Nvidia plans to acquire Hugging Face for about $12.9 billion, according to The Decoder, in a deal that would put the chipmaker in control of the central distribution hub for open AI models. The Decoder reports that Nvidia CEO Jensen Huang announced the agreement on September 3, 2026, after reports about a possible transaction surfaced in late August. The deal has not closed. According to The Decoder’s account of an SEC filing, the transaction is expected to close in the first half of 2027 and still needs regulatory approval. The reported value includes about $11.9 billion in purchase price plus a stock program of up to $1 billion intended to retain staff. The strategic asset is not just Hugging Face’s brand; it is the platform’s role in the open-model ecosystem. Nvidia says, according to The Decoder, that more than 18 million developers use Hugging Face, which hosts more than 3 million models, 500,000 datasets, and 1 million applications. Nvidia also says more than 200,000 companies rely on the service. Huang is promising continuity. The Decoder reports that Nvidia says Hugging Face will remain open and hardware-neutral, with no requirement to use Nvidia hardware and continued support for other clouds and chips. That commitment matters because Hugging Face is where many developers, companies, universities, and government users find, share, test, and deploy open models. The item also notes a tension in the origin story. Huang said Hugging Face co-founder Clem Delangue came to him while thinking about the company’s next chapter, according to The Decoder. Hugging Face co-founder Thomas Wolf, by contrast, wrote on LinkedIn that Huang made Delangue the offer to build Hugging Face into an open, independent, hardware-neutral platform. Wolf called Nvidia the best-fitting partner for the company’s mission and said nothing changes for users today, The Decoder reports. For Nvidia, the logic is tied to compute demand. The Decoder frames the deal against a shift among the largest AI providers: Google, Amazon, OpenAI, and more recently Anthropic are building their own accelerators. If the biggest closed labs increasingly move workloads onto in-house silicon, open models remain a different market: distributed across clouds, enterprises, universities, and agencies that generally do not design their own chips. Nvidia has already been active on Hugging Face. The Decoder reports that Nvidia says it is the platform’s largest contributor of open models, with more than 500 models and more than 250 datasets. Nvidia has also released Nemotron models with weights, large parts of the training data, and training and post-training recipes, and is participating in the Nemotron Coalition with companies including Mistral AI, Thinking Machines Lab, Perplexity, Black Forest Labs, Cursor, LangChain, Reflection AI, and Sarvam. The Decoder’s analysis is careful not to describe Nvidia as the open-model leader. It says Nvidia trails stronger Chinese models, though its own models can run faster in certain scenarios. Nvidia also benefits from those Chinese models by offering NVFP4 versions of Kimi K2.6 and GLM-5.1 tuned for its chips, according to the report. The second angle is hosting. Hugging Face distributes models and also rents compute to run them, giving Nvidia a potential route to demand generation as well as model distribution. For now, the essential caveat is that the transaction is still pending, the commitments to neutrality are promises, and regulators have not yet approved the acquisition. Who benefits: Nvidia would gain ownership of the main open-model hub and another channel for compute demand. Hugging Face staff could benefit from the reported retention stock program if the deal closes as described. Who's exposed: Companies, developers, universities, and agencies that rely on Hugging Face are exposed to an ownership transition at a core AI infrastructure platform. Competing cloud and chip providers will be watching Nvidia’s hardware-neutral commitment closely.