Hugging Face CEO Vows to Push Harder for Open-Source AI After Jensen Huang's Nvidia Exit
Hugging Face CEO Clément Delangue said he will keep fighting for open-source AI, and will do so harder than ever, following the departure of Jensen Huang from Nvidia. TechCrunch reported the statement on November 12, 2026.
Delangue described Huang's departure as a “really sad day,” according to a post on X.
The departure lands on top of a pending transaction. In September 2026, Nvidia announced a plan to acquire Hugging Face for $12.9 billion. The deal is scheduled to close in the first half of 2027, so it has been announced but not completed. The U.S. Department of Justice has begun inquiring into how the acquisition would affect competition in the AI market.
At $12.9 billion, the deal would be Nvidia's second-largest acquisition. The largest is its $20 billion purchase of Groq's assets, completed in December 2025. Nvidia's recent deals have centered on compute, networking and software infrastructure, so a model hub with a large developer following is a different kind of asset. Hugging Face hosts the model weights (the trained numerical parameters that make an AI model function), along with datasets and tooling that a large share of the open-weight ecosystem depends on.
Huang was widely regarded as a strong advocate for open-source AI and open-weight models during his tenure as Nvidia's CEO. In July 2026, he published an AI policy letter urging the U.S. government not to over-restrict open-weight models. The letter was signed by 25 companies, including Microsoft, Meta, Hugging Face, IBM and Palantir. Hugging Face's signature meant its public position on open weights was, at that point, aligned with the Nvidia leader who is now leaving the company.
In my view, the questions that matter most for practitioners concern mechanics rather than headlines. Will the Hub's access terms, licensing presentation and model hosting policies change after a close, and if so, how? Will the DOJ inquiry alter the timeline or the conditions? The announced terms answer neither question, and it would be premature to assume an answer. What exists is a public commitment from Hugging Face's chief executive to continue the open-source effort, made in response to a leadership change at the acquirer.
The broader context here is that open-weight models have become a policy question as well as a technical one. The July letter shows that a coalition of large software and AI firms was prepared to argue for openness in a formal government submission. Whatever happens with the transaction, those companies have put their names to the position in public. The regulatory review and the scheduled close over the coming months will show whether those commitments hold through a change in ownership.
For engineers who build on open models, nothing in the code they run changes immediately. The more consequential shift is in the governance around it, and that will become clearer as the deal and the inquiry progress. Open distribution has widened the group of people who can fine-tune (retrain an existing model on their own data), evaluate and deploy capable systems without negotiating a license with a single vendor. If the stated commitment holds through the transition, that broader base of builders can keep working with frontier-class tooling. The open-source side has argued for exactly that outcome, and the next phase will test it.
