“250,000 units went out. 250,000, well, 249,000 came back. And, practically put us out of business.” That is Jensen Huang, on Sequoia Capital’s Crucible Moments podcast, describing what happened to Nvidia’s first product. Today the company is valued at $4.94 trillion.
A diner booth and a doomed chip
In 1993, Huang, Chris Malachowsky and Curtis Priem met at a Denny’s to talk about a chip that would bring realistic 3D graphics to personal computers. Huang places the booth “right at the corner of Capitol and Berryessa, this place in East San Jose.” His first day of work was February 17th, his birthday, when he had just turned 30.
The NV1 was the result. Under the hood, it processed quadrilateral primitives (forward texture mapping), while competitors preferred triangles. Then Microsoft introduced DirectX and announced that Direct3D would support only triangles. The technical reality was simple: a chip built around a rival geometry had no future on the PC platform that mattered. The NV1 failed to gain traction, and back it came.
Thirty days from the end
Huang has turned that close call into a habit of mind. He told Joe Rogan in December 2025 that he has used the phrase “30 days from going out of business” for 33 years. Former Sequoia partner Mark Stevens recalled Huang saying exactly that, and added that on a few occasions in the mid-nineties it was true.
Sega kept the lights on. Huang hedged his memory: “it was some 5 million dollars, I think it was, that they continued to pay us. It was all the money that we had.” The Next Web describes it as a $5m investment in 1995 that bought roughly six more months. That is not much runway for a hardware company.
RIVA 128 and what came after
According to The Next Web, Nvidia spent that time on the RIVA 128, which shipped in 1997 and sold one million units in four months. “Yeah, we saved the company,” Huang said. Nvidia went public on January 22, 1999. Sega later sold its Nvidia stock for $15 million, which, as ever, looks like a modest sum in hindsight.
From there came CUDA in 2006, the AlexNet deep learning success on Nvidia GPUs in 2012, several crypto cycles and the present AI buildout. Market commentator Josh Brown noted that “CUDA far predates AI, but CUDA is the thing that every professional in the field has standardized on.”
The numbers now
As of Oct. 30, 2026, Nvidia stood at $4.94 trillion, behind Apple at $5.03 trillion, a gap of $88 billion. Its latest quarter brought revenue of $96.22 billion, up 105.85% year over year, with Data Center contributing $89.02 billion. The company that once had about a month of cash now reports supply obligations of $279.00 billion, mostly memory for Vera Rubin.
Success has not cured the anxiety. “The sense of vulnerability, the sense of uncertainty, the sense of insecurity, it doesn’t leave you,” Huang told Fortune. Next on the calendar is the fiscal third-quarter revenue guidance of $108.0 billion, plus or minus 2%, which leaves out any Data Center compute revenue from China.
Sources
- In 1993, 3 Friends Met At Denny's To Talk About Making A Chip That Would Allow Computers To Create Realistic 3D Graphics. By The Mid-1990s, The Company Was Weeks From Running Out Of Cash. Today Nvidia Is Worth Nearly $5 Trillion
- Nvidia
- Not AMD. Not Marvell. Here's Which Company Nvidia Needs to Be Worried About.
- Not AMD. Not Marvell. Here's Which Company Nvidia Needs to Be Worried About. | The Motley Fool
- NVIDIA RTX Spark Powers Neural Rendering and AI Characters in Games | NVIDIA Technical Blog
- Nvidia forced to dial back key bandwidth targets on Vera Rubin hardware — tweaks will affect target specifications for memory and interconnect fabric
Photo: Anderseidesvik (CC BY-SA 4.0)
