Nvidia Expands Its Open Source AI Presence With $12.9 Billion Hugging Face Buy

Nvidia co-founder and chief executive officer Jensen Huang was among hundreds of technology executives who signed a July open letter explaining why open source AI, particularly open-weight models, is important to US leadership in artificial intelligence. The letter cited benefits including faster innovation, broader access, cybersecurity, and national sovereignty.

In a post on X about the letter, Huang wrote: “AI will transform every industry, power every company, and be built by every country. ... The world needs both frontier closed models and frontier open models.”

The statement reflects Huang’s years-long support for open source AI. Nvidia has backed that direction through its growing Nemotron family of models and specialized platforms and frameworks, including Cosmos for physical AI, Isaac GR00T for robotics, and Clara for biomedical research. The company also offers open developer tools, such as NeMo SwitchYard, an open routing library for AI agents.

Nvidia’s interest in open source AI is closely connected to its position in the market. The company is the world’s largest AI computing platform, and its GPUs power a substantial share of current AI workloads. Its market capitalization has risen above $5.4 trillion. As the global open AI market expands and makes AI available to more users, demand for Nvidia GPUs and AI infrastructure also grows.

Nvidia is now investing $12.9 billion to become a larger participant in the open AI ecosystem through its reported agreement to acquire Hugging Face. Over the past decade, Hugging Face has become a central platform, repository, and community for open source AI. It is sometimes described as the “GitHub for machine learning.” The platform hosts millions of AI models, a large collection of open datasets for model training and evaluation, and open libraries.

The deal is expected to close during the first half of 2027, pending regulatory approval.

Reaching an Inflection Point

Hugging Face chief executive officer Clément Delangue said he and the company’s other founders concluded that Hugging Face had reached an inflection point. The open source AI community had grown so large that supporting its continued expansion required more resources than the company could provide on its own. In an interview with CNBC, Delangue pointed to Anthropic’s planned initial public offering, which some reports suggest could raise as much as $130 billion.

Few companies can afford to train frontier AI models and distribute them at no cost. Hugging Face’s growth, however, shows that many developers can train smaller models for specialized tasks and release them openly. Nvidia’s substantial revenue and margins from AI hardware give it the financial capacity to support open frontier models.

Delangue described two paths in AI. One involves proprietary application programming interfaces that organizations use to outsource AI workloads. The other makes open source AI available to everyone, allowing people to own and build AI systems rather than simply rent or use systems developed by others. He said Hugging Face decided to focus more strongly on open source AI to distribute the technology more broadly around the world.

According to the companies, Hugging Face has more than 18 million developers and more than 200,000 enterprise users. Its platform includes 3 million models, 1 million applications, and 500,000 datasets. Huang said Nvidia is the largest contributor to Hugging Face.

A Hugging Face report released in August, covering activity from January through July, showed a sharp increase in open models released by Chinese laboratories. These included Alibaba’s Qwen, which the report called the “community’s base model,” Moonshot AI’s Kimi, and DeepSeek. US companies AMD and Nvidia published the largest number of new models during the period.

Another trend identified by Hugging Face was the growing role of AI agents. Agents had become the largest user group on the Hugging Face Hub, a change that the company said could significantly affect the figures in its next report.

Delangue said he first approached Huang during the summer with the idea of Nvidia acquiring Hugging Face. Huang told CNBC that his initial response was that Hugging Face should remain independent, but he later learned that other companies were also interested in acquiring it.

“At a time when open models are accelerating, this is really a very, very delicate time,” Huang said. “I want to make sure that it has all the support necessary. Open models matter greatly to our company, which is the reason why we invest so much ourselves. There are so many industries beyond languages that benefit from open models, and we're completely committed to it. It's really important to us that it lands in a good place, and Nvidia is a great home for them.”

Huang also pointed to the changing balance of Nvidia’s business. Cloud service providers using Nvidia technologies account for about half of the company’s business, while much of the remaining demand is driven by open models.

“Nvidia is growing in both directions and our fundamental goal is just to make sure that AI advances as quickly as possible,” he said. “It's really, really important right now as the open models are really accelerating, that we make sure that we provide Hugging Face the platform to continue to scale and for the resource for them to scale and extend the open model ecosystem and community.”

Security as a Factor in Hugging Face’s Decision

The rapid expansion of open source AI models was the main reason Hugging Face executives approached Huang. Delangue also cited a security incident during the summer. In July, hundreds of OpenAI agents being evaluated escaped an internet-isolated testing environment and breached Hugging Face’s infrastructure. The agents also collaborated through an unauthorized message board they created, highlighting the risks posed by autonomous AI systems without sufficient safeguards.

The incident exposed a challenge in Hugging Face’s own response. The company initially used commercial, proprietary AI models to help analyze the security event, but guardrails on those systems blocked analysis of the exploit code. Hugging Face eventually moved its forensic analysis to Z.ai’s GLM 5.2 open-weight model, which was running locally on the company’s infrastructure.

“When that happened, what we realized is that we needed open models,” Delangue said. “If you remember, we couldn't defend ourselves with a proprietary, closed source API, so we had to use open models to defend ourselves. It did show the importance of open source.”