Nvidia's early technical bet on quadrilateral polygons and underestimating Moore's Law nearly bankrupted the company when Microsoft standardized on triangles.

Nvidia and Sega's NV2 chip used quadrilaterals instead of the triangles Microsoft's DirectX would standardize on
Sega abandoned Nvidia for the Dreamcast in 1996
Nvidia's memory-efficient chip design became a cost disadvantage as memory prices fell faster than anticipated

“So it's interesting to note that NVIDIA by being first and not projecting out the exponential change that would come from Moore's Law was actually…”
40:00

#directx#microsoft#near bankruptcy#moore's law

Myth Buster1:05:30

Intel's Failed Attempt to Build Its Own Graphics Cards

Intel tried releasing dedicated external graphics cards to compete directly with Nvidia around 1999, and the products were poorly reviewed.

Intel's usual strategy was to integrate successful peripherals into the motherboard, commoditizing startups
Nvidia's programmable shaders made this harder to replicate
Intel's own external graphics cards were among the worst-reviewed of their era

“They come out with their own external cards right around this time, like 1999, to directly compete and like they suck.”
1:06:00

#intel#competition#integrated graphics

Hot Take· 1

Hot Take1:26:00

The Almost-Merger: When AMD Tried to Buy Nvidia

Before buying ATI, AMD's first choice was actually to acquire Nvidia — a deal that fell apart over who would run the combined company.

AMD approached Nvidia first, before ultimately acquiring ATI in 2006
Jensen insisted on being CEO of the combined company
That demand reportedly killed the deal, redirecting AMD to ATI instead

“Jensen in this merger acquisition talk with AMD insisted that he be the CEO of the combined company. And that is the thing that blew…”
1:27:00

#amd#ati#m&a#jensen huang

Explainer· 3

Explainer29:30

Why It's Called Nvidia: The Envy Origin

The founders picked the name from file names shortened to 'NV' and landed on the Latin word for envy.

Design files were saved as 'NV' for next version
Founders searched for words containing 'NV' and found the Latin word for envy
Green branding was later chosen to play on 'green with envy'

“they find the Latin word NVIDIA, I-N-V-I-D-I-A, which means envy”
30:00

#branding#naming#company history

Explainer57:30

Why GPUs Are Built for Parallel, Not Serial, Processing

An explanation of the core architectural difference between CPUs (serial) and GPUs (massively parallel) and why it matters for graphics.

CPUs process instructions serially, one after another through the pipeline
GPUs can process many independent pixel/lighting calculations simultaneously
This parallelism is the foundational mechanism that later enables scientific and machine learning workloads

“Like there's all these things that need to get outputted to the screen that do not depend on each other.”
58:30

#gpu architecture#cpu#parallel computing

Explainer1:09:30

The $500M Xbox Deal That Funded Programmable Shaders

Microsoft's secret Xbox project became a major supply deal for Nvidia, funding the GeForce 3 and the introduction of programmable shaders.

Microsoft signed Nvidia to a $500M/year deal with a $200M advance to supply Xbox graphics
The GeForce 3 introduced programmable shaders, enabling dynamic in-game lighting for the first time
This partnership protected Nvidia strategically from Intel's motherboard integration threat

“$500 million a year deal for NVIDIA to supply the graphics for the Xbox with a 200 million dollar advance.”
1:09:30

#xbox#microsoft#geforce 3#programmable shaders

Story· 6

Story07:00

From Taiwan to a Kentucky Reform School: Jensen Huang's Origin

Jensen Huang's childhood journey from Taiwan and Thailand to a boarding school in Kentucky that turned out to be a reform school, and how it shaped his resilience.

Jensen's father worked for Carrier and brought the family's American dream to life
His mother taught him English from a dictionary, 10 words a day
He was sent to Oneida Baptist Institute in Kentucky at age 9, which turned out to be a reform school
His roommate was a 17-year-old recovering from a knife fight; they became close friends and Jensen started weightlifting

“Now I don't get scared very often. I don't worry about going places I haven't gone before. I can tolerate a lot of discomfort.”
Jensen Huang · 09:30

#jensen huang#biography#origin story

Story19:30

The Denny's Pitch That Started Nvidia

Chris Malachowski and Curtis Priem pitch Jensen on a 3D graphics chip idea over dinner at Denny's, leading to Nvidia's founding in 1993.

Jensen's Sun Microsystems colleagues Chris Malachowski and Curtis Priem pitch him on a dedicated 3D graphics chip for PCs
The bet was that consumer PC gaming (Wolfenstein 3D, Doom) would become a massive market
The idea required building an entire developer ecosystem (APIs/SDKs) from scratch

“well, sure, let's go meet at my favorite spot, Denny's”
19:30

#founding story#3d graphics#denny's

Story28:00

Don Valentine's Blind Bet on Jensen

How a referral from LSI Logic's CEO helped Jensen secure funding from Sequoia's Don Valentine despite botching the pitch.

LSI Logic CEO Wilf Corrigan personally referred Jensen to Don Valentine at Sequoia
Jensen badly fumbled the pitch itself
Sequoia and Sutter Hill funded Nvidia at a $6M post-money valuation

“So, against my best judgment, based on what you just told me, I'm gonna give you money. But if you lose my money, I'll kill…”
Don Valentine · 28:30

#venture capital#sequoia#fundraising

Story1:02:00

The Letter That Won Over TSMC's Morris Chang

After sales channels ignored Nvidia, Jensen mailed a physical letter to TSMC's founder, who called him directly and became Nvidia's long-term foundry partner.

Jensen wrote directly to Morris Chang after TSMC's sales team ignored Nvidia
Chang called Jensen personally at Nvidia's office during a chaotic testing crunch
The relationship led to a landmark multi-year manufacturing deal still in place today

“everybody shut up. Morris Chang is on the phone.”
Jensen Huang · 1:03:00

#tsmc#morris chang#supply chain

Story1:23:30

The Apocryphal Story That Foreshadowed GPU Computing

Jensen's likely-composite anecdote about a Stanford researcher who used off-the-shelf GeForce cards to replace weeks of supercomputer time with hours.

A quantum chemistry researcher's models that took weeks on a supercomputer finished in hours using GeForce cards
The story foreshadows Nvidia's future pivot toward scientific computing and machine learning
The hosts note it's likely a composite/apocryphal story but 'true in spirit'

“I just want to thank you, Jensen, for making my life's work achievable in my lifetime.”
1:25:00

#scientific computing#gpu compute#foreshadowing

Story1:45:00

Nvidia's Early Bet on Keyhole, Later Google Earth

Jensen personally saw the potential in Keyhole, a struggling startup building a graphical model of Earth, which Google later acquired to become Google Earth.

Keyhole couldn't raise money from other investors but got backing tied to Nvidia's leadership
The company became Google Earth after acquisition
Signals Nvidia's early recognition that graphics/simulation had value far beyond gaming

“Google acquired it. It became Google Earth. And NVIDIA was one of the early investors.”
1:46:00

#keyhole#google earth#simulation#investing

Tool· 1

Tool47:00

The Million-Dollar Emulator: Testing Chips at One Frame Every 30 Seconds

With nine months of cash left, Nvidia used an unproven software emulator to design and test its next chip without a physical prototype.

Nvidia spent roughly a third of remaining cash on emulator hardware and software
Testers had to watch the emulator render one frame every 30 seconds to verify correctness
The emulator startup Nvidia used went out of business shortly after, having had Nvidia as essentially its only customer

“This emulator runs at one frame every 30 seconds.”
48:00

#chip design#emulation#tapeout

Takeaway· 1

Takeaway1:27:30

From 29% to 66%: Nvidia's Gross Margin Turnaround

A key data point showing Nvidia commanded little pricing power in 2004 versus the differentiated, high-margin business it would later become.

Nvidia's gross margin was only 29% in 2004 despite years of rapid revenue growth
By contrast, Nvidia's gross margin at the time of recording was 66%
The gap illustrates the difference between revenue growth and true economic power

“Back in 2004, that gross margin was only 29% that they were able to command as a premium on their cards.”
1:28:00

#gross margin#financials#pricing power

Frameworks in this episode

The Bet-the-Company Reinvention Cycle

Value Creation vs. Value Capture Grading

Insights & moments

⚡Myth Buster· 2

Nvidia's First Near-Death: Betting on Quadrilaterals