In 1995, NVIDIA was a month from bankruptcy. This conversation saved it.
In 1995, NVIDIA was about a month from bankruptcy. Not because of bad luck, because its own founders had made three confident technical bets, and all three were wrong, while roughly a hundred competing startups in Silicon Valley got it right. What saved the company wasn’t a better idea. It was radical honesty in leadership, delivered in person, at real personal risk, to someone who had every reason to walk away.
Three Confident Bets. All Three Wrong.
NVIDIA’s first chip, the NV1, made three major architecture decisions, and it got all three backwards relative to where the industry was headed. It used forward texture mapping instead of inverse. It rendered curved surfaces instead of triangles. It had no Z-buffer, meaning the application itself had to sort depth instead of the hardware doing it automatically. When the industry standardized on triangle-based rendering shortly after, NVIDIA’s approach became incompatible with where every other serious competitor was going.
This wasn’t a case of one bad call. It was three well-reasoned technical decisions, made by smart engineers, all wrong at the same time, while a hundred other companies chasing the exact same opportunity chose correctly. NVIDIA had been first to the starting line and finished dead last.
The Conversation Nobody Wanted to Have
By the time the company recognized the mistake, it was in serious trouble, down to roughly a month of cash, having already laid off more than half its employees. NVIDIA also had an outstanding contract with Sega: NVIDIA was supposed to deliver a chip for a Sega game console, and Sega still owed a final payment on that contract.
Jensen Huang, then in his early thirties, flew to Japan and sat down with Sega’s president, Shoichiro Irimajiri, a 20-year Honda executive, including a stint as president of Honda of America, who had joined Sega a couple of years earlier. Huang told him three things, directly: the technology NVIDIA had promised didn’t work. NVIDIA should be released from finishing the contract, because completing it would waste Sega’s money on something broken. And, even though he was asking to be let out of the contract, NVIDIA still needed Sega to hand over the final payment, or the company would collapse immediately.
Huang’s own assessment of the odds, in his words: “What are the chances that giving NVIDIA $5 million, that we would develop the right strategy, that he would get a return on that $5 million or even get it back? 0%. You do the math, it’s 0%.” That flight to Japan was radical honesty in leadership in its earliest, least polished form, years before either man had language for it.
The Answer Nobody Expected
Irimajiri took a few days to think it over. Then he said yes anyway. That yes is best understood as a return on radical honesty in leadership, not a return on a five-million-dollar bet, the math never justified the money.
That $5 million bought NVIDIA roughly six more months, just enough time to build the RIVA 128, a chip built the right way this time, which shipped in 1997 and sold a million units in four months. It saved the company. Sega eventually sold its NVIDIA shares for about $15 million, tripling its money. Had it held on, that stake would be worth something in the neighborhood of a trillion dollars today.
Radical Honesty in Leadership Actually Saved the Company
It’s tempting to file this away as a story about a lucky investor. That misses the actual lesson. Sega didn’t say yes because the pitch was compelling, by Huang’s own math, the odds were zero. Sega said yes because Huang walked in and told the truth about a failure that was entirely his own company’s fault, instead of spinning it, delaying it, or letting Sega find out on its own.
That’s the whole mechanism. Not a better technical idea in the moment, an admission that the idea had been wrong, delivered with enough honesty and humility that the person on the other side of the table decided to bet on the person instead of the technology.
Radical Honesty in Leadership: The Lesson Underneath the Story
Being confident and being right are not the same thing, and a room full of smart, well-credentialed people can be confidently wrong together, NVIDIA’s engineers were. The instinct in that moment is almost always to protect the story: delay the bad news, soften it, hope the technology can be patched before anyone notices. Huang did the opposite. He walked into the room with the least favorable version of the truth and let the other person decide what to do with it.
That’s radical honesty in leadership in its most literal form, not a communication style, but a decision made under real pressure, with real consequences, by someone who had far more to lose by being honest than by staying quiet a little longer. It’s also a case study in the accountability, resilience, agility, and perseverance that separate the organizations that survive a wrong bet from the ones that don’t.
Where the 5 Kryptonite Ingredients Come From
The five ingredients referenced below, IDEAS, SPEED, TALENT, DISTINCTION, and LEADERSHIP AT ALL LEVELS, are explained in full in Two Brothers, Zero Employees, $401 Million.
Run it through the Kryptonite ingredients and two carry the weight: IDEAS, the willingness to recognize, quickly and completely, that the original idea was wrong, rather than defend it past the point of usefulness, and LEADERSHIP AT ALL LEVELS, the willingness to have that conversation in person, without deflecting responsibility, while there was still something left to save.
IDEAS and LEADERSHIP AT ALL LEVELS is obvious, this story also reveals the other three. SPEED: Huang didn’t call a meeting or run another round of analysis — with a month of cash left, he got on a plane. TALENT: the conversation itself was a masterclass in the human skills no algorithm can fake — honesty, humility, and the integrity to argue against his own company’s short-term interest in the room. DISTINCTION: Irimajiri wasn’t betting on a chip, he was betting on NVIDIA — and that bet only makes sense if the company had already earned enough credibility and differentiation for a 20-year Honda executive to say yes at 0% odds.
Radical honesty in leadership doesn’t start as a company value on a wall. It starts as one uncomfortable conversation someone chooses to have before they’re forced to.
What’s the conversation in your organization right now that everyone’s avoiding because the comfortable story is easier than the honest one?
This is exactly the kind of shift I walk through in Distinct or Extinct, where the real threats and opportunities of disruption actually sit, and the 90-Day Deployment Plan for responding to them before your competitors do. If you want a clearer read on where your organization stands, the Kryptonite Scorecard will show you: realmikeevans.com/scorecard