As Washington intensifies scrutiny of advanced artificial intelligence, Silicon Valley firms are hardening into rival camps over whether AI models should be “open-weight” or “closed-weight.” Analysts say the split tracks commercial interests more than ideology. Nvidia, the dominant AI chip supplier, corralled dozens of companies into a new alliance backing open-weight tools—an effort unveiled days after a high-profile hack roiled the policy debate. Notably missing were leading builders of closed systems such as OpenAI and Anthropic.
The emerging alignment underscores divergent business models: hardware and platform players that benefit from broad adoption and developer access versus model providers seeking to protect intellectual property, manage safety risks, and lock in customers. The clash comes as lawmakers weigh new rules for model disclosure and security, raising the stakes for an industry where control over data, distribution, and compute increasingly determines who gets paid.
Related article:
