Nvidia to Acquire Hugging Face for $12.93 Billion
Nvidia agreed on September 3, 2026 to acquire Hugging Face, a major platform for AI developers, for $12.93 billion (£9.57bn), making it one of the largest deals in the chipmaker's history The Guardian.
Under the deal's terms, Nvidia will pay approximately $11.9 billion to Hugging Face's investors and offer up to $1 billion in equity-based retention packages for employees who join Nvidia as part of the transaction The Guardian. Reuters had reported on August 27 that Nvidia was in talks to acquire the company for roughly $13 billion, citing a report by The Information Reuters.
Hugging Face, founded in 2016 by Clément Delangue, Julien Chaumond and Thomas Wolf, is a New York-based startup backed by Intel, AMD and Amazon The Guardian. The platform hosts AI models — the trained programs that power applications like language translation or image generation — alongside datasets, software libraries and cloud services that developers use to build and deploy AI applications The Guardian.
Nvidia CEO Jensen Huang said Hugging Face will remain an open platform for the entire AI ecosystem and that Nvidia chips would not be required to build on or deploy through Hugging Face The Guardian. Nvidia already operates the open AI model Nemotron and has positioned itself as a vocal supporter of open AI technology The Guardian.
The broader context here matters. Think of Hugging Face as the app store of the AI world: a central hub where developers discover, share and deploy thousands of models, datasets and libraries, regardless of which hardware or cloud provider they use. Its investor roster — Intel, AMD and Amazon — reflects the breadth of that ecosystem. Nvidia's acquisition therefore places a company whose revenue is overwhelmingly tied to GPU sales in direct ownership of a platform that sits at the intersection of competing chip architectures and cloud infrastructures.
Huang's explicit commitment that Nvidia silicon will not be a prerequisite for building on Hugging Face is the most consequential guarantee in the announcement. Without it, the deal would raise immediate concerns about vendor lock-in — the scenario where a platform subtly or overtly pushes users toward one company's hardware, making it costly to switch. The pledge parallels Nvidia's stewardship of Nemotron, its own open model, and signals an intent to frame the acquisition as an ecosystem play rather than a channel strategy for its hardware business. Whether that commitment holds in practice, and how regulators and rival chipmakers interpret it, will shape the competitive dynamics of the AI infrastructure layer for years.
The deal also tightens Nvidia's grip on the full AI development stack — the chain of tools from the chips that train models, through the platforms where models are shared, to the services that deploy them in real applications. Nvidia already dominates GPU-based training and inference (the process of running a trained model to produce results). Adding Hugging Face gives it ownership of the distribution layer where models are discovered, shared and deployed. That vertical integration, from silicon through platform to deployment, is the structural shift to watch, regardless of Huang's open-platform assurances.
Several questions remain unresolved by the announcement. The deal's regulatory path, particularly in the European Union and the United States, will test antitrust authorities' appetite for consolidation across the AI stack. Competitors who rely on Hugging Face as a neutral distribution venue, including AMD and Intel, face a new relationship with their own platform partner's new owner. And the retention programme, while sizable, offers no guarantee that key engineering talent will stay through integration.
The transaction was announced on Thursday, September 3, 2026 The Guardian.
