The founding of NVIDIA
Three engineers, one Denny's booth, and a bet that graphics would eat computing.
The setup
By early 1993, Jensen Huang had spent nine years in the chip industry — at LSI Logic, where he ran the CoreWare business, and at AMD as a microprocessor designer — plus a fresh Stanford master’s degree earned at night. He was 30, established, and employable for life.
He quit to start a company with two fellow graphics engineers: Chris Malachowsky, a Sun Microsystems veteran, and Curtis Priem, formerly of IBM and Sun, one of the designers of the GX graphics chip. All three believed the same unfashionable thing: that 3D graphics was not a sideshow but the future of computing, and that a dedicated graphics processor — built by a company that did nothing else — could outrun the integrated offerings of the giants.
The Denny’s meeting
The founding conversation happened at a Denny’s restaurant in East San Jose in April 1993. The choice of venue was no accident in the retelling: Huang had washed dishes and waited tables at Denny’s as a teenager in Oregon. Starting his company in the same booths where he’d worked graveyard shifts became the founding myth he still tells — the dishwasher returning as a founder.
Whether a business plan was literally sketched on a napkin is the kind of detail that improves with retelling (see The myth machine). What is documented: the three men agreed on the company, the mission, and the name that month.
The name
“NVIDIA” combines NV — engineering shorthand for “next version” — with invidia, the Latin word for envy. Files from the company’s early days show the founders cycling through alternatives before landing on it.
The money
The founders put in their own capital — Huang has described the starting stake in interviews — and then raised venture funding. Sequoia Capital and Sutter Hill Ventures became early backers, with Sequoia leading the institutional round. The exact early figures are reported variously in profiles; what matters is the structure: a classic Silicon Valley venture bet on three technical founders with deep domain expertise.
The first office
NVIDIA set up in Sunnyvale, California, in the heart of Silicon Valley — close to the chip ecosystem Huang knew from LSI and AMD, and close to the manufacturing partner that would matter most: TSMC in Taiwan, the foundry that would fabricate NVIDIA’s chips for the next three decades.
Why it almost didn’t work
The founding bet had a flaw the founders couldn’t see: they bet on the wrong graphics standard. Their first chip, the NV1 (1995), rendered 3D with quadratic texture mapping — elegant, but incompatible with the triangle-based approach Microsoft’s DirectX was standardizing. Within two years the company was nearly dead (see Near-death).
Huang’s retrospective lesson, repeated in speeches: the founding vision was right (graphics would matter enormously) and the founding execution was wrong (they built for the wrong standard). Surviving the gap between the two nearly killed them.
The co-founders
- Chris Malachowsky — Sun Microsystems engineer; NVIDIA’s longtime technologist, still with the company decades later.
- Curtis Priem — graphics architect (Sun, IBM); left NVIDIA in the early 2000s and became a noted philanthropist, including major gifts to Rensselaer Polytechnic Institute.
Huang is the only founder who stayed as CEO — one of the longest continuous founder-CEO runs in Silicon Valley history.
Source notes
Founding details draw on Huang’s retellings (including the Sequoia Capital oral history and GTC keynotes), NVIDIA’s corporate history, and profiles of Malachowsky and Priem. The Denny’s location (East San Jose, April 1993) is consistently reported; napkin-level details are flagged as embellishment-prone on The myth machine.