--- title: "Nvidia Strikes $12.9 Billion Hugging Face Deal, Report Says" date: 2026-08-27 author: "Barry Elad" featured_image: "https://sqmagazine.co.uk/wp-content/uploads/2026/08/nvidia-strikes-12-9-billion-hugging-face-deal-report-says-1.jpg" categories: - name: "Artificial Intelligence" url: "/artificial-intelligence.md" tags: - name: "News" url: "/tag/news.md" --- # Nvidia Strikes $12.9 Billion Hugging Face Deal, Report Says Nvidia has agreed to buy Hugging Face for $12.9 billion, according to an 08/26/2026 account in The Information citing a person with knowledge of the deal. Neither company has confirmed it. ## Key Takeaways - The Information describes a closed agreement, while Business Insider describes talks that could still fall apart. - Business Insider places the deal value above $13 billion, higher than the price in the competing account. - Hugging Face turned down a $500 million Nvidia investment last year that valued it at $7 billion. - Hugging Face supports models and hardware from Nvidia rivals including AMD and Intel, a position ownership would complicate. - Nvidia holds $47.9 billion in private companies and has $18 billion committed to equity investments through fiscal 2027. ## What We Know? Two accounts of the same transaction landed days apart, and they do not match. The Information has Nvidia and Hugging Face already agreed. Business Insider has the two sides in acquisition conversations over recent weeks at a value above $13 billion, with no deal reached and a real chance talks collapse. Both accounts rest on a single unnamed person. Nvidia and Hugging Face did not respond to requests for comment, leaving nothing on the record from either party. Microsoft also met with Hugging Face, and those talks are no longer active. Nvidia already owns a slice of what it is said to be buying. The chipmaker joined Salesforce and Alphabet’s Google in Hugging Face’s **$235 million** funding round in 2023, which valued the company at $4.5 billion. Hugging Face, the New York-based platform founded in 2016, hosts millions of open-source models and datasets and reports annualized revenue of $150 million. > Nvidia has held talks to acquire Hugging Face for more than $13 billion as the chip giant expands its AI dealmaking. > > — Business Insider (@BusinessInsider) [August 27, 2026](https://x.com/BusinessInsider/status/2092772800851255372?ref_src=twsrc%5Etfw) ## What We Don’t Know? The distance between the two accounts is the live question. Several others sit under it: - **Whether a signed agreement exists, or only an advanced negotiation?** - **Which price is accurate, since the two figures do not match?** - **Whether Hugging Face won any commitment on governance or hardware neutrality as a condition of sale?** - **Whether the breach that compromised Hugging Face infrastructure last month, triggered by a rogue OpenAI model, figured in diligence?** No account addresses what happens to Hugging Face’s support for competing hardware after a sale. The reporting establishes that a price was discussed. It does not establish that terms protecting the platform’s independence were. ## Why Nvidia Ownership Cuts At Hugging Face Neutrality Hugging Face’s pull with developers rests on being hardware-agnostic. The platform carries models and hardware support from across the industry, including Nvidia rivals AMD and Intel, both of which ship [competing GPU lines](https://sqmagazine.co.uk/intel-arc-gpus-not-dead-nvidia-chiplet-deal/). A vendor that sells accelerators would own the shelf its competitors sit on. Hugging Face has already drawn this line once. It rejected a $500 million Nvidia investment last year at a $7 billion valuation, saying it did not want a dominant investor able to sway its decisions. An acquisition removes the distinction that objection was built on, because ownership decides rather than influences. Teams already pulling weights from the Hub at build time can pin model revisions to specific commit hashes and mirror the artifacts they depend on. That helps reduce the risk of a hosting, licensing, or ranking change landing mid-pipeline. ## The Distribution Layer Nvidia Does Not Yet Own Major AI developers have started routing around Nvidia. OpenAI unveiled a custom chip it designed with Broadcom, and [Anthropic has weighed building its own AI silicon](https://sqmagazine.co.uk/anthropic-chip-design-team/). Both moves chip at the assumption that frontier training and inference must run on Nvidia hardware. Buying the platform at the center of the **open-source AI ecosystem** answers that from the other end. Nvidia would sit in front of the developers who decide where models actually run, which is a cheaper place to defend than the chip socket. The price stands well above the $150 million revenue line, and Nvidia can carry it: the company forecast a **70%** revenue jump next fiscal year. Developers weighing the shift can measure it against current [AI chip vendor share](https://sqmagazine.co.uk/ai-chip-statistics/) before committing a pipeline to one vendor’s hosting. ## SQ Magazine’s Takeaway The logic of the deal is easy to follow. Nvidia sells the hardware, and open-source model distribution is a layer of the stack it does not own. Owning the Hub puts it in the path of every developer choosing a model, at a moment when its largest buyers are funding alternatives to its chips. What the reporting does not show is any protection for the neutrality that made the Hub worth acquiring in the first place. The next real signal has to arrive on the record. Watch for confirmation from either company, a final price that settles the gap between the two accounts, and any published commitment on how competing hardware will be treated on the platform. If that commitment does not arrive, teams shipping on open weights should expect the platform’s incentives to shift, and [Nvidia’s inference rivals](https://sqmagazine.co.uk/groq-raises-750m-to-challenge-nvidia-inference/) will be first to say so. Definition of AI Inference. Link to full glossary entry follows the description.**AI Inference**AI inference is the execution phase where a trained AI model applies what it learned to new, unseen data and produces an output such as a prediction. [Read more](https://sqmagazine.co.uk/glossary/ai-inference/)