Nvidia Closes $12.93B Hugging Face Deal — Open-Source AI’s Hub Stays Open, but the Balance Shifts

Nvidia has formally agreed to acquire Hugging Face, the open-source AI model repository, for $12.93 billion. The deal, confirmed on September 3, 2026, ends weeks of speculation and gives the chipmaker ownership of the platform often called the “GitHub for AI” — a library of millions of models, datasets, and tools used by developers worldwide. Source: The Verge

The acquisition is Nvidia’s largest in years and signals a strategic pivot: instead of merely selling the GPUs that power AI workloads, the company is now buying the infrastructure where many of those workloads are discovered, shared, and iterated upon. Hugging Face is not yet profitable, but its role as the de facto distribution layer for open-weight models makes it a high-leverage asset. Source: Ars Technica

What’s staying the same — and what isn’t

Nvidia CEO Jensen Huang promised that Hugging Face “will remain an open platform for the entire AI ecosystem.” In his announcement, he explicitly said developers can choose their own models, frameworks, clouds, inference providers, and computing platforms — and that “Nvidia compute will not be required to build on or deploy through Hugging Face.” Source: The Verge

That commitment is critical for the open-source community, which had reason to be wary after Nvidia’s earlier attempt to invest $500 million at a $7 billion valuation was rejected by Hugging Face precisely because the startup did not want a dominant investor that could sway decisions. Source: Business Insider

Still, ownership changes incentives. Hugging Face’s platform is model-agnostic today; under Nvidia, the company will have to work to maintain trust that it won’t gradually steer users toward Nvidia-optimized tooling. The deal also gives Nvidia a direct relationship with millions of developers — a foothold that could help it defend against hyperscalers like Amazon and Google, which are building their own AI chips, and against frontier labs like OpenAI and Anthropic that are investing in custom hardware to reduce dependence on Nvidia. Source: Ars Technica

Beyond GPUs: robotics and the platform play

Nvidia’s advantage is already moving beyond the GPU. As AI deployments scale to gigawatt levels, orchestration hardware — networking, storage, and inference accelerators like the Groq 3 LPX — has become a differentiator. The company’s Vera Rubin architecture, now rolling out, pairs GPUs with CPUs and specialized racks that make megascale data centers more efficient. Source: TechCrunch

Hugging Face adds another layer: it has been expanding into robotics models through its Pollen Robotics subsidiary. In late August, Pollen launched the Microduck, a $399 rollerskating robot that runs on a Rockchip RK3566 processor and uses open-source software. Source: The Verge Nvidia is already a major player in physical AI, and owning Hugging Face gives it a distribution channel for robot models and simulation tools — a natural fit for its Isaac robotics platform.

The deal in context

The final price — $12.93 billion — is slightly below the $13+ billion figure reported in earlier talks, but still represents a roughly 3x premium over Hugging Face’s $4.5 billion valuation in its 2023 funding round, in which Nvidia participated. Source: Business Insider Nvidia has $47.9 billion in private holdings and committed $18 billion to equity investments for the rest of its fiscal year, so the deal is easily financed.

Microsoft also reportedly met with Hugging Face, but those talks are not ongoing. Source: Business Insider

For developers, the near-term impact is minimal: Hugging Face’s platform continues to work as before. The longer-term question is whether Nvidia can resist the temptation to use its ownership to tilt the playing field — and whether the open-source community will accept those assurances at face value.

This article updates our earlier reporting from August 27, 2026: “Nvidia's $13B Hugging Face Bid Puts a Price on Open-Source AI’s Infrastructure.”