Nvidia acquires Hugging Face for $12.93B in a deal priced to match the hugging face emoji

Nvidia acquires Hugging Face for $12.93B

Nvidia is buying Hugging Face for $12,930,300,000—a price deliberately set to match the decimal code of the hugging face emoji, the company's icon. The acquisition caps a ten-year journey from consumer chatbot startup to the central infrastructure layer for open-source AI.

Hugging Face did not begin as an AI model hub. In 2016, founder Clément Delangue (previously at French computer vision startup Moodstocks, acquired by Google) teamed with a mathematician and a patent lawyer to build an AI chatbot positioned as an emotional digital friend for teenagers. The product reached a million messages a day and accumulated over 100 million messages by 2018, drawing investment from Kevin Durant and other backers. But the consumer business stalled.

The pivot came when Google released BERT in 2018. Rather than compete, Hugging Face converted the model from TensorFlow to PyTorch and released it freely. Developers adopted the conversion immediately. The team then shifted strategy entirely: instead of building a single application, they became infrastructure—hosting and versioning models, attaching datasets, enabling discussion and demos through a Spaces product, and offering private repositories. GitHub's playbook applied to AI models.

The network effect took hold. Hugging Face scaled to 18 million developers and 200,000 companies using the platform, with 3 million models and over 500,000 datasets. The company raised under $400 million across multiple rounds while maintaining capital efficiency—they neither bought chips nor served models directly, relying instead on organic growth through the flywheel. Hugging Face became profitable in 2025.

The rejected offer

Nvidia approached Hugging Face with a $500 million investment at a $7 billion valuation in late 2025. The company declined. The signal was clear: any deeper commitment required full control, not a minority stake. That conversation led directly to the acquisition.

Jensen Huang's open-source play

Nvidia CEO Jensen Huang has positioned himself as the loudest advocate for open-source AI development. His strategic logic: Nvidia sells chips to closed-source labs building proprietary models—labs that are increasingly developing their own ASICs, creating competitive pressure. The open-source channel is different. Hugging Face becomes the front door. Developers discover and fine-tune models on the platform, then graduate to buying Nvidia inference capacity, chips, or compute when they're ready to deploy.

The acquisition also sends a market signal. Hugging Face raised roughly $400 million and delivered a $12.93 billion exit. Retention packages for the team reportedly total around $1 billion. The message to the open-source AI community is concrete: you can build a sustainable company and exit meaningfully in this space.

At the time of Hugging Face's last fundraise in August 2023 at a $4.5 billion valuation, the cap table included Salesforce, Google, Amazon, Nvidia, AMD, Intel, Qualcomm, and IBM—what the speakers describe as a "murderer's row of potential acquirers." The company had positioned itself explicitly as Switzerland: chip-agnostic, framework-neutral, supporting competing clouds and infrastructure providers. That independence is now over. Nvidia owns the default home for open-source AI models.