NVIDIA has agreed to buy Hugging Face for $12.9 billion, adding one of the AI industry’s largest open-model platforms to its business.
More than 18 million developers, researchers and creators use Hugging Face, which has become a major hub for building with open AI. More than 200,000 companies also use the platform to find and deploy models.
The deal would give Hugging Face access to NVIDIA’s computing and engineering resources while keeping its existing open model approach in place.
According to NVIDIA, the platform will continue supporting technology from across the AI ecosystem after the acquisition closes.
Why It Matters: NVIDIA already powers much of the infrastructure companies use for AI. Hugging Face would extend its role into how companies find, evaluate and deploy models. For organizations using a mix of models, clouds and hardware, the acquisition adds another factor to decisions about AI architecture and vendor relationships. Hugging Face’s commitment to remain open will help preserve the flexibility many companies already have when building on the platform.
- Inside the Deal: About $11.9 billion will go to Hugging Face stockholders, with up to $1 billion set aside for equity awards aimed at retaining employees who join NVIDIA. The deal is expected to close in the first half of 2027, pending regulatory approval, and would become NVIDIA’s second-largest acquisition after its roughly $20 billion Groq purchase last year. Hugging Face was valued at $4.5 billion in 2023, when NVIDIA invested alongside companies including Google, Amazon, IBM and Salesforce.
- How It Came Together: Hugging Face CEO Clément Delangue approached Jensen Huang over the summer after deciding the company needed more compute, support and visibility to keep growing. The two sides reached an agreement within weeks. Founded in 2016, Hugging Face has since raised more than $395 million, and recent reporting puts its annualized revenue at about $150 million. Delangue said in July that the company was getting close to profitability.
- Keeping the Platform Open: Developers will still be able to use models and technology from across the industry. Hugging Face will continue working with different clouds, inference providers and computing platforms, and NVIDIA hardware will not be required. The two companies were already closely connected before the acquisition, with NVIDIA publishing more than 500 models and 250 open datasets on Hugging Face. Huang says NVIDIA’s resources can help improve the platform’s reliability, safety, model evaluation and deployment.
- Investing in Open Models: The acquisition follows NVIDIA’s $6 billion agreement with Poolside and more than $50 billion the company says it has invested in frontier AI labs. Huang publicly advocated for open weights, arguing that they give organizations a way to adapt existing AI to their needs without building every model themselves. Buying Hugging Face gives NVIDIA ownership of one of the largest communities built around that approach.
- Regulatory and Security Considerations: Rules around open-source AI could affect how Hugging Face operates under NVIDIA. In its SEC filing, the company noted that widely used models developed around the world, including in China, are routinely shared and modified through platforms like Hugging Face. Open models have also played a role in Hugging Face’s own cybersecurity work. Delangue recently said an NVIDIA version of a Chinese open model helped the company respond to an AI-related security incident after proprietary models failed.
Go Deeper -> NVIDIA to Acquire Hugging Face – NVIDIA
Hugging Face approached Nvidia’s Huang weeks ahead of $12.9B acquisition, CEO tells CNBC – CNBC
Nvidia confirms it will buy Hugging Face for $12.9 billion – TechCrunch
Nvidia to buy open-source AI platform Hugging Face for $13bn – FinancialTimes
