Nvidia Is Reportedly Buying Hugging Face for $12.9 Billion — and the Open-Weights World Is Watching

The platform that hosts the open-weights ecosystem may be about to belong to the company that sells the compute it runs on. On Wednesday, The Information reported — citing a person familiar with the deal — that Nvidia has agreed to acquire Hugging Face for $12.9 billion. Reuters, CNBC, TechCrunch, and Forbes all carried the report within hours.

Important framing first: this is a reported deal, not a confirmed one. Neither Nvidia nor Hugging Face has issued an announcement, and both declined comment when asked by CNBC and Reuters. TechCrunch notes the talks had “not yet produced a signed agreement” as of Wednesday night and could still fall apart.

What Is Actually Reported

The consistent picture across outlets:

Price: $12.9 billion, per The Information’s sourcing; Business Insider characterized talks as valuing the company at “more than $13 billion.”
Status: agreed per a source, unsigned per TechCrunch’s reporting, unconfirmed by both companies.
Context: Hugging Face’s last known valuation was $4.5 billion, from its 2023 round — which Nvidia itself participated in, alongside Salesforce Ventures, Google’s GV, IBM, AMD, Intel, and Qualcomm. The reported price is nearly 3x that.
Business scale: The Information reports Hugging Face was recently generating about $150 million in annual revenue, up from roughly $100 million two months earlier, and is “close to profitability” per CEO Clem Delangue’s July comments to TechCrunch.
History: Hugging Face previously turned down a $500 million Nvidia investment at a $7 billion valuation, telling the Financial Times it did not want a dominant investor.

Why Nvidia Would Want It

The strategic logic runs in several directions at once:

A way back into cloud. Nvidia reportedly scaled back DGX Cloud about a year ago. The Information’s reporting suggests owning Hugging Face — whose Inference Providers already route developer workloads to rented compute — gives Nvidia a channel to sell unused committed capacity rather than building a cloud from scratch. TechCrunch adds the sharper financial angle: Nvidia has promised to help cover tens of billions in customer cloud commitments, and Hugging Face could monetize capacity those customers don’t use.

Protecting the open-weights flank. Huang has spent the summer defending open models in Washington — Nvidia signed the industry letter urging against restrictions on open-weight releases, alongside Hugging Face and two dozen other companies. If open weights face regulatory pressure, owning the main distribution platform gives Nvidia a seat at the center of that fight rather than a lobbying position outside it.

Vertical integration. As one fund manager told CNBC: Nvidia wants to be “integrated in the entire stack vertically, going from energy to foundational models and also to applications.” The chips, the networking, the inference software, and now — reportedly — the model hub and its 5M+ models and datasets.

What It Would Mean for Open-Weights Developers

Hugging Face is de facto infrastructure: model hosting, datasets, Spaces, inference routing, and the hub every lab from Z.ai to Meta ships through. Three concerns will dominate discussion if the deal closes:

Neutrality. The hub’s value is that it serves every lab — including Nvidia’s competitors in silicon (AMD’s ROCm ecosystem lives there) and in models. Whether hub governance, featured-model placement, or inference-provider routing stays neutral under Nvidia ownership is the obvious question. Neither company has said anything about platform commitments.
The timing optics. The reported deal lands a month after the OpenAI agent incident in which Hugging Face was the target — an episode that produced our earlier coverage of the postmortem reports — and right as Hugging Face’s revenue was accelerating. A near-3x valuation premium over 2023 is what a strategic buyer pays when the asset is becoming critical infrastructure.
Precedent. This would be one of the largest acquisitions of the generative AI era, and it follows Stripe’s reported $7+ billion acquisition of OpenRouter earlier this month. AI infrastructure is consolidating around distribution points, not just model builders.

The counterweight: Hugging Face spent years refusing exactly this kind of dominance. Whether its independence thesis changed because of regulation risk, revenue reality, or a price too large to refuse is unknown — and until either company speaks, so is most of the deal’s substance.

What Happens Next

Watch for three signals, in order of likelihood: a joint confirmation (or a quiet collapse of the talks, which Business Insider explicitly flagged as possible); regulatory review, where a $12.9B acquisition of the open-source AI hub by the dominant AI chipmaker would draw scrutiny in both the US and EU; and — only after either of those — any statement about platform neutrality commitments, licensing, or pricing for Hub services.

Until a signed agreement or an official statement exists, treat every specific claim about post-acquisition plans as speculation. STACKDUST will follow up when either company confirms.

Update: Acquisition Formally Confirmed (September 3, 2026)

On September 2 and 3, 2026, Nvidia and Hugging Face officially announced a definitive merger agreement, followed by Nvidia filing Form 8-K (Item 8.01) with the US Securities and Exchange Commission.

The official terms confirm and clarify the transaction structure:

Total Consideration: Approximately $12.93 billion, consisting of roughly $11.9 billion in cash payable to Hugging Face stockholders and up to $1.0 billion in an equity-based retention program for employees joining Nvidia.
Closing Timeline: Expected to close in the first half of 2027, subject to customary closing conditions and regulatory approvals in the US and Europe.
Platform Neutrality Commitments: In public statements, Nvidia CEO Jensen Huang and Hugging Face CEO Clem Delangue emphasized that Hugging Face will remain an open, vendor-neutral platform. Nvidia compute will not be required to build or deploy models on Hugging Face, and the hub will continue supporting alternative accelerators, including AMD and Intel, alongside multi-cloud deployments.