Nvidia’s Hugging Face Acquisition Is a $12.9 Billion Bet on Open-Source AI
| Source: Wired AI
Tags: Nvidia, Hugging Face, Jensen Huang, Clément Delangue, open-source, acquisition, open-weights
Nvidia acquires Hugging Face for $12.9 billion, gaining control of the internet's largest open-source AI repository — 3 million models, 18 million developers — while pledging the platform stays cloud-agnostic, repositioning Nvidia from chip vendor to AI platform owner.
Details
Nvidia has formally closed a $12.93 billion deal to acquire Hugging Face, the platform hosting over 3 million open-source AI models, 1 million applications, and 500,000 datasets used by 18 million developers. The acquisition, rumored for weeks, was made official Thursday. The strategic logic is clear: as Amazon, Meta, and other hyperscalers design custom chips to reduce Nvidia GPU dependence, Nvidia is embedding itself in the software layer. Controlling Hugging Face gives Nvidia a critical position in the open-source AI stack — developers who build on the platform will gravitate toward Nvidia infrastructure, even without a mandate. CEO Jensen Huang committed that 'Nvidia compute will not be required to build on or deploy through Hugging Face' and pledged to maintain open standards. Notably, Hugging Face had rejected a $500 million Nvidia acquisition offer as recently as last year, according to the Financial Times. At $12.93 billion against $150 million in annualized revenue, the final price reflects how much the strategic calculus shifted as custom-chip competition intensified. Nvidia already has 500+ models and 250 open datasets on the platform. For enterprises, near-term continuity is expected. The longer-term risk is governance drift: an Nvidia-owned Hugging Face may gradually optimize for Nvidia GPU compatibility in ways that disadvantage competing hardware, regardless of current pledges.