A coalition of 25 technology companies, including industry giants Nvidia, Microsoft, and Meta, has called on policymakers to refrain from implementing hasty regulations targeting open-weight artificial intelligence models. The joint letter, released Friday, warns that restrictive measures could hamper competitive innovation and push development of these technologies to other countries. Open-weight models differ from proprietary systems in that users can download, modify, and operate them independently on their own infrastructure.
The initiative comes amid growing competition from Chinese AI developers. Moonshot AI, a Chinese startup, recently unveiled a model that demonstrated superior performance compared to some leading American systems across key industry benchmarks. Concerns about Chinese technological advancement have prompted discussion about potential restrictions on access to these models within the United States, with Treasury Secretary Scott Bessent indicating the administration may investigate intellectual property theft allegations.
However, the signatory companies contend that open-weight models promote broader competition and prevent artificial intelligence capabilities from concentrating among a limited number of firms. They argue that relying exclusively on closed systems creates vulnerabilities, as proprietary models can be compromised or malfunction without external visibility. Notably, OpenAI and Anthropic—both valued at approximately one trillion dollars and preparing for potential initial public offerings—did not sign the letter, though their leaders have expressed support for increased AI accessibility.
About the Author
