A petition with converging interests
Since Friday, a petition to return Jensen Huang to the helm of Nvidia has been circulating on social media; by Monday, the initiative had gathered more than 115,000 names, including Elon Musk and several influential researchers at the company, as MIT Technology Review reports. Among the signatories, however, one specific industry contingent stands out: executives from Hugging Face, a hub for sharing AI models, and representatives of the Andreessen Horowitz and Sequoia Capital funds.
According to the source's account, these names have a direct interest in the development of open-source AI—that is, models that can be downloaded and used independently of the big technology companies, unlike ChatGPT or Claude. Supporters of this approach say that concentrating AI in the hands of a few players would harm everyone else; the source observes that the concern would resemble that of industry critics, if many of these supporters were not also investors who would benefit from startups able to build on cheaper models without paying the big tech companies.
Huang, MIT Technology Review writes, had become a sort of reference point for those defending open source before policymakers, against fears that such models lack safeguards and that they cede AI leadership to China, where the most powerful and widely used open-source models are born.
Chinese chips and the Commerce rule
It would not be only the personal circumstances cited by the source that undermine Huang's credibility as an open-source advocate. On Thursday, the day of his resignation, former OpenAI researcher Will DePue published online findings according to which Chinese AI companies had gained access to Nvidia's advanced Blackwell chips to a much greater extent than had been estimated. China cannot access the company's most cutting-edge chips, a measure designed to protect the U.S. lead in the sector; it is unclear how the chips would have been obtained. DePue maintains that the findings were verified by senior researchers at OpenAI and Anthropic, but an independent analysis did not confirm them; neither DePue nor Nvidia responded to requests for comment.
The source specifies that the conclusions do not demonstrate Nvidia's involvement in supplying the chips and that the broader smuggling phenomenon was already known and is beyond the company's control. However, by suggesting that Nvidia may have profited from the rise of Chinese startups, whose models are increasingly used by American companies in place of domestic alternatives, the findings put Huang in an awkward position with respect to his defense of open source as an instrument of American technological strength.
The source also cites a new rule announced Monday by the U.S. Department of Commerce, designed to prevent Chinese companies from training models on Nvidia chips accessible remotely in data centers in countries such as Malaysia: a shortcut that, according to the account, has contributed to the improvement of Chinese open-source models. In the same context, the source recalls that in September Nvidia announced the acquisition of Hugging Face for nearly $13 billion and that its support for open source also responds to a direct economic interest: the more companies build their own models instead of paying Anthropic and OpenAI, the more Nvidia chips it sells.
A return that is far from assured
U.S. Treasury Secretary Scott Bessent said last week that Nvidia should not consider an internal candidate to replace Huang, according to the source. The new indications about chips would further reduce the chances of his return. The question posed by the same source remains open: if Huang exits the scene, who will take on his role as interlocutor with policymakers, on a front that appears internally divided over the rise of Chinese open-source models.
