Support for open-weight AI models is growing across the technology industry, but companies remain divided over how to balance openness with security. According to Business Standard, major firms including NVIDIA, Microsoft, Google, and OpenAI backed a policy letter urging the U.S. government not to impose broad restrictions on open-weight AI models, arguing that open access encourages innovation, research, and competition. However, these companies have taken different approaches when it comes to formal security commitments and governance frameworks.
The article notes that NVIDIA has taken the strongest public stance in favor of open AI by launching the Open Secure AI Alliance, a coalition focused on developing open-source cybersecurity tools and security standards for AI systems. Microsoft has also supported open-weight models while investing in enterprise security and responsible AI practices. Although Google and OpenAI endorsed the industry's policy position on avoiding premature regulation, they did not join NVIDIA's new security alliance, highlighting differing views on how AI security should be managed.
Meanwhile, Anthropic has emerged as one of the leading voices urging greater caution around frontier open-weight models. While the company has clarified that it does not support banning open-weight AI altogether, it argues that the most advanced models could increase risks related to cybersecurity, biological threats, and misuse by malicious actors. Instead of broad bans, Anthropic advocates targeted safeguards such as tighter controls on advanced AI chips, stronger protections against large-scale model distillation, and rigorous safety testing before releasing highly capable models.
The article concludes that the AI industry broadly agrees that open-weight models have an important role in driving innovation, but consensus breaks down over the level of security required before releasing increasingly powerful systems. As governments consider new AI regulations and competition with China intensifies, the debate is shifting from whether open AI should exist to how it can remain both accessible and secure. The differing strategies adopted by leading AI companies are likely to shape future AI governance and the evolution of open-weight models worldwide.
