Anthropic Faces Backlash for Failing to Endorse Open-Weights

In Focus

• Debate about open-weight AI models started after China’s Moonshot unveiled Kimi K3
• Anthropic prefers gated access to frontier AI models
• Major U.S. tech companies have endorsed open-source AI

Anthropic has attracted criticism for failing to endorse open-weight AI. Several Silicon Valley tech companies criticized the Claude Code developer as the U.S. considers restrictions on certain Chinese AI Models. Anthropic faced open-source AI backlash weeks after the U.S. government ordered it to restrict foreign access to its advanced AI models. The company restored access on July 1 after the government lifted the directive.

Which Tech Companies Support Open Weights?

More than two dozen tech companies signed the “Open Weights and American AI Leadership” letter last week while others did so over the weekend. The companies include Meta, Nvidia, OpenAI, Google, Microsoft and SpaceX. The open-weights endorsement means that developers can download, edit, and deploy AI models on their infrastructure.

By July 27, Anthropic was the only leading frontier AI company that had not endorsed the industry-wide campaign to open-source AI models. The company has faced serious backlash online for its decision.

“The entire tech industry (save for Anthropic) has come out in favor of open source AI. They won’t stop until they kneecap open source. The rest of the industry needs to watch these guys like a hawk,” Former AI czar for the Donald Trump administration, David Sacks said in a post on X.
OpenClaw founder, who now works at OpenAI, Peter Steinberger also criticized Anthropic’s reluctance to sign the open weights letter on X.

What Triggered the Open-Weight Debate?

The discussion about open-weight AI models was triggered by the unveiling of the Kimi K3 AI model. Developed by Chinese startup Moonshot AI, the open-weight AI model rivals leading closed models’ performance in certain benchmarks.

China has adopted a more open approach to AI compared to the US. Last week, some senior official in President Donald Trump’s administration accused Moonshot of AI distillation practices. The official claimed that the Chinese startup trained Kimi K3 by distilling Anthropic’s Fable AI model.

Treasury Secretary Scott Bessent indicated that the U.S. could impose sanctions on Chinese open-source AI models that distill American AI models. Such a move could deepen the AI rivalry between the two, where approaches to open-source AI development increasingly diverge.

The open-weight AI debate accelerated after an OpenAI AI agent hacked Hugging Face systems autonomously. Last week, Hugging Face said it relied on a model from China’s Z.ai to defend its systems against OpenAI’s rogue AI agent as restrictions prevented access to closed AI models.

Why is Anthropic Not Open-Source?

Previously, CEO Dario Amodei has supported Anthropic’s closed source AI. Amodei argued that developers effectively lose control of frontier AI models once their weights become publicly available.

Instead, the Claude Code developer prefers gated access to frontier AI models. Anthropic adopted this approach with Claude Mythos when it launched the Project Glasswing initiative. Under this initiative, the AI developer gives early Mythos access to vetted companies.

Anthropic’s decision to avoid endorsing open weights places the company at the centre of a growing debate over the future of AI development. While rivals push for greater openness, Anthropic is betting that controlled access and stronger safeguards are the better path for managing advanced AI risks.