Innovation & Wealth › The information age › The AI boom: from deep learning to ChatGPT
Nvidia (Jensen Huang)
founded April 1993· Company or fortune· Companies & fortunes
Well documentedSolid records, texts or physical evidence.
Nvidia, co-founded by Jensen Huang in 1993 to make graphics chips for video games, became the main supplier of chips for artificial intelligence.
See it on the Innovation & Wealth map →What happened
Huang, Chris Malachowsky and Curtis Priem planned Nvidia at a Denny's in San Jose. Its GeForce cards (1999) came to dominate PC gaming. In 2006 it released CUDA, software that let its chips do general math in parallel; in 2012 researchers trained the breakthrough AlexNet image system on Nvidia cards. Since ChatGPT (2022), tech giants have bought its data-center chips by the hundreds of thousands. Nvidia designs; TSMC manufactures. Its market value passed $5 trillion in October 2025.
Impact on society and the economy
It made a gaming part the core tool of the AI boom, made Huang's roughly 3.5% stake worth over $100 billion (estimates vary), and put chips at the center of US export controls.
How we know
Company filings (data-center revenue rose from about $15 billion in fiscal 2023 to about $115 billion in fiscal 2025), the 2012 AlexNet paper, and stock-market data.
Still debated
Whether its huge profit margins can last as Google, Amazon and others design their own AI chips, and whether AI spending is a bubble, are open questions.
Connections (3)
▸TSMC (Morris Chang) manufactures the chips of Nvidia (Jensen Huang)Well documented
The connection
Nvidia owns no chip factories. Since about 1998 it has relied mainly on TSMC to build its graphics and AI chips, including the H100 and Blackwell lines. TSMC's lead in advanced manufacturing and chip packaging is a main reason Nvidia can ship AI chips at scale, and also its main supply bottleneck.
When and how
Hsinchu, Taiwan, and Santa Clara, California, since the late 1990s; Chang has written that he asked Huang to become TSMC's CEO in 2013; Huang declined.
How we know
Nvidia's filings name TSMC as its main manufacturer, and both companies' leaders describe the partnership publicly.
Debates
Whether Nvidia will shift some production to Samsung or Intel, and how risk around Taiwan affects it, are open questions.
▸Nvidia (Jensen Huang) supplies the chips for The AI boom: from deep learning to ChatGPTWell documented
The connection
Training and running AI models takes trillions of simple calculations in parallel, which Nvidia's graphics chips do well. Its CUDA software (2006) made them programmable, and researchers adopted them after AlexNet (2012). Since ChatGPT its data-center chips, costing tens of thousands of dollars each, became the scarce input; it holds an estimated 80–90% of the AI chip market.
When and how
Santa Clara, California; Toronto in 2012; a demand surge from late 2022 led by Microsoft, Meta, Google, Amazon and OpenAI.
How we know
Nvidia's data-center revenue rose from about $15 billion in fiscal 2023 to $115 billion in fiscal 2025, and research papers name the hardware used.
Debates
Whether rival chips from Google, AMD, Amazon and Chinese designers will erode Nvidia's lead and its unusually high profit margins.
▸Nvidia (Jensen Huang) sells most of the chips in AI data center building boomWell documented
The connection
AI data centers are built around Nvidia hardware: racks of its GPUs linked by its networking gear, often the largest single cost of the project. Nvidia has also invested in its own customers, including a planned investment of up to $100 billion in OpenAI announced in 2025 and stakes in cloud renter CoreWeave, which critics call circular financing.
When and how
Worldwide from 2023; major buyers include Microsoft, Meta, Amazon, Google, Oracle, xAI and CoreWeave.
How we know
Nvidia's filings show a handful of huge customers supply much of its revenue, matching those firms' disclosed capital spending.
Debates
Whether Nvidia's investments in customers inflate demand, and whether data centers built around today's chips will age fast as new ones arrive yearly.