Anthropic Backs Open-Weight AI Models but Refuses to Sign Nvidia CEO's Letter

AI company Anthropic has voiced support for open-weight AI models while declining to sign the open letter from Nvidia CEO Jensen Huang that calls for defending open-sourced AI. The seemingly nuanced position underscores the company's careful navigation of the open-source debate.

Anthropic stated that it recognizes the value of open-weight models for the AI ecosystem, arguing that weight openness promotes research transparency and innovation. This aligns with the company's previous approach of partially open-sourcing its Claude model series.

However, Anthropic did not join Huang's open letter, which the Nvidia CEO launched during his debut on X. The decision indicates that despite supporting openness in principle, the company has reservations about the specific initiative's form and content.

This stance places Anthropic on a distinct path compared to other AI players. Meta has aggressively pursued fully open-source releases, while OpenAI has moved toward closed commercial models. Anthropic is charting a middle course championing the concept of open weights without binding itself to any particular industry letter or coalition.

Analysts note that Anthropic's core mission of building safe AI systems may explain its caution. Open weights enable external auditing and safety research, but also risk misuse by bad actors.

Anthropic's position may also reflect commercial considerations. As an API-driven AI company, it needs to attract the developer community while maintaining a competitive moat.

The industry will watch whether other AI companies join Huang's open letter and whether Anthropic launches its own open-source initiative. In the rapidly evolving AI governance landscape, each company is calibrating its own approach to openness.