NVIDIA, Microsoft, and Meta spearheaded an 'Open Secure AI Alliance' on July 27, advocating for open-weight AI models. Conspicuously absent initially, OpenAI later backed the initiative. However, Anthropic's CEO, Dario Amodei, clarified on July 28 that his company's concern isn't open-weight models, but rather a call for banning AI chip sales to Chinese companies due to national security fears, setting a divergent path in the AI policy debate.