Nvidia leadership change raises questions for open-weight AI

Nvidia CEO Jensen Huang has resigned, with chief financial officer Colette Kress named interim CEO. The leadership change has raised questions over whether Nvidia will sustain the open-weight AI priorities Huang championed, including the Nemotron model family, investments in open-model startups and a proposed acquisition of Hugging Face for $13 billion.
Huang had framed open source AI as important to safety, innovation, diffusion and sovereignty. His departure arrives while major Nvidia customers, including Google and OpenAI, are developing their own chips, increasing Nvidia’s incentive to serve a wider base of enterprises, governments and other organisations that train, fine-tune and run their own models.
Open models were part of Nvidia’s product and investment agenda
Within Nvidia, Huang pushed development of Nemotron open-weight models for long-running agentic AI systems and reasoning workloads. The company also invested in Reflection AI and Thinking Machines Lab, two startups intended to help American open models compete more effectively.
Nvidia has not finalised its proposed $13 billion deal for Hugging Face. Hugging Face CEO Clem Delangue said the company would continue to fight for open source regardless of what happens next. Nvidia separately agreed to pay $6 billion for a non-exclusive licence to the technology of AI coding startup Poolside.
The company’s public stance has also been visible in industry debate. In an open letter signed by 25 companies, including Microsoft, Meta, IBM and Palantir, Huang urged the US government not to over-restrict open-weight models, including many developed in China. That position placed Nvidia among the most prominent opponents of stronger controls advocated by closed-model providers such as Anthropic and OpenAI.
Succession puts policy and deal choices in focus
The issue is more consequential because Nvidia debate on open and closed AI models reflects Nvidia’s role in the debate over open and closed AI models, while a new leadership team must decide how strongly to maintain that role. Community members have expressed concern that a successor could take a more cautious approach to open-model support.
Huang’s resignation also adds scrutiny to his advocacy for softer chip export controls on China. The source reports that restrictions on Chinese open-weight models could reduce competition for American companies, particularly those that keep their own frontier models closed.
What businesses should watch
Nvidia’s support for an open AI ecosystem is not presented solely as an ideological commitment. A broader set of organisations training and deploying models can expand demand for GPUs and reduce Nvidia’s reliance on a small group of frontier labs. Businesses building model strategies should therefore monitor Nvidia’s decisions on Nemotron, Hugging Face and open-weight policy as practical indicators of how much continuity the interim leadership will provide.

