Nvidia Confirms Its $13 Billion Acquisition of Hugging Face

What the filing confirms

The deal rumored in late August is now official. On Thursday, September 3, Nvidia announced the acquisition in a securities filing, and CEO Jensen Huang published a blog post confirming a total value of $12,930,300,000, roughly $13 billion. Of that, $11.9 billion goes to Hugging Face stockholders, and up to about $1 billion funds an equity retention plan for employees who join Nvidia. [1]

The price matches the reporting from late August, when The Information, TechCrunch, and CNBC first revealed that Nvidia had agreed to buy the company for approximately $12.9 billion. [3] [4] What changed this week is the paperwork: the filing makes the agreement binding and public, and both CEOs spent Thursday explaining it. Huang appeared unmoved by the idea of losing bidders, telling CNBC, "It doesn't matter who the other bidders were; it only matters who wins." [1]

What Nvidia promises about openness

The immediate worry for anyone who downloads open models is whether Hugging Face stops being neutral infrastructure. Both companies addressed that head on, stating the platform will remain open and compute-agnostic, meaning models trained on or served by non-Nvidia hardware should keep their home there. [1]

Huang's blog post framed the deal as scaling open models, not absorbing them: "Together, we will scale Hugging Face's platform, strengthen its infrastructure" and "expand access to AI for developers and institutions worldwide." He added that "as the opportunity for open models accelerates, Hugging Face can serve the global AI community at unprecedented scale." [1]

There is also a strategic reading. Nvidia has watched closed-model labs like OpenAI and Anthropic pull ahead on consumer mindshare, and in July it signed a letter urging US policymakers to support open-source AI. Owning the hub where open models live is a hedge against being squeezed out if closed platforms win. [1] CNN characterized the acquisition as a massive bet by the chipmaker to expand the reach of open-source AI. [2]

The founder's view, and one awkward callback

Hugging Face CEO Clement Delangue said in a post on X that the company will gain more compute and support from Nvidia under the new ownership. [1] On CNBC he reached back to a rough episode in the platform's recent history, calling the July OpenAI agent hack of Hugging Face a "turning point" and saying the company responded by leaning on open models for its own defense. [1]

That incident still stings for users. It showed that even the open-source community's central hub can be a target, and it is part of why this publication keeps suggesting you keep verified copies of the models you rely on. [1]

What local AI users should watch

For now, nothing about your workflow needs to change: downloads, model cards, and datasets work as they did yesterday. The questions that matter arrive over the next year, as the two companies integrate and the open-source community holds them to those promises. [1] Three things are worth watching.

First, watch whether Hugging Face's licensing and takedown policies stay community-driven or start to favor Nvidia's commercial interests. Second, watch compute grants: Nvidia ownership could mean more free inference and training credits, which helps small model builders. Third, watch rival hubs like ModelScope, which could pick up projects that want a home independent of any chip vendor. Open weights have been the quiet engine behind every good on-device model of the past two years, and whoever holds the hub holds a lot of leverage over that ecosystem. [1] [4]