Nvidia agreed on September 3, 2026, to acquire Hugging Face for $12.9303 billion, a transaction that would move the chipmaker deeper into the software and developer communities built around open artificial-intelligence models. Nvidia Chief Executive Jensen Huang announced the agreement, while Hugging Face Chief Executive Clément Delangue described the deal to CNBC as a rapid response to the need for more scale and resources. The reviewed announcements describe an agreement, not a completed closing, so the transaction’s current status is proposed and pending completion. [1][2][3]
The deal would combine Nvidia’s computing infrastructure and capital with Hugging Face’s model repository, datasets, software libraries and hosted tools. Nvidia says the platform will remain open to the wider AI ecosystem and that users will be free to choose models, frameworks, clouds, inference providers and computing platforms. Reuters reported that about $11.9 billion would go to Hugging Face investors and that a retention programme worth up to $1 billion would be offered to employees joining Nvidia. [1][2]
Key points
- Nvidia announced the agreement on September 3, 2026, at a stated price of $12.9303 billion. [1][2]
- Reuters reported approximately $11.9 billion for existing investors and an equity-based employee retention programme of up to $1 billion. [1]
- Nvidia says Hugging Face will remain open and will support multi-cloud and multi-accelerator development and deployment. [2]
- Nvidia says its chips will not be required to build on or deploy through the platform. [2]
- The sources reviewed describe a signed agreement and do not report that the acquisition has closed. [1][2][3]
What happened
Huang’s September 3 announcement said Nvidia had agreed to acquire Hugging Face for exactly $12,930,300,000. He framed the transaction as a way to scale the platform, strengthen its infrastructure and expand access to AI for developers and institutions. [2] Reuters separately reported the same total, describing it as one of Nvidia’s largest acquisitions and as a bet on open models that can be downloaded, run and customised rather than accessed only through closed services. [1]
The announcement followed discussions initiated by Hugging Face. Delangue told CNBC that his company approached Huang during the summer and that the parties reached an agreement a few weeks later. He said Hugging Face and open-source AI had reached a turning point that required more resources, scale and visibility. Those comments are the CEO’s explanation of the rationale, not an independent forecast of the deal’s outcome. [3]
Financial structure and deal status
The reported consideration has three important elements. First, the headline value is $12.9303 billion. Second, Reuters said about $11.9 billion would be paid to Hugging Face investors. Third, the agreement includes an equity-based retention programme worth up to $1 billion for staff who join Nvidia. Because the retention amount is described as “up to,” it should not be treated as a guaranteed cash payment or as a separate purchase price. [1][2]
Neither the Reuters report nor the companies’ September 3 announcement reviewed for this report says that the transaction has closed. CNBC likewise described Nvidia as having agreed to buy Hugging Face. The appropriate market status is therefore announced agreement pending completion, rather than completed acquisition. The reviewed sources do not provide a closing date, regulatory timetable or detailed conditions. [1][2][3]
Verified background and context
Hugging Face is not only a model library. Reuters described it as a platform offering models, datasets, software libraries and cloud services for building and deploying AI applications. Nvidia’s announcement said more than 18 million developers, researchers and creators use the platform to share more than 3 million models, 500,000 datasets and 1 million applications, while more than 200,000 companies use it to discover, evaluate, customise and deploy AI. Those usage figures are Nvidia’s stated figures and are presented here as company claims. [1][2]
The strategic context is a shift in where AI value is captured. Nvidia has historically been identified with accelerated computing hardware, but Reuters reported that customers including Meta, OpenAI and Microsoft were developing their own AI chips, increasing the incentive for Nvidia to broaden its role. CNBC characterised the Hugging Face deal as a move beyond hardware and further up the AI stack. That description is reporting and analysis, not evidence that Nvidia will stop prioritising chips. [1][3]
Open models also create a governance and competition question. Nvidia’s public position is that Hugging Face will support open-source and open-weight models from across the ecosystem, including work by different model builders. It also says the service will remain multi-cloud and multi-accelerator. These are commitments made by Nvidia; the sources do not establish how they would be enforced operationally after closing. [2]
Why it matters
For developers, the immediate issue is access. If Nvidia’s stated policy is implemented, users should be able to continue selecting models, frameworks, cloud services and accelerators without a formal requirement to use Nvidia hardware. That could reduce uncertainty for teams whose applications run across different vendors. It does not guarantee equal performance, pricing or feature availability across those choices, and the reviewed sources do not make such a guarantee. [1][2]
For Nvidia, ownership would provide influence over a central discovery and deployment layer rather than only over the processors used to run models. Reuters quoted analyst Axel Rudolph as saying the company was buying strategic influence as much as current earnings. A BlueRock executive cited by Reuters raised the possibility that technical methods could eventually favour Nvidia hardware despite the public assurances. That is an analyst or industry concern, not a finding that Nvidia has breached its commitment. [1]
What remains uncertain
The sources leave several material questions unanswered. They do not state when the transaction will close, whether any regulatory review is required, how the consideration will be funded in detail, or how the employee retention programme will be allocated. They also do not specify governance arrangements for Hugging Face, the duration of the hardware-neutral commitments, or an independent mechanism for resolving disputes over platform access. [1][2][3]
There is also a difference between formal openness and practical neutrality. Nvidia says developers may choose rival chips and cloud providers, but the sources do not supply service-level commitments, pricing terms, performance parity or audit procedures. Developers should therefore distinguish the company’s current public position from independently verified post-closing behaviour. [1][2]
What happens next
The next verifiable milestone is completion of the agreed transaction, if the parties satisfy its undisclosed conditions. Nvidia and Hugging Face would then need to communicate how the platform is organised, how staff retention awards operate and how multi-cloud and multi-accelerator support will be maintained. Those are implementation questions, not announced outcomes. [1][2][3]
Until a closing announcement or fuller transaction documentation is published, the most defensible reading is that Nvidia has announced a $12.93 billion expansion into open-model infrastructure, while Hugging Face remains a proposed acquisition target. The central test will be whether the promised openness survives integration in ways developers can observe and verify. [1][2]
