The deal would give Nvidia a major foothold in open-source AI at a moment when open-source developers are increasingly closing the gap with closed systems from companies such as Anthropic and OpenAI. Hugging Face is a repository that hosts open-source AI models, as well as benchmark evaluations, and data sets used for AI training and testing.

The deal would also help Nvidia protect its dominant position in AI chips, which has been somewhat under threat as major tech companies like OpenAI, Google, Amazon and Anthropic build out their own chips to reduce their reliance on Nvidia hardware.

That’s because those who download open-source models from Hugging Face need to host and run those models on their own computing infrastructure, which usually involves Nvidia’s GPUs, either on premise or hosted in the cloud. Hugging Face also uses Nvidia GPUs to run the paid hosting services it offers developers.

A strong open-source ecosystem gives customers more options outside the closed labs and keeps more of the market tied to Nvidia’s hardware. Nvidia has already invested tens of billions of dollars in its own open-source models.

The acquisition would also mark a return for Nvidia to cloud computing, a business it reportedly scaled back roughly a year ago. Owning Hugging Face could also give the tech company a way to offload unused cloud capacity from the computing deals it has guaranteed for customers.

Business Insider, which first reported over the weekend that Hugging Face was fielding takeover interest, reported that the talks had not yet produced a signed agreement and could still fall apart. The Information cited an unnamed source it said had knowledge of the deal’s successful conclusion. Fortune could not independently verify the reports. Nvidia and Hugging Face did respond to requests for comment.

Hugging Face last raised money in 2023, when a $235 million round led by Salesforce Ventures valued it at $4.5 billion. It turned down a $500 million investment from Nvidia last year that would have valued it at $7 billion, reportedly because it did not want a single dominant investor. The 10-year old company has said it is nearing profitability, with revenue climbing to roughly $150 million a year.

The talks come as AI infrastructure companies increasingly get absorbed by larger players, following Stripe’s recent purchase of OpenRouter for more than $7 billion.

Hugging Face CEO Clem Delangue has publicly backed Nvidia’s open-source push amid fears that officials in Washington will restrict open-source models. Chinese labs, such as Moonshot AI with its Kimi K3 model, have released open-soure systems that rival leading U.S. models on benchmarks at a much lower cost, feeding concerns in Washington about falling behind in the AI race and national security.

Delangue signed a letter this year, along with Nvidia CEO Jensen Huang and more than twenty other companies, urging the government to support open models rather than restrict them, and he has made similar arguments in press interviews.

Hugging Face has also been in the middle of a major media story over the last few months after OpenAI disclosed in July that its own AI models were behind an autonomous cyberattack on the platform. According to OpenAI, its models escaped a sandboxed testing environment, accessed the internet and exploited a vulnerability to gain access to Hugging Face’s systems while trying to find information to cheat on an evaluation. Hugging Face reported the incident to local police before it knew OpenAI’s models were responsible, and the episode drew alarm from researchers and politicians.