Top AI companies and executives are pushing back against a potential Trump administration move to ban Chinese open-weight models, arguing the policy could undermine cybersecurity, safety, and innovation. The dispute centers on the claim that open models could strengthen defensive capabilities and reduce fragmentation across ecosystems—while government officials are weighing concerns about cybersecurity incidents and the availability of high-capability tools. Reporting says Nvidia, Meta, Microsoft, OpenAI and others signed an open letter titled “Open Weights and American AI Leadership.” The letter argues for keeping open models available and frames broader availability as a pathway to improved safety, faster innovation, and “sovereignty.” Meanwhile, open-source advocates and researchers are also calling for more transparency around how safety frameworks apply across model access and licensing. For universities experimenting with model access, the outcome could affect procurement policies, campus lab compliance, and guidance for research workflows that rely on open-weight tools.