Nvidia Buys Hugging Face for $12.93 Billion

Nvidia just made the second-biggest acquisition in its history — and it did not buy a chip company. On September 3, CEO Jensen Huang confirmed a $12.93 billion deal to acquire Hugging Face, the open-source platform that has quietly become the default home for AI models, datasets, and applications. The chipmaker is stepping well outside hardware for the first time at this scale.

Server room with rows of black and white network cabinets under blue cables
Photo by Taylor Vick on Unsplash

What Hugging Face Actually Is

If you have never used it directly, Hugging Face is still likely touching software you use every day. Founded in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf, the platform now hosts more than 3 million AI models and 500,000 datasets, serving over 18 million developers, researchers, and creators. More than 200,000 companies rely on it to find, test, modify, and deploy AI models — making it something close to GitHub’s role for code, but built specifically around machine learning.

Why Nvidia Wanted It

Nvidia built its trillion-dollar valuation on selling the chips that train and run AI models. Buying Hugging Face pushes it a layer higher — into owning a piece of how those models actually get discovered, shared, and deployed in the first place. Huang described the logic plainly to CNBC: “This is such a large growth driver of our company, and together we can scale the open community even faster than they’re able to do today.” Nvidia is already Hugging Face’s single largest contributor of open models and data, having published more than 500 models and 250 datasets on the platform before this deal was ever announced.

Hugging Face Approached Nvidia First

According to Delangue, the deal did not start as an acquisition pitch from Nvidia — Hugging Face came to Huang. “During the summer, I think we realized that Hugging Face and open-source AI in general was at the turning point, and that it needed more, more resources, more scale, more visibility,” Huang told CNBC. Delangue said discussions moved quickly once he raised the idea, framing Nvidia as “a perfect home.” Notably, Hugging Face had reportedly turned down a much smaller $500 million offer from Nvidia back in late 2025, suggesting the company held out for a deal roughly 25 times larger less than a year later.

The Deal Nvidia Says Won’t Lock Out Competitors

The most pointed question hanging over the acquisition is whether Nvidia will quietly turn Hugging Face into a walled garden favoring its own hardware. Nvidia’s answer, at least on paper, is no. The company says Hugging Face will keep supporting competing accelerators, including hardware from AMD and Intel, along with multiple cloud providers and AI frameworks. “Hugging Face will remain an open platform for the entire AI ecosystem,” Huang wrote in a blog post. “Developers will choose the models they want, the frameworks they want, the clouds and inference service providers they want and the computing platforms they want.”

How the $12.93 Billion Breaks Down

Roughly $11.9 billion of the deal goes to Hugging Face’s investors, with up to $1 billion set aside specifically for employee retention — a structure aimed at keeping the team that built the platform in place through the transition. In context, the price tag is a rounding error for Nvidia: the company’s market capitalization currently sits near $5.5 trillion, meaning this deal represents roughly a quarter of one percent of Nvidia’s total value. It is Nvidia’s second-largest acquisition ever, trailing only the $20 billion purchase of chipmaker Groq’s assets in December 2025, and well ahead of its previous largest deal, the near-$7 billion purchase of Israeli chipmaker Mellanox back in 2019.

Bottom Line

This deal is less about Nvidia needing new revenue and more about controlling distribution. Chips train the models; Hugging Face is where a huge share of the world actually finds, tests, and ships them. If Nvidia keeps its promise to stay hardware-neutral on the platform, this mostly strengthens open-source AI’s infrastructure without changing who can use it. If that promise slips over time, it becomes a much bigger story about how much of the AI stack one company can quietly control. For now, Hugging Face keeps its name, its team, and its openness — just with Nvidia’s balance sheet behind it.

Leave a Reply

Your email address will not be published. Required fields are marked *