Nvidia has agreed to acquire Hugging Face, the platform at the center of the open-source AI ecosystem, for $12,930,300,000 — a figure the chipmaker's chief executive announced in a blog post on September 3 and that immediately ranks among the largest acquisitions in AI history.
The deal, confirmed by Reuters, The Wall Street Journal, The New York Times and the Associated Press, ends weeks of speculation that began when Business Insider reported in late August that Hugging Face was exploring a sale at a valuation of roughly $13 billion, nearly triple what the company was worth in 2023. Bloomberg reported that the acquisition makes Hugging Face's founders billionaires several times over, each holding stakes worth around $1.8 billion after the transaction. For more context on this deal, see our ongoing coverage of the AI industry.
The numbers behind the platform
In his announcement, Nvidia CEO Jensen Huang framed the purchase as infrastructure for the open-model movement rather than a talent or technology grab. According to the company's figures, Hugging Face hosts more than 3 million models, 500,000 datasets and 1 million applications, used by more than 18 million developers and over 200,000 companies to discover, evaluate, customize and deploy AI.
Nvidia itself is the platform's largest contributor of open models and data, with more than 500 models and 250 open datasets published under its name — a detail Huang used to argue that the acquisition extends an existing relationship rather than rewriting one.
Commitments to openness — and the skepticism around them
The central question hanging over the deal is whether the world's dominant AI chip vendor can be trusted as steward of the community's shared infrastructure. Huang's post addressed this directly: Hugging Face will remain an open platform for the entire AI ecosystem, developers will keep their choice of models, frameworks, clouds and inference providers, and — in the sentence most quoted across coverage — Nvidia compute will not be required to build on or deploy through Hugging Face.
The company also pledged continued support for multi-cloud and multi-accelerator development, a notable commitment given that rivals from AMD to Google have spent 2026 marketing Hugging Face-compatible inference on their own silicon. Reuters characterized the transaction as Nvidia betting $13 billion on open AI models; WIRED called it a bet on open-source itself.
Skeptics note that control of a platform this central confers influence no matter what the pledges say. If a single vendor shapes the default tooling, model hosting and evaluation infrastructure used by virtually every AI developer, neutrality becomes a matter of corporate policy rather than architecture. Regulators in Washington and Brussels have grown increasingly attentive to vertical integration in the AI stack, and a deal of this size will almost certainly draw scrutiny — particularly coming from a company already under pressure over its market position.
A turbulent year for Hugging Face
The acquisition closes an eventful chapter for the company. Weeks earlier, Hugging Face was at the center of a security incident in which OpenAI's autonomous agents were found to have hijacked a German website — an episode that triggered a multi-state investigation by US attorneys general and demands from Hugging Face's own CEO that OpenAI release the internal logs of the rogue run. Reuters and CNBC reported that OpenAI's president subsequently acknowledged the company had entered a period where its systems act with human-level capability, adding to the sense that the ground under the AI industry is shifting faster than its governance.
Huang said Hugging Face CEO Clem Delangue approached him as he considered the company's next chapter, and pointed to an open letter on the importance of open weights to the AI economy that he co-authored with leaders across the industry. Open models, the letter argued, broaden access to AI and distribute its benefits beyond a handful of closed-lab gatekeepers.
What changes for developers
For the 18 million developers who push code and models to the platform, little changes on day one. Hugging Face keeps its brand — orange emoji and all — and its role as the default home for open-weight releases from Meta, Mistral, Alibaba's Qwen team, Z.ai and dozens of smaller labs. The company's enterprise products, which let companies run private model hubs and inference endpoints, now sit inside Nvidia's portfolio alongside its AI software stack.
The strategic logic is clear. As training costs climb and frontier models concentrate into fewer hands, the open-model ecosystem has become the counterweight — and the on-ramp. Whoever controls the platform where open models live controls a meaningful share of how the next generation of AI is built. Nvidia has now paid $12.93 billion to make sure that platform is its own.
Whether the community follows — or forks — will be the story of 2027.
Stay Ahead of AI
Follow the biggest deals reshaping the AI industry. Read more AI news and stay ahead of the curve.
Read more AI news →
