Nvidia has agreed to acquire Hugging Face, the online platform where developers share and download open-source artificial intelligence models, for $12.9 billion, according to reports published on 26 and 27 August 2026. Whether the deal has formally closed, and whether regulators have been asked to approve it, is not yet known.
Hugging Face was founded in 2016 and has since grown into the dominant repository for open-source AI models: a kind of GitHub for the machine-learning world, where researchers and engineers upload, share, and build on each other's work. Acquiring it gives Nvidia a stake in a layer of the AI stack that sits well above the chips it manufactures.
That matters because the chip layer is under pressure. Nvidia currently holds roughly 85 per cent of the AI chip market, but that share faces erosion as the largest AI companies develop their own processors. OpenAI, Google, Amazon, and Anthropic are all building custom AI chips in-house. Google already uses its proprietary Tensor Processing Units, or TPUs, exclusively to train its Gemini AI models. In August 2026, Anthropic moved further in the same direction, hiring Amir Salek, a former TPU team-lead at Google, to build a new in-house chip division.
The logic of the Hugging Face deal becomes clearer in that context. A company whose customers are designing their way around its core product has reason to find new points of leverage. Hugging Face, as the place where vast numbers of developers discover and adopt AI models, is one such point. Developers who rely on the platform to find models are, in some sense, already in Nvidia's ecosystem even before they run anything on Nvidia hardware.
The acquisition sits alongside two other moves that together sketch out Nvidia's wider positioning. In July 2026, Nvidia helped lead an open letter defending open-source AI and urging Washington to leave it unrestricted. Nvidia chief executive Jensen Huang made the case personally: in what was described as his first post on X, he wrote that "open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty." Separately, Nvidia successfully lobbied Washington to loosen export restrictions on its chips to China, expanding the addressable market for its hardware.
The combination is coherent, if not straightforward. Nvidia champions open-source AI, owns the largest hub for open-source models, and sells chips to anyone allowed to buy them. The incentives align: open models spread widely, developers congregate on Hugging Face to find them, and more compute gets sold to run them.
What the deal means for Hugging Face's open-source mission is an open question. The platform's value to the developer community rests partly on its independence. Whether engineers at rival AI companies, several of whom are building chips specifically to reduce their dependence on Nvidia, will continue to use and contribute to a Nvidia-owned platform is unknown. The terms of any planned integration have not been disclosed.
Comments (0)