On Tuesday, Nvidia ($NVDA) said its stakes in private companies are now worth $65.5 billion, up from $22 billion in January, a rise that came from new investments and markups of existing portfolio companies. A markup is Nvidia raising its own estimate of what a stake is worth; no cash comes back when it happens.
Most of the jump came from spending. Nvidia put $16 billion into equities in the last quarter alone and $58 billion over the first nine months of its fiscal year, a figure that includes shares of public companies as well as private ones. In November 2025, the company said it had spent $5 billion on equity over the same nine months of the prior year. PitchBook counts nearly 350 startups that took Nvidia money in the three years since late 2023, among them Anthropic, OpenAI, Thinking Machines, Reflection, Lambda, Nscale, and Crusoe.
That spending now has no settled owner. Jensen Huang is leaving, so an interim chief executive and a successor not yet named will decide the next quarter’s checks. My argument is that the $65.5 billion is a record of one founder’s appetite for deals, and that each new bet will now be a single judgment call by people who have said they want to make fewer of them.
One Founder’s Judgment Set the Pace
In late 2023, Nvidia said the goal of its corporate investing was to expand the AI ecosystem by funding startups it saw as able to reshape their markets. As stated, that goal sets no ceiling, no share of cash flow, and no return target, so it could justify $5 billion over nine months or more than ten times as much. The same open mandate reached past startups: Nvidia has given $118 billion in guarantees, $105 billion of it a backstop for OpenAI’s $500 billion data center in Ohio. A guarantee here is a promise to pay a lender if the borrower cannot.
According to TechCrunch, Huang’s pace of dealmaking worried Kress, who spoke for the company on Tuesday’s earnings call. Investors told the publication they expect her to be more selective with new investments and to write smaller checks into companies Nvidia already owns a piece of. VCs, in the publication’s words, are bracing for “a meaningful pullback in capital.”
A Successor Inherits Stakes, Not Reasons
An organization is safest when every key person has a designated successor who has already apprenticed in the job1. Nvidia’s investing ran the other way. Its rule for when to write a check lived in Huang’s judgment, and it isn’t clear who at Nvidia learned to make those calls alongside him.
A museum whose director chose its art on personal taste for a decade faces the same handover. The collection stays on the walls, the way Nvidia’s stakes stay on its balance sheet. Artists who sold to the museum every year lose their steadiest patron, as startups that counted on Nvidia’s checks would. The appraised value of the collection also leaned on the director’s purchases, so a successor who buys less can lower what the museum owns without selling a single painting.
Investors Pay for Recurring Customers
Kress said Nvidia would continue providing financing assistance to strong customers, which means some of this money goes to companies that also pay Nvidia. She added that “as the market matures and additional sources of capital become available,” the company would be “rigorously evaluating future investments, commitments, and guarantees.” That sentence describes a lender stepping back as others step in, not a program growing.
Public investors reward a business that collects recurring payments from lots and lots of customers2. Financing a customer can help create exactly that revenue, which is why a pullback would show up twice: once in the portfolio and once in sales. Kress drew the line herself: “We intend to return all excess free cash flow net of strategic uses to shareholders during this last quarter and going forward.”
What the Next Quarter Will Show
Nvidia did not build an investing program that outlives its founder; it built a portfolio of his decisions, and its future now sits inside four words, “net of strategic uses.” The money doesn’t stop, since Kress committed to financing strong customers. What changes is the default. Under Huang, a deal needed a reason to be refused; under Kress, a deal needs a case for why it is strategic, and the first claim on spare cash belongs to shareholders.
If that reading is right, the $16 billion quarterly pace could fall sharply under new leadership, and the next disclosure might show the $65.5 billion rising mostly on markups rather than new money. Startups such as Thinking Machines and Reflection may then turn to other investors for their next financing. Those investors, not Nvidia, would set the prices that Nvidia’s markups depend on, and the first deals Nvidia sits out could test some of the gains from $22 billion to $65.5 billion.
Huang turned Nvidia’s spare cash into stakes in nearly 350 startups by deciding, deal after deal, that the next one was strategic; his successor inherits every one of those stakes and none of the habit that produced them.
Sources
1 Dalio, Ray. Principles.
2 Mehta, Nick, Dan Steinman, Lincoln Murphy. Customer Success.
“Nvidia’s startup stakes ballooned to $65B, but Huang’s exit puts future deals in doubt.” techcrunch.com.

Leave a Reply