man, insisted. So Huang boxed up one of the first units, loaded it into his car, and drove it to San Francisco himself.

In 2016, he walked into a small upstairs room filled with researchers— Berkeley robotics pioneer Pieter Abbeel, OpenAI cofounder Ilya Sutskever, and others—working in a cramped little office. That room turned out to be OpenAI, long before it became the most discussed AI organization in the world. Huang left the DGX1 with them and drove home.

Looking back, even as the CEO of a $4.5 trillion company who now draws crowds and autograph-seekers wherever he goes, he doesn’t describe any of this as foresight or heroism. To him, it’s simply the continuation of the risks his parents took when they sent two boys across the world with almost nothing.

“We really believed it, and so if you believe in that future, and you don’t do anything about it you’re going to regret it for your life,” Huang said.

“The people who first made Nintendo’s hits are still working at the company,” Keza MacDonald, the author of Super Nintendo, a forthcoming book about the developer, told me recently. “For the last 50 years, these people have been passing down knowledge and training up a new generation of Nintendo creatives.” 

Both Nintendo’s business and creative leaders have long tenures at the company. Current president Shuntaro Furakawa joined the company in 1994 as an accountant. Shigeru Miyamoto, the brains behind franchises like “Super Mario” and “The Legend of Zelda,” joined as a staff artist in 1977. 

There is a risk that companies that rely too much on institutional knowledge get stuck in their ways. Yet Nintendo, according to MacDonald, has combined institutional knowledge with fresh ideas to continuously replenish its pipeline of fun games: “It’s not like the oldest guy gets to decide what’s a good idea and what isn’t. Everyone puts ideas in.”

Nintendo has its share of flops, failed experiments, and puzzling business decisions–as does every firm. Yet the company maintains its share of the highly competitive video game industry against bigger, deeper-pocketed rivals like Sony and Microsoft. 

The few designers who’ve left Nintendo still have fond feelings about their time there. As Lee Schuneman, a former Nintendo game designer and now Efekta Education Group’s chief product officer, told our Brainstorm Design audience this week, “I got to work with some of the most talented game designers in the world, including people like [Shigeru Miyamoto] at Nintendo, and [learn] a whole range of lessons about how to make playful experiences.”

That goodwill may be the result of Nintendo avoiding the industry’s boom-bust churn and valuing the expertise its workforce accumulates.

Nintendo “is still, to this day, making games differently from everyone else,” MacDonald says. You can check out the rest of our mainstage sessions from Brainstorm Design here.—Nicholas Gordon

Contact CEO Daily via Diane Brady at diane.brady@fortune.com

Top news

Netflix to acquire Warner Bros. Discovery studios 

The online streamer and the maker of the  Superman and Harry Potter franchises are expected to announce a sale of Warner’s studios and HBO Max business to Netflix, the WSJ reports. Paramount Skydance chief David Ellison lobbied the White House against the deal even though Netflix offered a richer valuation, according to the New York Post.

“China’s Nvidia” stages IPO

Moore Threads, a maker of GPUs based in Beijing went public today at a valuation of $1.1 billion and its stock rose by 400% on day one.

$10 billion a week on U.S. national debt

The calendar year may have a few weeks left to tick off, but as far as the government’s budget is concerned, we’re in fiscal 2026. The Treasury has already paid out a 12-figure sum to service the nation’s debt. Unlike the tax and calendar year, the government’s financial calendar runs to the end of September. According to Treasury data, in the nine weeks since, it has spent $104 billion in interest on its $38 trillion borrowing burden. That’s more than $11 billion a week, and already represents 15% of federal spending in the current fiscal year.

Poor labor data may have locked in Fed cut

Analysts may not have necessarily digested this week’s lackluster labor data with glee—but it sure didn’t dampen their spirits either. Wall Street is hoping for a Christmas miracle with a final interest rate cut from the Fed, bringing the base rate down to 3.5% to 3.75%, and recent jobs reports may just have sealed the deal.

U.S. lobbied against E.U. seizing Russian money

American officials urged Europe not to use frozen Russian assets as the basis of loans that would fund Ukraine’s defense against Moscow’s invasion of its Eastern flank. The funds could be used as an incentive to end the war, Washington argued.

January 6 pipe bomb suspect arrested

Brian Cole Jr., 30 of Woodbridge, Virginia, was the subject of a five-year-long investigation by federal officials. 

Wall Street forecasts S&P will hit 7,500

Analysts are publishing their notoriously unreliable annual stock market forecasts and this year nine investment banks are guessing that the market will rise about 10% in 2026.

The markets

S&P 500 futures were up 0.17%  this morning. The last session closed up 0.11%. STOXX Europe 600 was up 0.18% in early trading. The U.K.’s FTSE 100 was up 0.19% in early trading. Japan’s Nikkei 225 was down 1.05%. China’s CSI 300 was up 0.84%. The South Korea KOSPI was up 1.78%. India’s NIFTY 50 is up 0.55%. Bitcoin fell to $91.4K.

Around the watercooler

How a Texas gas producer plans to exploit the ‘mega trend’ of power plants for AI hyperscalers by Jordan Blum.

Battle for sports betting market heats up as Polymarket announces return to the U.S. by Carlos Garcia.

Nvidia CEO Jensen Huang admits he works 7 days a week, including holidays, in a constant ‘state of anxiety’ out of fear of going bankrupt by Jessica Coacci.

Kim Kardashian shaped Skims into a $5 billion brand—now she wants to help other entrepreneurs mold their skills for success by Emma Hinchliffe.

CEO Daily was compiled and edited by Jim Edwards and Lee Clifford.

“I went to a very large institutional investor and explained our gas-to-power strategy in our business, and I got berated for like 30 minutes about how it was such a foolish thing for me to go into power,” BKV founder and CEO Chris Kalnin told Fortune.

Fast forward to today, and BKV’s stock has spiked 50% since going public in September 2024—rising from a small cap to a mid-market cap value of $2.5 billion. BKV is on the brink of making a deal with a hyperscaler to provide immediate gas-fired power to an AI data center campus and continue, according to analysts, before buying and building more power plants.

“It was a pretty controversial decision for us to buy power. It’s been, honestly, one of our best investments ever,” Kalnin said. “Hyperscalers need more generation. They used to talk about hundreds of megawatts. Now the conversations start with gigawatts. Can you give me gigawatts?”

“We’re going to have to build power plants. If they want gigawatt power, we’re going to have to add more power,” he said.

BKV was founded a decade ago in partnership with Thailand’s Banpu Power—BKV being short for Banpu Kalnin Ventures—so there was some built-in power expertise. After focusing on natural gas production, four years ago BKV started buying up two power plants in Temple, Texas—located between Austin and Dallas—which now provide a total of 1.5 gigawatts of electricity generation capacity—enough to power more than 1.1 million homes, or a major data center campus. There is room to expand.

BKV is in the process of increasing its ownership stake in the joint venture with Banpu from 50% to 75%—slated to close in the first quarter—to double down on the power business and to better disclose the financials to investors, now that BKV is public, Kalnin said.

Tim Rezvan, energy analyst for KeyBanc Capital Markets, said the new AI data center market is naturally finding its way to BKV and its power plants, as opposed to BKV chasing a boom from behind.

“It’s a lot of skill and a little bit of luck on top of that to take that power plant when they did,” Rezvan said. “They’re really in the catbird seat because they control these merchant power plants that can redirect power, in theory, the next day. The market is eagerly waiting to hear what’s going to happen with a potential, behind-the-meter deal with a hyperscaler.”

Despite BKV continuing to grow as a gas producer, Rezvan said the power segment is now a majority of the value of the stock and 90% of what investors want to talk about.

BKV also has one of the most advanced carbon capture and storage programs in the energy sector to deliver the power more cleanly, which may be especially inviting to Big Tech.

“The ability to deliver almost an carbon-neutral natural gas molecule—and then they can tie that in with the big hyperscaler—is a unique suite of services they can offer carbon-conscious consumers of power,” Rezvan said.

Birthplace of shale

An alum of McKinsey, Kalnin first connected to Banpu through that network. Banpu wanted to invest in U.S. shale gas after seeing the U.S. ship cheap natural gas to Asia in the early days of the shale boom.

The industry’s modern drilling and hydraulic fracturing, or fracking, techniques were pioneered in the Barnett Shale near Dallas, but those companies quickly fled the more mature Barnett to Louisiana’s Haynesville Shale, Pennsylvania’s Marcellus Shale and, eventually, West Texas’ oily Permian Basin.

As everyone was moving out of the Barnett, BKV bought in cheaply and, over time, became the dominant player there. “The Barnett was heavily undervalued relative to its risk,” Kalnin said. “The only way you find deep value is you see something that other people don’t see in the fundamental value. And so that’s the origin story of BKV.”

“Basically, all the players left the Barnett in that 2010 timeframe and went to other plays. Those plays have evolved how you fracked, and how you drill, and how you do directional drilling, and how you target [gas] zones. None of that technology was reapplied back to the Barnett,” he said. Now, BKV is taking those advanced drilling and fracking techniques and getting much more value from the Barnett than was believed possible, he said.

A lot of gas goes toward power generation, so Kalnin saw that as a natural extension. Borrowing on his McKinsey background, Kalnin said he sought to identify “mega trends” and take advantage. “It’s the idea of a glacier moving in a direction, and you can’t stop it. The idea is you want to get into that path or get the benefit from that trend. You’re not going to figure out all the nuances, but you’re going to get the direction correct if you think through it deeply.”

That mega trend Kalnin identified was the anticipated growth of U.S. power demand and the underinvestment in the sector after more than two decades of flat demand.

“Natural gas is the baseload of U.S. power, and not just as a bridging fuel, but actually as a core fuel for the future,” Kalnin said. “Most of the research at the time was showing that it was going to be all renewables. I called absolutely BS on that, and I said I’m going to double down on gas.”

Kalnin admittedly did not foresee the AI data center boom, but he did see the need for more power from population growth and greater electrification and manufacturing.

“We saw the trend,” he said. “And then, of course, the AI train has kicked things into high gear.”

What’s next?

BKV just expanded in the Barnett through a $370 million acquisition from Bedrock Energy Partners, and the company is launching another carbon capture project next year as part of its “Barnett Zero” emissions effort.

But the real focus is on power and attracting a hyperscaler. The pitch is for “closed-loop, net-zero power” from the gas fields and pipelines to the power plants and carbon capture.

“Not only do we have the power side, we can do the pipelines, we can do the gas, we can do the grid connection, we can do the whole thing soup-to-nuts in one company, and, by the way, we can decarbonize it with carbon capture,” Kalnin said.

“I can give somebody a fixed power price for 20 years because we can produce and sell the gas to ourselves and fix the gas price. Think about the ability to do that with a hyperscaler.”

Rezvan agrees. But, if BKV hasn’t signed a major power deal in the next six months or so, investors may start to get antsy.

“There will be some pressure to deliver,” Rezvan said. “The question is, why would a mega-cap tech company partner with a small energy company? The answer is because they have power right now. It’s not a greenfield project that would take many years. They could literally turn around and deliver that power in short order.

“This is a power-starved market. The ability to deliver on short notice is what gets them a deal I believe.”