Key Notes
- Kevin Bass accused Anthropic and political allies of a campaign against Huang without substantiating coordination.
- James Lewis offered separate policy hypotheses and did not identify Anthropic or OpenAI as organizers.
- OpenAI signed Nvidia’s open-weights letter, while Anthropic publicly rejected a blanket ban on open models.
Jensen Huang’s departure from Nvidia (Nasdaq: NVDA) has pulled the industry’s argument over open AI models into the debate about his downfall. Commentator Kevin Bass has accused Anthropic and its political allies of using the scandal to weaken a champion of open models. His post does not substantiate that allegation.
The public policy record is more complicated than a simple battle between Nvidia and AI model developers. OpenAI is listed among the supporters of Nvidia’s open-weights letter. Anthropic has explicitly rejected a blanket ban on open models, while arguing for tighter controls on chips and mandatory safety testing.
What Bass and Lewis Actually Said
In a November 11 post on X, Bass alleged that reporting about Huang formed part of a campaign by Democrats and Anthropic to restrict open AI and entrench a regulated monopoly. The post makes an accusation about motive and coordination; it does not present documents, communications or other evidence establishing that either group directed the reporting.
Cybersecurity and technology-policy specialist James Andrew Lewis offered a different explanation. In posts on November 12, he suggested that officials opposed to the administration’s policy on China and chips might have leaked information, and that analysts might have used the controversy to defend semiconductor provisions in the National Defense Authorization Act.
Lewis described these as hypotheses. In a separate post, he argued that media coverage and national-security concerns can be used to revive a losing policy position in Washington. He did not identify Anthropic or OpenAI as organizers.
The Companies’ Published Positions
Huang has publicly defended broadly available AI models. The July 24 coalition letter hosted by Nvidia argues that downloadable models widen access, strengthen competition and give customers more control over where they run AI. OpenAI appears on the published signatory list.
That matters when assessing claims that OpenAI is trying to defeat Nvidia’s open-model agenda. Its support for the letter does not resolve every disagreement in the industry, but it complicates a portrayal of the two companies as opposing camps on model openness.
AI will transform every industry, power every company, and be built by every country.
– Jensen Huang (@JensenHuang) July 24, 2026
Anthropic CEO Dario Amodei laid out his position on July 27. He said the company had never advocated a ban on open-weights models as a category. He instead backed restrictions on advanced chips reaching authoritarian states, action against industrial-scale distillation and mandatory testing of sufficiently capable models, whether open or closed.
Amodei agreed that open models can broaden access and customer control. His dispute with the coalition concerned safety: he challenged the assumption that wider access necessarily benefits cybersecurity defenders more than attackers.
Why Openness Matters to Nvidia’s Business
Downloadable models can let businesses run AI on their own infrastructure or through a cloud provider of their choice. For Nvidia, wider deployment creates more potential demand for computing hardware. For laboratories charging for model access, capable downloadable alternatives can increase competition on price and customization.
For shareholders, the policy stakes are concrete: which models businesses may run, where the computing happens and who captures the revenue. Disagreements over safety rules can shift those advantages, even when companies remain commercial partners.
Nvidia’s immediate business test comes on November 17, when it is scheduled to report earnings under interim CEO Colette Kress. Shareholders will be listening for her plans for a company whose hardware powers both proprietary AI services and the open models that compete with them.
