Nvidia announced plans to acquire Hugging Face for $12.9 billion, its second-largest purchase after spending $20 billion on Groq's assets.
Hugging Face operates as the primary hub for publishing, sharing and running open-source AI models. The platform hosts more than 3 million models and 500,000 datasets used by over 18 million developers and 200,000 companies, according to Nvidia.
Nvidia CEO Jensen Huang told CNBC the valuation reflects competitive pressure to win the asset. He said open models drive roughly half of Nvidia's revenue, and that Nvidia is the largest contributor of open-source models globally.
The acquisition grants Nvidia weeks of advance visibility into which models developers are building, which datasets they prioritize and which architectures are gaining traction before that information reaches mainstream tech news, according to Naveen Chhabra, principal analyst at Forrester.
Gil Luria, head of technology research at D.A. Davidson, characterized the deal as defensive—analogous to Microsoft's 2018 acquisition of GitHub for $7.5 billion. Luria said that if a rival lab such as Google owned Hugging Face, it could suppress open-source AI development while allowing proprietary platforms to consolidate market power.
The move extends Nvidia's reach beyond chips into the software and development layers of the AI stack. But it creates a structural problem: Nvidia is acquiring a platform historically seen as hardware-agnostic, which could trigger customer concerns about whether Nvidia will favor its own silicon or restrict access for competitors' chips.
