Nvidia Reportedly Nears $12.9B Hugging Face Acquisition

Introduction

Nvidia may be moving toward a deal that would reshape the business around open-source AI. The Information reported that Nvidia had agreed to buy Hugging Face for approximately $12.9 billion. Business Insider, however, said the companies had not yet signed an agreement and that the talks could still collapse. TechCrunch said it had contacted both companies and had not received a response at the time of publication. The transaction should therefore be treated as an unconfirmed negotiation, not a completed acquisition.

Key points

• A sharp valuation increase. Hugging Face was valued at about $4.5 billion after raising $235 million in 2023. A deal near $12.9 billion would represent a substantial premium over that last known valuation.
• A direct foothold in open-source AI. Hugging Face is one of the most widely used places for developers to share, download, and run open-source models. Owning the platform would give Nvidia a more direct role in model distribution, tooling, and deployment.
• A hedge around Nvidia’s chip dominance. OpenAI, Google, Amazon, and Anthropic are all working on in-house AI chips to reduce their dependence on Nvidia. A strong open-source ecosystem could give customers alternatives to closed AI providers while preserving demand for Nvidia hardware across a broader range of deployments.
• A possible return to cloud computing. Nvidia has reportedly scaled back DGX Cloud. Hugging Face already helps developers run models using rented computing resources, potentially giving Nvidia a customer and product base for re-entering cloud services without building everything from scratch.
• A way to use excess capacity. Nvidia has committed to helping customers cover cloud-computing contracts worth tens of billions of dollars. If customers do not consume all of that capacity, Hugging Face’s user base could become a channel for reselling or reallocating unused resources.

Why the timing matters

Hugging Face CEO Clem Delangue has recently aligned himself with arguments in favor of open models and has backed calls for the U.S. government not to broadly restrict open-weight systems. As Washington debates the competitive and security implications of open-source AI, the platform’s role has become more strategically important. Delangue has also said the company’s annual revenue had grown from roughly $100 million to about $150 million and that it was approaching profitability.

The company previously rejected a $500 million Nvidia investment that would have valued it at about $7 billion. One reported concern was that a dominant investor could influence Hugging Face’s decisions. A full acquisition creates a different control structure from a minority investment, which may help explain why a transaction could now be considered, though that remains an interpretation rather than a confirmed reason.

What it could mean

If completed, the deal would give Nvidia more than a model community. It would acquire an entry point connecting developers, open model weights, deployment tools, and computing services. That could help Nvidia build a broader software-and-cloud ecosystem around its chips and strengthen open-source AI as a competitor to closed model providers.

The integration would still be difficult. Nvidia would need to balance monetizing compute with preserving Hugging Face’s open and relatively platform-neutral developer culture. If users see the platform as too tightly tied to one hardware vendor, trust could suffer. Until the companies confirm an agreement, the final price, structure, and completion of the transaction remain uncertain.

Source: TechCrunch AI