Huang Bet the Company on a Chip He Never Touched

July 13, 2026 - 21:34

In 1997, Nvidia was running out of money. The company had burned through most of its venture capital, and its first product had flopped. Co-founder and CEO Jensen Huang faced a brutal choice: either bet everything on a new graphics chip that existed only inside a computer simulation, or watch his startup die.

Huang ordered mass production of the RIVA 128. The problem was that no physical prototype had ever been built. Every test, every performance benchmark, every compatibility check had been run inside a software simulator. If the simulation was wrong, the chips would be useless, and Nvidia would be bankrupt.

"I never held the chip in my hands before we committed to manufacturing," Huang later recalled. The gamble was enormous. The company had just enough cash left to pay for the first production run. There was no room for error.

The RIVA 128 shipped and became a hit. It outperformed competitors from 3dfx and ATI at a lower price point. The chip saved Nvidia and set the stage for the GeForce line that would dominate PC gaming for decades.

Huang's willingness to bet the company on a simulation reflected his engineering background. He trusted the math. But it also reflected desperation. "We would have been out of business if it didn't work," he said. The chip that saved Nvidia was born not in a clean room, but inside a computer model.