Why Nvidia loves funding startups

Nvidia founder and CEO Jensen Huang.

No company looms over the AI startup ecosystem like Nvidia. The $5.5 trillion chipmaker has backed a startup once every five days this year, participating in $266 billion worth of deals as of mid-August.

Don't expect the startup lovefest to change anytime soon.

Nvidia's earnings this week shed light on why Jensen Huang's behemoth cares about the littler guys, which is rooted in its experience backing AI labs. The company has invested nearly $50 billion to date in frontier labs, a sum that shockingly represents a "small fraction" of its free cash flow over the period, Huang said. In return, these AI labs are expected to contribute a quarter of Nvidia's business next fiscal year.

Several of those AI labs have outgrown Nvidia, a primary driver of the company's push to raise $500 billion of third-party capital in partnership with Apollo Global Management, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR.

Nvidia is now focused on what comes after the frontier labs.

One front is open models: Huang said "nearly all" open models run on Nvidia, thanks to its ubiquity and flexibility as a platform. He sees open models as key to helping startups and enterprises build domain-specific, proprietary AI.

This week, The Information reported that the company agreed to pay $12.9 billion for Hugging Face, a kind of GitHub for hosting open-source AI models and datasets.

Another front is agentic AI, which Huang cited as a major driver of the demand that Nvidia expects to double in the coming year. "We have 40,000 employees, roughly. In the future, we'll have 400,000 agents, 4 million agents. And those agents are running continuously," Huang said.

AI agents will gobble up compute, and the startups making them will be great customers for Nvidia. The company estimates that around 70% of AI venture dollars are spent on compute, which would mean that more than $280 billion went from startups to the pockets of Nvidia, its customers and its competitors in the first half of 2026.

As Big Tech and the frontier labs graduate to owning their own infrastructure, Nvidia is focused on further expanding its addressable market by supporting everyone from startups to neoclouds, enterprises and governments.

This article originally appeared on PitchBook News