Podcast Review | NVIDIA: Jensen Huang – How I Built This with Guy Raz

NVIDIA: Jensen Huang. From near collapse to becoming the world’s biggest company – How I Built This with Guy Raz

I spent an hour or so over the weekend listening to the latest How I Built This episode with NVIDIA CEO Jensen Huang. For anyone with an interest in NVIDIA's history, or what it actually takes to build a generational technology company, it is well worth the time.

What stood out to me was just how close NVIDIA came to going out of business in the mid-1990s. At one point the company was circa 30 days from insolvency with approximately US$30,000 in the bank. The cause was the NV1 chip, which used 'forward texture mapping' while the rest of the industry, including Microsoft's DirectX in Windows 95, standardised on 'inverse texture mapping'. The lifeline came from a trip to Japan, where Huang persuaded Sega to convert ~US$5 million of an existing US$12 million contract into an equity investment, buying NVIDIA the runway to keep operating.

The CUDA section was the part I found most relevant. NVIDIA launched CUDA in 2006 and made the deliberate choice to embed it into every GeForce gaming chip, seeding millions of devices with general-purpose GPU compute years before there was any commercial AI workload to justify it. Huang frames it as solving the 'chicken or the egg' problem for new platforms — you cannot wait for the application before laying down the infrastructure. It is hard to think of many platform decisions of the last twenty years that have aged better.

From a leadership and resilience perspective, a few things stood out. The first is Huang's willingness to admit, publicly and to a major customer, that the product had failed and that the company needed help to survive. That conversation with Sega was an act of commercial vulnerability very few founders would be comfortable with, and it likely saved the business. The second is the discipline of staying invested in CUDA for nearly a decade without any obvious payoff. Huang did not appear to waver despite years of NVIDIA being labelled a misallocator of capital — a useful reminder that the market's tolerance for long-duration investments is often far shorter than the time it takes for those investments to compound. The third is his attribution of much of his temperament — self-reliance, patience, the ability to sit with discomfort — to the years he and his brother spent at the Oneida Baptist Institute in Kentucky, where they had been enrolled by mistake.

Huang also points to The Innovator's Dilemma as a foundational influence — the idea that disruptive technologies often look 'toyish' before becoming 'good enough' to displace incumbents. A useful frame to keep in mind.

‍

‍Link to Podcast

View more NAOS podcasts and book reviews‍