Nvidia Sells the Chips Behind AI and Funds the Buyers Too
Its stakes in leading AI labs, plus seed funding and infrastructure bets, link the industry's growth to a single supplier.

Marketplace · Nvidia was AI's linchpin. Now what?The recent leaps in artificial intelligence have relied heavily on computing power, by making models bigger and giving them more computing to work with.
That computing power comes overwhelmingly from one company. Nvidia makes the specialized chips that train large AI systems and handle the heavy math when those systems answer questions.
Nvidia does not just sell those chips. It also helps finance the companies that buy them, putting it at the center of both supply and demand.
The arrangement is sometimes called circular financing, a web of investments and dependencies with Nvidia at the middle. Money flows out from Nvidia to AI companies, and can flow back as chip orders or rentals of systems packed with its chips.
Part of that web is direct ownership. Meghan McCarty Carino, host of Marketplace Tech, says Nvidia "has taken major equity stakes in AI companies to which it also sells chips," including OpenAI, xAI, and Anthropic.
Another part is early stage investing. Nvidia operates a large venture capital arm, essentially an in-house fund for startup bets, that provides seed money to young companies. Many of those startups then rent or buy Nvidia-powered computing to build their products.
Nvidia is also a major investor further down the chain, in the companies building AI infrastructure such as data centers and related systems. Those ties give Nvidia two roles in the growth of AI computing, as a supplier and as a backer.
Questions people ask
Why do AI models need so much computing power?
Training an AI model means adjusting billions of numerical values over and over until the system predicts well. That takes thousands of chips running in parallel for weeks or months, so access to computing often determines who can build the most capable model.
Why do economists worry about circular financing?
When one company owns part of its customers, demand can look stronger than it would on its own. If startups struggle or chip demand slows, losses can travel quickly through investments, loans, and purchase agreements that tie the same companies together.
What does it mean to take an equity stake in an AI company?
An equity stake means owning part of a company and sharing in its gains and losses. For Nvidia, a stake can align interests: if the AI developer grows and needs more computing, Nvidia may benefit both as an investor and as a chip supplier.

