Nvidia has confirmed it is acquiring Hugging Face, the de facto home of open-source AI, for approximately $12.93 billion, ending weeks of speculation and marking the chipmaker's largest move yet beyond hardware. The confirmation came Thursday morning Pacific time, along with a pledge from Nvidia CEO Jensen Huang that the platform will remain open to the entire AI ecosystem.
Hugging Face hosts three million models, one million applications used by more than 18 million developers, and half a million datasets, scale that has made it the default distribution point for open-weight AI. That reach is precisely what Nvidia is buying. For more context on this story, see our ongoing breaking AI news.
Huang's Pledge: No Forced Lock-In
The central question hanging over the deal has been whether Nvidia would fold Hugging Face into its commercial stack. Huang addressed it directly in a blog post announcing the acquisition.
"Hugging Face will remain an open platform for the entire AI ecosystem," Huang said. "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. Nvidia compute will not be required to build on or deploy through Hugging Face."
The assurance matters because of Nvidia's position: the company supplies the dominant hardware platform for both AI training and inference, and controlling a layer of the stack that developers touch daily gives it influence no competitor can match. Critics of large platform acquisitions will watch whether Hugging Face's tooling, licensing defaults and featured models gradually favor Nvidia's ecosystem in practice, even if no formal requirement exists.
Huang pointed to Nvidia's own record on the platform as evidence of good faith, saying the company has released more than 500 models and 250 open datasets on Hugging Face.
From Rejection to Record Deal
The acquisition caps a striking reversal. According to the Financial Times, Hugging Face rejected a $500 million deal from Nvidia just last year. Founded in 2016, the company has raised more than $395 million to date, with its last round closing in 2023 at $235 million, led by Salesforce Ventures with participation from Google, Amazon, IBM and Nvidia itself.
Hugging Face CEO Clem Delangue, who according to CNBC approached Huang weeks before the deal was struck, framed the sale as a way to scale the open-source alternative rather than abandon it. In a post on X, he thanked the community for showing that open models could compete with closed-source APIs.
"But for it to happen at [a] larger scale, it needs more compute, more support, more collaboration, and more visibility," Delangue wrote. "That's why we went to talk to Jensen, who offered to do exactly that with us."
The platform's business fundamentals made the valuation discussable. The Information reported last month that Hugging Face is generating roughly $150 million in annualized revenue, and Delangue told TechCrunch in July that the company's growth rate had it "close to profitability."
The Security Saga That Shadowed the Deal
Unusually for a blockbuster acquisition, the announcement arrives in the aftermath of a security crisis on the platform itself. Days before Nvidia's confirmation, OpenAI admitted that its own agents had breached Hugging Face, an incident that prompted the lab to pause some model development for roughly two weeks this summer.
Delangue said in July that an open model from Nvidia helped Hugging Face defend against cyberattacks after proprietary models failed to protect the platform. Huang echoed the security theme on Nvidia's earnings call, arguing that frontier models are vital to cybersecurity and that autonomous defense systems emerging across the industry "couldn't do it without open models."
A Larger Open-Model Strategy
The purchase fits a broader Nvidia campaign around open-weight AI. Last month, the Wall Street Journal reported that Nvidia struck a $6 billion deal with coding startup Poolside to develop open models, and the company said on its recent earnings call that it has infused more than $50 billion into frontier AI labs. Huang has also co-signed public letters advocating open-weight models as a strategic asset for the United States against rivals like China.
The logic is straightforward: if nearly all open models run on Nvidia hardware, as Huang told analysts, then owning the platform where those models are distributed tightens the relationship between the world's most valuable chipmaker and the developers who program it.
What It Means for Developers
For the 18 million developers who use Hugging Face, nothing changes immediately. Huang's commitments are explicit, and regulators are likely to scrutinize whether they are kept. The meaningful tests will come over time: whether Hugging Face's governance remains independent, whether model hosting stays neutral across cloud providers, and whether the community that built the platform's value continues to treat it as neutral ground.
The deal still requires regulatory review in multiple jurisdictions, and antitrust scrutiny of any Nvidia expansion is all but guaranteed given the company's market position. Neither company has publicly projected a closing timeline.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
