Nvidia has agreed to buy Hugging Face for $12.9 billion, The Information reported Wednesday night, folding the neutral distribution hub for open-source AI, more than 2 million models, tens of thousands of training datasets, and over 13 million developers who pull from it, into the chipmaker whose GPUs those models mostly run on. Bloomberg, citing Business Insider, put the potential valuation closer to $13 billion and cautioned that no agreement is signed. A source familiar with the matter told CNBC only that they could “confirm acquisition [by Nvidia] has been part of ongoing and recent talks.”
The price sits against a revealing trajectory. Salesforce Ventures led a 2023 round at a $4.5 billion valuation. Last year Nvidia offered $500 million at a $7 billion valuation, and Hugging Face turned it down. Roughly two months ago, annual revenue was near $100 million; today it’s approaching $150 million a year. The refusal has aged into leverage.
For Nvidia, the logic isn’t mysterious. OpenAI, Google, Amazon, and Anthropic are all building their own silicon, the four customers most likely to design Nvidia out of the frontier stack. Owning the platform where open-weight models are discovered, fine-tuned, and deployed is a hedge against that. It’s also a distribution channel the company has quietly wanted for a while, roughly a year after it scaled back DGX Cloud and shifted toward guaranteeing tens of billions in cloud commitments on behalf of customers rather than selling compute directly.
What changes for the broader ecosystem is subtler and probably more consequential. Hugging Face’s value to the open-model economy has been its posture: neutral, model-agnostic, hardware-agnostic, a place where Spaces and Inference Endpoints monetize without gatekeeping the catalog. A chipmaker with a direct interest in developer lock-in now controls the roadmap, the access rules, and the compute pricing on the platform where the affordable AI stack lives. Every downstream tool built on pulls from that hub, the Apache-2.0 release of Alibaba’s Qwen3 among them, inherits whatever terms the new owner chooses to set.
The strategic frame around the deal echoes the mid-2010s cloud-consolidation cycle, when neutral infrastructure layers were absorbed into the platforms that most needed them not to be neutral. GitHub in 2018 is the cleanest parallel: a horizontal developer commons acquired by a company with obvious downstream products to sell. The commons kept working. The defaults changed.
Whether this one closes is still open. Business Insider’s caveat matters: no signature, and deals at this scale collapse regularly under regulatory scrutiny, especially when the buyer is already the object of antitrust attention across three continents and the target is the connective tissue of an entire open-source category. Nvidia turning down at $7 billion and coming back at $12.9 billion is the kind of number that invites a second look, from Hugging Face’s board and from everyone else.
Sources
- https://www.bloomberg.com/news/articles/2026-08-27/nvidia-discussed-buying-ai-startup-hugging-face-insider-says
- https://www.cnbc.com/2026/08/27/nvidia-hugging-face-acquisition.html
- https://techcrunch.com/2026/08/26/nvidia-closes-in-on-hugging-face-acquisition/
- https://fortune.com/2026/08/27/nvidia-hugging-face-billion-dollar-deal-open-source-ai/
- https://siliconangle.com/2026/08/27/nvidia-reportedly-acquires-ai-project-hosting-platform-hugging-face-for-12-9b/