Nvidia Buys Hugging Face for $12.9 Billion

Nvidia announced on September 3, 2026 that it is acquiring Hugging Face, the leading open-source AI platform, for roughly $12.9 billion. The deal, confirmed on Nvidia's official blog and reported by Bloomberg, CNBC and TechCrunch, reshapes an ecosystem that thousands of SMEs rely on to host, test or deploy open-weight AI models.

In brief

Nvidia is acquiring Hugging Face for $12.9 billion, roughly $11.9 billion paid to shareholders plus up to $1 billion set aside for employee retention.
Hugging Face counts more than 18 million developers, researchers and creators, and more than 200,000 companies use the platform to deploy AI.
Nvidia has pledged to keep Hugging Face open and multi-cloud: no requirement to run Nvidia hardware to host or deploy a model.
The deal is expected to close in the first half of 2027, pending regulatory approval.
For SMEs using Hugging Face models or Spaces, nothing changes immediately, but relying on a single dominant player across the AI stack is worth watching.

What Nvidia is buying, and for how much

Over a few years, Hugging Face became the central hub of open-source AI: a software library (transformers), a host for models and datasets, and a marketplace where researchers and companies publish their work. According to Nvidia's announcement, the platform hosts more than 3 million models, 500,000 datasets and 1 million applications (its "Spaces").

$12.9Bdeal value, including $1B set aside for employee retention18M+developers, researchers and creators active on Hugging Face200,000+companies deploying AI through the platform

According to CNBC, Hugging Face's own CEO approached Nvidia's Jensen Huang weeks before the deal was signed. For Nvidia, this ranks among the largest acquisitions in its history, part of a broader strategy to control more of the AI value chain: chips, cloud infrastructure and now the software layer where models are published and shared.

Part of a wider industry shift

This deal is not an isolated move. Since 2025, the sector's biggest players have been pushing to control more links in the AI chain, from silicon all the way to model distribution.

Aug 26, 2026

Talks reported

TechCrunch reports that Nvidia is close to a deal to acquire Hugging Face.

Sep 3, 2026

Deal confirmed

Nvidia and Hugging Face confirm the $12.9 billion deal on their respective blogs.

H1 2027

Expected close

The deal is expected to close, pending regulatory approval.

Nvidia stresses one point in its announcement: Hugging Face is meant to stay a neutral platform. Developers will keep the freedom to choose their models, frameworks, cloud providers and chips without being required to use Nvidia hardware. It is a strong commitment, one that will need to hold up over time once the deal closes.

What this actually changes for SMEs

Many SMEs use Hugging Face without necessarily realizing it: a transcription model embedded in an internal tool, a demo Space tested before a project, an inference API called from an n8n or Make workflow, or a self-hosted open-source model chosen to avoid dependence on a proprietary API.

Before the deal

An independent ecosystem, funded through standard venture rounds. Roadmap set by the community and Hugging Face's existing investors.

After the deal (expected)

Access to stronger funding and GPU infrastructure. Nvidia commits to keeping the platform open and multi-cloud, with deeper integration into its own software stack (NIM, CUDA).

Two readings coexist. On the upside: a better-funded platform, potentially more reliable and better equipped for enterprise use, with easier access to the latest Nvidia hardware optimizations. On the cautious side: concentrating such a central piece of open-source AI infrastructure under the world's leading AI chipmaker raises questions about the platform's long-term neutrality, and about possible pricing changes once the deal closes.

Key takeaway for SMEs

If you use Hugging Face models, datasets or Spaces in production, expect no immediate changes before the deal closes, likely in the first half of 2027. Use this window to document your technical dependencies (which models, which APIs, which costs) and identify alternatives if needed.

What to watch before the deal closes

Three things are worth tracking over the coming months: the outcome of the antitrust review (the deal concentrates two already-dominant players), whether the openness pledge holds once integration actually begins, and any future changes to Hugging Face's paid services (model hosting, Inference Endpoints, private Spaces). For an SME, the safest posture is to keep using the tools that work today while keeping a clear record of what might need to be replaced if the platform's policies change.

FAQ

What is Hugging Face?

Hugging Face is an open-source AI platform hosting models, datasets and demo applications ("Spaces"). It is used by more than 18 million developers and more than 200,000 companies to find, test and deploy AI models.

Why is Nvidia buying Hugging Face?

According to Nvidia's official announcement, the goal is to strengthen the platform's infrastructure and expand access to AI for developers and institutions worldwide. The deal is part of a broader Nvidia strategy to control more of the AI value chain, from chips to model distribution.

Will Hugging Face stay free and open?

Nvidia says it intends to keep Hugging Face open, multi-cloud and multi-accelerator, without requiring users to run Nvidia hardware. This commitment, made at signing, will need to be verified over time as the two companies integrate.

When will the acquisition close?

The deal is expected to close in the first half of 2027, pending regulatory approval across the relevant jurisdictions.

This deal shows how far the open-source AI ecosystem, long seen as neutral ground, is now being pulled into the industry's larger power dynamics. To go further on choosing between proprietary and open-source AI, see our guide to open-source AI models or explore our AI resources for SME leaders.