A petition calling for Jensen Huang to return as Nvidia’s CEO drew more than 115,000 names by Monday, November 16, MIT Technology Review reports. Elon Musk and several influential Nvidia researchers were among the reported signatories. Tom’s Hardware says Musk confirmed signing a petition on X, but it is unclear which one; it also notes that the identities on a widely shared petition have not been verified.
The counts differ: MIT Technology Review cites more than 115,000 names, while Tom’s Hardware reports one petition had passed 130,000 signatures, after also citing another with more than 91,000. The outlets do not establish whether these are the same petition. Huang’s departure has prompted supporters of open-source AI to push for his return, seeing him as a powerful voice with policymakers.
Why open-source supporters back Huang
Open-source AI models can be downloaded and run without relying on a large technology company’s service. Their supporters argue this lets startups, universities and researchers build and experiment more cheaply. MIT Technology Review notes that venture capital firms Andreessen Horowitz and Sequoia Capital, along with senior figures at Hugging Face, are among the movement’s supporters. Startups using these models may spend less on services from companies such as OpenAI and Anthropic, while still buying the computing hardware needed to train and run AI.
Nvidia’s commercial interests are part of the picture: the company sells the chips used to power AI development. MIT Technology Review also says Nvidia announced in September that it would acquire Hugging Face for nearly $13 billion. Huang has publicly argued that open-source AI helps US startups, universities and scientists compete. US Treasury Secretary Scott Bessent said last week that Nvidia should not choose an internal candidate to replace Huang, according to MIT Technology Review.
Chip-access questions complicate the push
A new concern comes from findings posted online by former OpenAI researcher Will DePue on Thursday, the day MIT Technology Review says Huang resigned. DePue suggested that Chinese AI companies had much more access to Nvidia’s advanced Blackwell chips than previously estimated. He said senior researchers at OpenAI and Anthropic checked the findings, but MIT Technology Review says no independent analysis has confirmed them. The findings do not show that Nvidia helped Chinese firms obtain chips; how the firms accessed them is unclear.
The US Department of Commerce announced a rule Monday intended to stop Chinese companies from training models remotely on restricted Nvidia chips in data centers abroad, including in Malaysia. MIT Technology Review says this kind of access has helped China’s open-source models improve. The rule could limit that route, though the supplied reporting does not say when it takes effect or how much it will reduce access. For open-source advocates, Huang’s departure may leave them without a prominent industry voice as policymakers weigh both the benefits and risks of these models.
Why it matters
For startups and researchers who use downloadable AI models, a change in access to Nvidia chips could affect how easily they can train or run models. The Commerce rule targets remote access by Chinese companies, but its start date and practical effect are not reported, and the chip-access allegations do not establish that Nvidia supplied those firms.