NVIDIA announced the Open Secure AI Alliance, uniting industry leaders, academics, researchers, and developers to build and share open tools and techniques that improve AI software, agent security, and responsible use. The move favors an ecosystem of shared tools over monopoly lock-in, citing risks to national security and community defense when models remain closed. Supporters note that open access strengthens forensics and collective security work, as seen in recent incidents involving frontier models. Founding members include NVIDIA and partners focused on collaborative AI safety efforts.