Nvidia acquires Hugging Face for $12.9 billion, raising questions about open model hosting

The acquisition, first reported by The Register's Dan Robinson, brings together the dominant AI hardware maker and the go-to platform for open-weight and open-source AI models. Hugging Face, often described as the 'GitHub of AI,' hosts thousands of models and datasets and maintains an active community of developers.

During a conference call with journalists, Hugging Face CEO Clement Delangue said the deal would provide the resources needed to reach 100 million developers, up from an estimated 80 million users today. He stressed that Nvidia would take a hands-off approach, allowing Hugging Face to continue its existing mission. 'They've entered into this agreement which is due to close next year, provided that they meet all the regulatory requirements,' Robinson noted on The Register's Kettle podcast.

However, skepticism is widespread. Senior reporter Tom Claburn and host Brandon Vigliarolo highlighted concerns that Nvidia, already a powerhouse in AI infrastructure, could eventually exert influence over the repository. 'As we've been seeing, there are quite a few concerns around that, as Nvidia is a very powerful company in the AI market,' Robinson said. Vigliarolo drew parallels to Microsoft's acquisition of GitHub, which has since grown to around 180 million developers.

The deal underscores the escalating battle for control over the AI ecosystem, with Hugging Face positioned as a neutral hub for open models. Whether that neutrality survives Nvidia's ownership remains to be seen.

Why This Matters

The acquisition concentrates power over AI model distribution in Nvidia, which already dominates AI hardware and software stacks.
Developers who rely on Hugging Face for open models face uncertainty about long-term access and neutrality.
The deal could shape the direction of open AI development, especially if Nvidia prioritises its own platforms.

Background

Hugging Face emerged as the central repository for open AI models, hosting everything from small research checkpoints to large language models like Llama. Its community-driven model made it a key resource for academics, startups, and hobbyists. Nvidia, meanwhile, has built a near-monopoly on AI training and inference hardware and has been expanding into software with tools like CUDA and NeMo. This acquisition marks its most significant move into the platform layer.

Key Perspectives

Nvidia and Hugging Face leadership: Argue the deal brings much-needed compute and storage resources, allowing Hugging Face to scale to 100 million developers while maintaining operational independence. Critics and skeptics: Warn that Nvidia's track record suggests eventual integration and control, potentially limiting openness and favouring Nvidia's own models and libraries. Regulators: Will need to scrutinise the deal for anti-competitive effects, especially given Nvidia's dominance in AI hardware and Hugging Face's central role in model distribution.

What to Watch

Regulatory approval processes in major jurisdictions (US, EU, UK) and any conditions imposed.
Any changes to Hugging Face's model hosting policies or community governance.
Reactions from major open model contributors, including Meta and Mistral, who may reconsider their reliance on the platform.

ZOTPAPER