Jensen Huang and Nvidia: From a Denny’s Booth to a $5 Trillion AI Giant

Nvidia was sketched out in late 1992 in a Denny’s on Berryessa Road in East San Jose by Jensen Huang, a director at LSI Logic who had designed microprocessors at AMD, and two frustrated Sun Microsystems engineers, Chris Malachowsky and Curtis Priem. Their spouses wanted all three to commit, so Priem resigned on New Year’s Eve 1992 and Huang quit on February 17, 1993, his 30th birthday. They started in Priem’s Fremont townhouse with $40,000, rejected names like Primal Graphics, found NVision taken by a toilet paper maker, and settled on a name from the Latin word for envy. With $20 million in venture capital they faced about 70 rival graphics startups. Their first chip, the NV1, rendered quadrilaterals just as Microsoft’s DirectX standardized on triangles, and it flopped.

Sega dropped Nvidia’s Dreamcast chip, but Sega’s president Shoichiro Irimajiri persuaded his company to invest $5 million in Nvidia anyway, a stake later sold for $15 million. Huang cut staff from 100 to 40, and the triangle-based RIVA 128 shipped in August 1997 with one month of payroll left, selling a million units in four months. Huang then spent over $1 billion building CUDA, and in 2009 the Google Brain team led by Andrew Ng found it sped deep learning about 100 times. Nvidia now holds more than 80 percent of the AI GPU market, passed $2 trillion in 180 days in early 2024, and in 2025 became the first company worth $4 trillion and then $5 trillion. In January 2025 the DeepSeek release triggered a loss of nearly $600 billion in one day, the largest single-day loss for a U.S. company.

  • In September 2023, according to Larry Ellison, he and Elon Musk spent an hour at Nobu in Palo Alto asking Huang for chips.
  • For years Huang opened presentations by telling staff the company was 30 days from going out of business.
  • Nvidia chips supply over 75 percent of the world’s top 500 supercomputers.
  • Huang has about 60 direct reports and refuses to hold one-on-one meetings.
  • A Cisco executive noted that Nvidia chips are delivered to data centers in armored cars.

Leave a Reply

Discover more from pplpod

Subscribe now to keep reading and get access to the full archive.

Continue reading