There’s a classic trick in the soda industry: give restaurants the fountain machines for cheap, then make a fortune selling them syrup forever. Nvidia appears to be running a far more expensive version of the same playbook, except the machines cost billions and the syrup is GPU compute.

The chipmaker has been pouring money into the companies that buy its hardware at a pace that makes venture capital firms look timid. In 2026 alone, Nvidia’s total equity investments in AI companies surpassed $40 billion, anchored by a massive $30 billion contribution to OpenAI’s funding round and a $2 billion stake in cloud provider CoreWeave.

The money pipeline

Let’s start with the headline number. Nvidia dropped $30 billion into OpenAI as part of a $110 billion funding round that valued the AI lab at $730 billion pre-money. To put that in perspective, $730 billion is roughly the GDP of Switzerland.

That wasn’t even the first big move. In January 2026, Nvidia invested $2 billion in CoreWeave, buying Class A shares at $87.20 per share. CoreWeave, for the uninitiated, is a cloud infrastructure company that runs massive GPU clusters, primarily Nvidia’s own chips, and rents out that compute power to AI companies. CoreWeave has plans to build over 5 GW of AI factories by 2030.

In English: Nvidia is bankrolling the company that builds the data centers that buy Nvidia’s chips. The circularity is not subtle.

These two deals form the backbone of that $40 billion-plus investment figure, though Nvidia has positions in other AI ecosystem companies as well. The company reported strong fiscal results early in 2026, with data center revenue growth continuing to climb alongside the relentless demand for AI accelerators.

Why “central bank” isn’t just a metaphor

Nvidia has inserted itself as the financier at multiple points in the AI capital chain. By investing in OpenAI, it ensures one of the world’s largest consumers of AI compute has the capital to keep buying hardware. By investing in CoreWeave, it ensures there’s a well-funded infrastructure provider building the physical facilities where its chips get deployed.

The company isn’t hiding from this characterization either. Its strategy explicitly positions it as both the dominant hardware supplier and a financial backer of its own ecosystem. It’s supplier and banker, rolled into one $3 trillion-plus market cap entity.

What this means for investors

For shareholders, the upside case is straightforward. Every dollar Nvidia invests into AI companies is essentially a dollar that catalyzes demand for its core product line. If OpenAI spends even a fraction of its $110 billion raise on Nvidia hardware, the return on that $30 billion investment gets compounded through chip sales on top of any equity appreciation.

The financial engineering is genuinely clever. Nvidia gets equity upside in what could become some of the most valuable companies on Earth. It simultaneously locks in hardware demand that feeds its income statement. And it builds switching costs into the ecosystem, because when your chip supplier is also your investor, changing vendors becomes a much more complicated conversation.

But there are real risks worth flagging. The circular nature of this arrangement creates concentration risk that could make regulators nervous. When the company supplying mission-critical hardware is also financing the buyers, the potential for conflicts of interest is obvious.

There’s also the question of what happens if AI spending cools. Nvidia would be exposed on two fronts simultaneously: declining chip sales and declining portfolio values.

No other chipmaker has the financial resources or market position to replicate this strategy. AMD and Intel are competing on silicon. Nvidia is competing on silicon, capital, and ecosystem control all at once.