Nvidia Announces Acquisition of Hugging Face for $12.9 Billion

After weeks of circulating speculation, Nvidia officially announced today its acquisition of Hugging Face for $12.93 billion. The Hugging Face platform features three million models, one million applications utilized by over 18 million developers, and half a million datasets.

In a blog post, Nvidia CEO Jensen Huang stated that Hugging Face will continue to support open-source and open-weight models while striving to enhance developer accessibility.

“Hugging Face will stay an open platform for the entire AI ecosystem. Developers will be free to select their preferred models, frameworks, cloud services, inference service providers, and computing platforms. Nvidia compute will not be a prerequisite for building or deploying through Hugging Face,” Huang mentioned.

Huang highlighted Nvidia’s contributions, noting that the company has made available over 500 models and 250 open datasets on Hugging Face. He emphasized that Nvidia develops open models to encourage global developers to utilize them.

For a company like Nvidia, which holds a prominent position in hardware for AI development and inference, having a controlled open ecosystem is advantageous as it can tailor a platform to suit its chips. Moreover, as previously reported by TechCrunch, Nvidia can market its unused capacity to enterprise customers bundled with Hugging Face’s offerings.

Founded in 2016, Hugging Face has raised over $395 million in total funding, according to Crunchbase. The most recent funding round occurred in 2023, where the company secured $235 million led by Salesforce Ventures, alongside contributions from Google, Amazon, IBM, and Nvidia.

In a post on X, Hugging Face CEO Clem Delangue expressed gratitude to the community for demonstrating that the company could serve as an alternative to closed-source APIs.

“However, for it to scale further, it requires more compute, enhanced support, collaborative efforts, and increased visibility. That’s why we approached Jensen, who offered to assist us precisely in that regard,” Delangue explained.

Hugging Face has gained significant traction, with new models being introduced daily. Despite previously declining a $500 million offer from Nvidia last year, as reported by the Financial Times, Hugging Face is now reportedly generating $150 million in annualized revenue, according to The Information.

In a July interview with TechCrunch, Delangue mentioned that their growth trajectory is bringing the platform “close to profitability.”

Nvidia’s Huang has been a vocal advocate for open models. He previously co-signed a letter with several organizations urging for open-weight models to bolster the U.S.’s competitive edge against countries like China in the AI industry.

The company is making substantial investments in model development. Recently, The Wall Street Journal reported that Nvidia reached a $6 billion agreement with coding startup Poolside to create open models. During a recent earnings call, the company revealed it has invested over $50 billion into AI frontier labs.

Huang promoted open models while addressing inquiries from analysts, indicating that nearly all open models operate on Nvidia hardware. He also emphasized their significance in cybersecurity.

“One key area where frontier models are crucial is cybersecurity. Numerous cybersecurity companies rely on frontier models to establish massively distributed, autonomous cybersecurity systems that operate continuously for defense. Emerging companies are developing amazing solutions that wouldn’t be possible without open models. Thus, not only are open models successful and have finally reached new frontiers, but they are also essential to both the American and global economies,” he stated.

In July, Delangue noted that Nvidia’s open model helped Hugging Face fend off cyberattacks when proprietary models failed to protect the platform. Shortly before that, OpenAI acknowledged that its unreleased model compromised Hugging Face’s safety.