Nvidia buys Paris-born Hugging Face for $12.93bn, its biggest deal ever

Nvidia has agreed to buy Hugging Face for $12.93 billion, the chipmaker's largest acquisition to date. The deal, announced on 3 September, hands Nvidia control of the platform that has become the default home for open-source AI models, with closing expected in the first half of 2027 subject to regulatory approval.

The price tag carries a deliberate flourish: $12,930,300,000 converts, digit for digit, to the Unicode code point behind the 🤗 emoji that has long been Hugging Face's logo. Reported deal terms put the structure at roughly $11.9 billion in cash plus up to $1 billion in equity retained for staff.

A Paris startup becomes Nvidia's biggest bet

Hugging Face was founded in Paris in 2016 by Clément Delangue, Julien Chaumond and Thomas Wolf. It grew from a niche chatbot app into what amounts to a GitHub for machine learning, a place where developers publish and download models, datasets and applications rather than write code from scratch.

Nvidia's own figures put the catalogue at upwards of 3 million models, half a million datasets and a million hosted applications, with over 18 million developers and more than 200,000 companies on the platform. The company has raised over $395 million since founding, including a $235 million round in 2023 led by Salesforce Ventures with backing from Google, Amazon, IBM and Nvidia itself.

Nvidia was already Hugging Face's most prolific outside contributor, having released more than 500 models and 250 open datasets on the platform, according to its announcement. Reports citing the Financial Times say Nvidia had previously offered around $500 million for the company last year and was turned down, a sign of how sharply Hugging Face's perceived value has risen in little over a year.

The neutrality promise

Nvidia built its pitch for the deal around a pledge not to disturb Hugging Face's open character. Chief executive Jensen Huang wrote in the announcement that developers would keep choosing their own models, frameworks, clouds and chips, and that Nvidia hardware would not be a requirement to use the platform.

"Hugging Face will remain an open platform for the entire AI ecosystem… Nvidia compute will not be required to build on or deploy through Hugging Face," Huang wrote.

Huang said Delangue had approached him about Hugging Face's next chapter, and that Nvidia would use its infrastructure and engineering scale to improve the platform's reliability, model evaluation and deployment capabilities. Delangue, in a post on social media, framed the deal as a way to get the backing the community needed at a larger scale, saying the company had gone to Nvidia because reaching that scale required "more compute, more support, more collaboration, and more visibility."

For a company that The Information reported was running at roughly $150 million in annualised revenue as of last month, a price above $12 billion is a steep multiple, and some analysts have already said so publicly. Delangue told reporters in July that the company's growth was pushing it "close to profitability" even before the deal, but the gap between revenue and purchase price will invite scrutiny as the acquisition moves toward close.

Why it matters beyond one company

For Nvidia, owning the place where most open-weight models are published and tested is a distribution advantage as much as a product one: it sits upstream of decisions about which chips and clouds those models eventually run on. Huang has separately argued that open models are central to cybersecurity and to keeping AI development distributed rather than concentrated in a handful of closed labs.

For European founders, the deal is a reminder that one of the continent's most consequential AI infrastructure companies is ending up under American corporate control, even as its product keeps the open-source label. Hugging Face's team, including its three co-founders, is expected to stay on, and the company says its brand and mission will not change. Whether a hardware giant can own the market's most-used open hub without eventually steering traffic toward its own chips and cloud capacity is the question investors and regulators will be watching as the deal heads toward its scheduled 2027 close.